UNDERSTANDING INSTITUTIONAL DIVERSITY
  
  
UNDERSTANDING INSTITUTIONAL
DIVERSITY
Elinor Ostrom
P R I N C E T O N U N I V E R S I T Y P R E S S P R I N C E T O N A N D O X F O R D
  
Copyright © 2005 by Princeton University Press
Published by Princeton University Press, 41 William Street,
Princeton, New Jersey 08540
In the United Kingdom: Princeton University Press,
3 Market Place, Woodstock, Oxfordshire OX20 1SY
All Rights Reserved
Library of Congress Cataloging-in-Publication Data
Ostrom, Elinor.
Understanding institutional diversity / Elinor Ostrom.
p. cm.
Includes bibliographical references and index.
ISBN-13: 978-0-691-12207-6 (cloth : alk. paper)
ISBN-10: 0-691-12207-5 (cloth : alk. paper)
ISBN-13: 978-0-691-12238-0 (pbk. : alk. paper)
ISBN-10: 0-691-12238-5 (pbk. : alk. paper)
1. Diversity in the workplace. 2. Multiculturalism. 3. Interpersonal
relations. 4. Organizational behavior. 5. Social norms. I. Title.
HF5549.5.M5O88 2005
302.3′5—dc22 2004065772
British Library Cataloging-in-Publication Data is available
This book has been composed in Sabon
Printed on acid-free paper. ∞
pup.princeton.edu
Printed in the United States of America
10 9 8 7 6 5 4 3 2 1
  
To the fabulous colleagues, students, visiting scholars,
and staff who have made scholarship at the Workshop
over the past thirty years highly productive, challenging,
and rewarding
  
  
Contents
List of Illustrations xi
Acknowledgments xiii
PART I: AN OVERVIEW OF THE INSTITUTIONAL ANALYSIS
AND DEVELOPMENT (IAD) FRAMEWORK 1
One
Understanding the Diversity of Structured Human Interactions 3
Diversity: A Core Problem in Understanding Institutions 4
Is There an Underlying Set of Universal Building Blocks? 5
Holons: Nested Part-Whole Units of Analysis 11
Action Arenas as Focal Units of Analysis 13
Zooming Out to an Overview of the IAD Framework 15
Viewing Action Arenas as Dependent Variables 16
Institutional Frameworks, Theories, and Models 27
The Limited Frame of This Book 29
Two
Zooming In and Linking Action Situations 32
An Action Situation as a Focal Unit of Analysis 32
Example of a Simple Action Situation 35
The Basic Working Parts of Action Situations 37
Linking Action Arenas 55
Predicting Outcomes 64
Evaluating Outcomes 66
Three
Studying Action Situations in the Lab 69
The Trust Game in the Experimental Laboratory 70
A Commons Dilemma in the Experimental Laboratory 78
Structural Changes in the Laboratory 85
Replications and Extensions of Commons Dilemma
Experiments 93
Conclusions 97
Four
Animating Institutional Analysis 99
Animating Open, Competitive Processes 100
   C ON TE NT Sviii
The Challenge of Imperfect Information 101
Assumptions Used in Animating Participants 103
Variety and Complexity: An Asset or a Liability? 116
A Focus on Collective Action to Overcome Social Dilemmas 119
Norms Fostering Collective Action 121
Emergence and Survival of Norms in Evolutionary Processes 125
Conclusion 131
PART II: FOCUSING ON RULES 135
Five
A Grammar of Institutions, Sue Crawford and Elinor Ostrom 137
Parsing Institutional Statements 137
The Syntax of a Grammar of Institutions 139
The Syntax Components 140
Applying the Grammar 152
Using the Grammar in Empirical Field Research 171
Some Next Steps 173
Six
Why Classify Generic Rules? 175
Solving Babbling Equilibrium Problems 176
The Policy Analyst’s Need to Understand How to
Reform Situations 180
Moving beyond Slogan Words to Describe Institutions 181
Coping with the Immense Diversity by Identifying
Generic Rules 181
The Role of Rules as Information Transformation
Mechanisms 184
An Underlying Universality? 185
Seven
Classifying Rules, Elinor Ostrom and Sue Crawford 186
The Horizontal Approach: Classifying by the AIM of a Rule 187
Position Rules 193
Boundary Rules 194
Choice Rules 200
Aggregation Rules 202
Information Rules 206
Payoff Rules 207
Scope Rules 208
Default Conditions: What Happens if No Rules Exist
Related to Components of an Action Situation? 210
  C ON TE NT S ix
The Vertical Approach: Operational, Collective-Choice,
and Constitutional-Choice Levels of Analysis 214
Using Rules as Tools to Change Outcomes 215
PART III: WORKING WITH RULES 217
Eight
Using Rules as Tools to Cope with the Commons 219
Field Research on Common-Pool Resources 221
What Rules Are Found in Self-Organized Common-Pool
Resource Regimes? 222
Contemporary Approaches to Resource Policy 236
Coping with Complexity: A General Problem 242
Changing Rules as an Adaptive Process 243
Theoretical Puzzles 251
Summing Up 253
Nine
Robust Resource Governance in Polycentric Institutions 255
Design Principles and Robust Social-Ecological Systems 258
Threats to Robust Governance of Common-Pool Resources 271
Modest Coping Methods for Dealing with Threats to
Sustainability 279
The Advantage and Limits of Polycentric Systems in Coping
with Design and Long-Term Sustainability of Systems 281
The Capabilities of Polycentric Systems in Coping with
Tragedies of the Commons 283
Conclusion 287
Notes 289
References 307
Index 351
  
  
Illustrations
Figures
1.1 The focal level of analysis—an action arena 13
1.2 A framework for institutional analysis 15
1.3 Four basic types of goods 24
2.1 The internal structure of an action situation 33
2.2 The Snatch Game 36
2.3 Levels of analysis and outcomes 59
2.4 Relationships of formal and informal collective-choice arenas 62
3.1 The Trust Game 71
3.2 Trust and reciprocity under three experimental conditions 77
4.1 The relationship between information, action-outcome
linkages, and internal mental models 105
4.2 The impact of communication, vividness, and salience on
the relationship between information, action-outcome
linkages, and internal mental models 108
4.3 The Snatch Game with norms 122
5.1 Authorized relationships: authority to act 145
5.2 Base game 156
5.3 Repeated game with shared strategies 157
5.4 Game with a norm and monitoring 159
5.5 Equilibrium diagram: game with a norm and monitoring 160
5.6 Game with a rule 163
7.1 Rules as exogenous variables directly affecting the elements
of an action situation 189
7.2 The Snatch Game with minimal property rights 213
Tables
2.1 Action sets assigned to positions in several “typical” action
situations 46
2.2 Results if collective-choice rules specify that a tie vote
retains the status quo 47
2.3 Results if collective-choice rules specify that a tie vote leads
to a change 47
3.1 Summary results across environments 90
4.1 Description of various decision strategies 115
   I LL US TR AT IO NSxii
5.1 Game summaries 155
5.2 Delta parameters and normative concepts used in recent
literature 169
6.1 Shared strategies, norms, and rules as used in ADICO syntax
and in recent literature 178
7.1 The AIM component of each type of rule 191
7.2 Default conditions 211
7.3 Rules changed to create elementary property rights for
agricultural commodities 212
8.1 Attributes and conditions used in boundary rules to define
who is authorized to appropriate from a common-pool
resource 224
8.2 Choice rules used to allocate common-pool resources 229
8.3 Water allocation rules in Chitwan and Tanahun districts 232
  
Acknowledgments
ANY BOOK DRAWING on unpublished and published papers that an author
has drafted and thought about for multiple decades requires many acknowledgments.
As I mention in several places in the volume, sections of
this book were started in the 1980s when I was a research scholar at the
Center for Interdisciplinary Research at Bielefeld University in Germany.
I was fortunate enough to visit Bielefeld during 1982 and again in 1988.
The opportunity to express some of my early thinking about nested sets of
rules and action situations with colleagues participating in both Bielefeld
groups is deeply appreciated and helped me immensely. I can remember
the puzzlement of several colleagues when I decided to entitle a lecture
given at Bielefeld in 1982 “The Hidden Structure behind the Structure.”
In some respects, one may think of this book as my effort to uncover that
hidden structure and to answer their many questions.
My major home base throughout this entire period, of course, has been
the Workshop in Political Theory and Policy Analysis at Indiana University,
Bloomington. Many earlier draft sections of what is now in this book
have been presented at our colloquiums, at miniconference sessions, and
in various working group meetings (e.g., the CPR Rules Coding Group,
the Evolutionary Theory Working Group) at the Workshop. It has been
an intellectual home for me where one could present radical ideas and get
them taken seriously at the same time as they are being challenged and
debated. The support of the Workshop by Indiana University is deeply
appreciated.
Vincent Ostrom cofounded the Workshop with me more than thirty
years ago, and his ideas permeate all that we do at the Workshop. None
of us could have achieved what we have done without his inspiration,
criticism, encouragement, and help, and there is no way of really adequately
thanking him for years of lively discussions about the theoretical
issues discussed in this book.
Throughout this period, we have been fortunate to receive funding
from the National Science Foundation (SES-8619498; SES-8921888;
SBR-9521918; SES-0083511), the Ford Foundation (950–1160–1), and
the MacArthur Foundation (00–63798-GSS). Work with colleagues in
the Resilience Alliance (supported by the McDonnell Foundation) has
proved important during recent years and has helped to stimulate work
on the last sections of the book. I was fortunate to spend a month with
colleagues at the Max Planck Center for the Study of Common Goods in
Bonn, Germany, during 2001. The staff there was very helpful and as-
   A CK NO WL ED GM EN TSxiv
sisted me in getting a number of notes that had been written in the 1980s
into text form.
In recent years, I have given a number of seminars at different universities
on various sections of this book. These include: the Sanford School
of Public Policy at Duke University in November 2001; the Department
of Political Science, University of California, San Diego, in March 2002;
as part of the Walker-Ames Lecture Series at the University of Washington
in April 2002; Kathmandu University in April 2002; the Department of
Economics and the Complex Systems Group at the University of Michigan
in November 2002; the Max Planck Institute in Jena, Germany, in
June 2003; Purdue University in October 2004; Cornell University in November
2004; Case Western University in January 2005; and at the Workshop
itself on October 18, 2004, as well as in my seminar during the fall
semesters of 2003 and 2004.
During the many years that colleagues and I have been trying to understand
institutions, the occasion to write a number of papers and chapters
have enabled coauthors and me to develop initial versions of many of the
arguments in this volume. Thus, there are partial sections of earlier papers
that are drawn on in multiple chapters of this volume, including:
Chapters 1 and 2: Elinor Ostrom, “Institutional Rational Choice: An
Assessment of the Institutional Analysis and Development Framework,”
in Theories of the Policy Process, ed. Paul A. Sabatier (Boulder, CO: Westview
Press, 1999), pp. 35–71; and Elinor Ostrom, “Doing Institutional
Analysis: Digging Deeper Than Markets and Hierarchies,” in the Handbook
of New Institutional Economics, ed. Claude Me´nard and Mary Shirley
(Dordrecht, the Netherlands: Springer, 2005), pp. 797–926.
Chapter 4: Elinor Ostrom, “Collective Action and the Evolution of Social
Norms,” Journal of Economic Perspectives 14 (3) (2000): 137–58.
Chapter 5: Sue E. S. Crawford and Elinor Ostrom, “A Grammar of Institutions,”
American Political Science Review 89 (3) (September 1995): 582–
600. (Reprinted with the permission of Cambridge University Press.) Reprinted
in Polycentric Games and Institutions: Readings from the Workshop
in Political Theory and Policy Analysis, ed. Michael McGinnis (Ann
Arbor: University of Michigan Press, 2000), pp. 114–55. Figures and tables
reprinted with the permission of the University of Michigan Press.
Chapter 8: Elinor Ostrom, “Reformulating the Commons,” in Protecting
the Commons: A Framework for Resource Management in the
Americas, ed. Joanna Burger, Elinor Ostrom, Richard B. Norgaard, David
Policansky, and Bernard D. Goldstein (Washington, D.C.: Island Press,
2001), pp. 17–41.
Chapters 8 and 9: Elinor Ostrom, “Coping with Tragedies of the Commons,”
Annual Review of Political Science 2 (1999): 493–535. (Reprinted
with permission of the Annual Review of Political Science.)
  A CK NO WL ED GM EN TS xv
Chapter 9: Elinor Ostrom, “Institutional Analysis, Design Principles,
and Threats to Sustainable Community Governance and Management of
Commons,” in Law and the Governance of Renewable Resources: Studies
from Northern Europe and Africa, ed. Erling Berge and Nils Christian
Stenseth (Oakland, CA: ICS Press, 1998), pp. 27–53.
As I work on this book, I am deeply appreciative of the long conversations
that I have had with colleagues about the Institutional Analysis and
Development (IAD) framework, which many of us have had a direct role
in creating. The many conversations and memos shared with Larry Kiser
and Sue Crawford are particularly important in affecting my own thinking,
and working with Roy Gardner and James Walker to model these
ideas using game theory and build experimental games to test our predictions
has been invaluable. Roger Parks has repeatedly given me extensive
comments on earlier manuscripts that have challenged me to work still
harder to explain the concepts. Marco Janssen and Mike McGinnis have
both read and commented extensively on sections of this manuscript.
Working with T. K. Ahn on multiple papers has also added to my understanding.
Chuck Myers at Princeton University Press has given me many
useful suggestions and has been extremely helpful through the final year
of moving an amorphous manuscript toward being a real book. Cindy
Crumrine, who copyedited this book for Princeton University Press, did
an outstanding job. The anonymous reviewers for Princeton University
Press also provided useful advice. Eric Coleman has read through the final
draft with a critical and useful eye and developed the index for this book.
Without Patty Lezotte’s excellent editing skills, thoughtful insights, and
cheerful help under high stress, I do not know how I could have finished
this book.
In addition, there have been a number of people who have given one
or more chapters a very serious read and given me comments. These include:
Ryan Adams, Arun Agrawal, Alexander Alexeev, Marty Anderies,
Krister Andersson, Ed Araral, Robert Axelrod, Brian Bartels, Avner Benner,
Robert Bish, Juan-Camilo Cardenas, Robert Christensen, Daniel
Cole, Eric Coleman, Meritxell Costeja, Shanon Donnelly, Ginger Elliot,
Ernst Fehr, Susan Fitzpatrick, Bruno Frey, Norman Frohlich, Jessica
Gerrity, Gerd Gigerenzer, Werner Gu¨th, Tobias Haller, Grant Hemphill,
Pamela Jagger, Kenneth Koford, Kudan Kumar, Wai Fung Lam, Joe
Oppenheimer, Mike Radcliffe, Eric Rasmusen, Bo Rothstein, Fritz
Scharpf, John Schiemann, Edella Schlager, Ganesh Shivakoti, Shui Yan
Tang, Ingela Ternstrom, Catherine Tucker, Frank Van Laerhoven,
Tommaso Vitale, James Wilson, Lihau Yang, and Eric Zeemering.
  
  
Part I
A N O V E RV I E W O F T H E I N S T I T U T I O N A L
A N A LY S I S A N D D E V E L O P M E N T ( I A D )
F R A M E W O R K
  
  
One
Understanding the Diversity of Structured
Human Interactions
TO UNDERSTAND institutions one needs to know what they are, how and
why they are crafted and sustained, and what consequences they generate
in diverse settings. Understanding anything is a process of learning what
it does, how and why it works, how to create or modify it, and eventually
how to convey that knowledge to others. Broadly defined, institutions are
the prescriptions that humans use to organize all forms of repetitive and
structured interactions including those within families, neighborhoods,
markets, firms, sports leagues, churches, private associations, and governments
at all scales. Individuals interacting within rule-structured situations
face choices regarding the actions and strategies they take, leading
to consequences for themselves and for others.
The opportunities and constraints individuals face in any particular
situation, the information they obtain, the benefits they obtain or are excluded
from, and how they reason about the situation are all affected by
the rules or absence of rules that structure the situation. Further, the rules
affecting one situation are themselves crafted by individuals interacting
in deeper-level situations. For example, the rules we use when driving
to work every day were themselves crafted by officials acting within the
collective-choice rules used to structure their deliberations and decisions.
If the individuals who are crafting and modifying rules do not understand
how particular combinations of rules affect actions and outcomes in a
particular ecological and cultural environment, rule changes may produce
unexpected and, at times, disastrous outcomes.
Thus, understanding institutions is a serious endeavor. It is an endeavor
that colleagues and I at the Workshop in Political Theory and Policy Analysis
have been struggling with for at least three decades.1
After designing
multiple research projects; writing numerous articles; developing ideas in
the classroom; learning from eminent scholars in the field, from students,
and from colleagues; and making diverse attacks on this problem, it is
time to try to put thoughts on this subject together within the covers of a
book, even though I am still not fully satisfied with my own understanding.
Consider this a progress report on a long-term project that will be
continued, I hope, by many others into the future.
  
C HA PT ER ON E4
Diversity: A Core Problem in Understanding Institutions
A major problem in understanding institutions relates to the diversity of
situations of contemporary life. As we go about our everyday life, we
interact in a wide diversity of complex situations. Many of us face a morning
and evening commute where we expect that others, who are traveling
at great speeds, will observe the rules of the road. Our very lives depend
on these expectations. Others depend on our own driving behavior conforming
in general to locally enforced rules about speeding, changing
lanes, and turn-taking behavior at intersections. Those of us who work
in large organizations—universities, research centers, business firms, government
offices—participate in a variety of team efforts. In order to do
our own work well, we are dependent on others to do their work creatively,
energetically, and predictably, and vice versa. Many of us play
sports at noontime, in the early evening, or on the weekends. Here again,
we need to learn the basic rules of each of the games we play as well as
find colleagues with whom we can repeatedly engage in this activity. During
the average week, we will undertake activities in various types of market
settings—ranging from buying our everyday food and necessities to
investing funds in various types of financial instruments. And we will
spend some hours each week with family and friends in a variety of activities
that may involve worship, helping children with homework, taking
care of our homes and gardens, and a long list of other activities undertaken
with family and friends.
Somehow as individuals we implicitly make sense of most of these diverse
and complex situations. We do so even today, with all of the new
opportunities and risks that were not even conceivable a few generations
ago. We now expect to watch the Olympic games and other international
competitions as they happen, no matter where they are located or where
we are in the world. We have become accustomed to buying bananas,
oranges, and kiwi fruit at any time of the year in almost any market we
enter around the globe. Not only do millions of us drive to work regularly,
many of us also fly to other parts of the globe on a regular basis, trusting
our lives to the knowledge and skills of pilots to know and utilize the
many do’s and don’ts of flying airplanes.
If we are considered to be adults and sane, we are expected to be able
to reason about, learn, and eventually know what to do in many diverse
situations that we confront in today’s world. We know that when we are
shopping in a supermarket that we can take a huge variety of goods off
the shelf and put them in a pushcart. Before we put these same goods in
our car, however, we need to line up at a counter and arrange to pay for
them using cash or a credit card (something else that was not so widely
available a few years ago). When we are shopping in an open bazaar in
  
D IV ER SI TY AN D S TR UC TU RE D I NT ER AC TI ON S 5
Asia or Africa, however, the do’s and don’ts differ. If we go at the end of
the market day, we may bargain over the price of the fruit that is left on
the stand—something we could never do in a supermarket where fruit
will be refrigerated overnight. If we are in the household goods section
of the bazaar, vendors would be astounded if we did not make several
counteroffers before we purchased an item. Try that in a furniture store
in a commercial district of a Western country, and you would find yourself
politely (or not so politely) told to leave the establishment. Thus, there
are many subtle (and not so subtle) changes from one situation to another
even though many variables are the same.
These institutional and cultural factors affect our expectations of the
behavior of others and their expectations of our behavior (Allen 2005).
For example, once we learn the technical skills associated with driving a
car, driving in Los Angeles—where everyone drives fast but generally follows
traffic rules—is quite a different experience from driving in Rome,
Rio, and even in Washington, D.C., where drivers appear to be playing a
bluffing game with one another at intersections rather than following traffic
rules. When playing racquetball with a colleague, it is usually okay to
be aggressive and to win by using all of one’s skills, but when teaching a
young family member how to play a ball game, the challenge is how to
let them have fun when they are first starting to learn a new skill. Being
too aggressive in this setting—or in many other seemingly competitive
situations—may be counterproductive. A “well-adjusted and productive”
adult adjusts expectations and ways of interacting with others in situations
that occur in diverse times and spaces.
Our implicit knowledge of the expected do’s and don’ts in this variety
of situations is extensive. Frequently, we are not even conscious of all of
the rules, norms, and strategies we follow. Nor have the social sciences
developed adequate theoretical tools to help us translate our implicit
knowledge into a consistent explicit theory of complex human behavior.
When taking most university courses in anthropology, economics, geography,
organization theory, political science, psychology, or sociology, we
learn separate languages that do not help us identify the common work
parts of all this buzzing confusion that surrounds our lives. Students frequently
complain—and justifiably so—that they have a sense of being in a
Tower of Babel. Scholars also see the same problem (V. Ostrom 1997, 156).
Is There an Underlying Set of Universal Building Blocks?
The core questions asked in this book are: Can we dig below the immense
diversity of regularized social interactions in markets, hierarchies, families,
sports, legislatures, elections, and other situations to identify universal
building blocks used in crafting all such structured situations? If
  
C HA PT ER ON E6
so, what are the underlying component parts that can be used to build
useful theories of human behavior in the diverse range of situations in
which humans interact? Can we use the same components to build an
explanation for behavior in a commodity market as we would use to
explain behavior in a university, a religious order, a transportation
system, or an urban public economy? Can we identify the multiple levels
of analysis needed to explain the regularities in human behavior that
we observe? Is there any way that the analyses of local problem solving,
such as the efforts of Maine lobster fishers for the last eighty years to
regulate their fisheries (see Acheson 1988, 2003; Wilson 1990), can be
analyzed using a similar set of tools as problem solving at a national
level (Gellar 2005; McGinnis forthcoming; Sawyer 2005) or at an
international level (Gibson, Anderson, et al. 2005; O. Young 1997,
2002)?
My answer to these questions is yes. This answer is, of course, a conjecture
and can be challenged. Asserting that there is an underlying unity is
easy. Convincing others of this is more difficult. I welcome exchanges with
others concerning the fundamental building blocks of organized human
interactions.
Many Components in Many Layers
The diversity of regularized social behavior that we observe at multiple
scales is constructed, I will argue, from universal components organized
in many layers. In other words, whenever interdependent individuals are
thought to be acting in an organized fashion, several layers of universal
components create the structure that affects their behavior and the outcomes
they achieve. I give a positive answer to these questions based on
years of work with colleagues developing and applying the Institutional
Analysis and Development (IAD) framework.2
Helping others to see the usefulness of developing a multilevel taxonomy
of the underlying components of the situations human actors face is
the challenge that I undertake in this volume. Scholars familiar with the
working parts used by mathematical game theorists to describe a game
will not be surprised by the positive answer. To analyze a game, the theorist
must answer a series of questions regarding universal components of
a game, including the number of players, what moves they can take, what
outcomes are available, the order of decisions, and how they value moves
and outcomes.
On the other hand, game theorists will be surprised at the extremely
large number of components identified in this book that create the context
within which a game is played. Further, if one drops the use of a universal,
simplified model of the individual, the number of options that a theorist
  
D IV ER SI TY AN D S TR UC TU RE D I NT ER AC TI ON S 7
must self-consciously make is even larger than experienced in the past.
While the usefulness of a universal model of rational behavior is challenged
in chapter 4, the assumption of a universal framework composed
of nested sets of components within components for explaining human
behavior is retained throughout the book.
Building a Framework
Game-theoretical analysis is drawn on and expanded in this book in several
ways. First, I do not confine analysis to those situations that are simple
enough to be analyzed as formal games. The core concept of an action
situation (discussed in chapters 2 and 3) can be formalized as a mathematical
game to represent many simple and important situations. Many other
significant situations—particularly where rules are the object of choice—
are too complex to be modeled as a simple game. (Agent-based models
and simulations of diverse types will provide the modeling tools we need
to capture patterns of interaction and outcomes in many of these more
complex settings [Janssen 2003].)
Second, I dig further to develop a consistent method for overtly analyzing
the deeper structures that constitute any particular action situation.
For some game theorists, this deeper structure is irrelevant once the structure
of a game itself is made explicit. Third, the narrow model of human
behavior used in game theory is viewed as one end of a continuum of
models of human behavior appropriate for institutional analysis. The
three basic assumptions of that model are used as a foundation for specifying
the type of assumptions that a theorist needs to make when animating
an institutional analysis.
The challenge for institutional theorists—as I discuss in chapter 4—is to
know enough about the structure of a situation to select the appropriate
assumptions about human behavior that fit the type of situation under
analysis. Thus, the approach presented here encompasses contemporary
game theory as one of the theories that is consistent with the IAD framework.
Also included, as discussed in chapter 4, are broader theories that
assume individuals are fallible learners trying to do the best they can in
the long term by using norms and heuristics in making their immediate
decisions.
As a scholar committed to understanding underlying universal components
of all social systems, I do not introduce complexity lightly. I view
scientific explanation as requiring just enough variables to enable one to
explain, understand, and predict outcomes in relevant settings. Thus, for
many questions of interest to social scientists, one does not need to dig
down through nested layers of rules that are examined in the last half of
this book. One can develop a good analysis of the situation (chapters 2
  
C HA PT ER ON E8
and 3), decide what assumptions to make about participants (chapter 4),
and predict outcomes. If the predictions are supported empirically, that
may be all that is needed.3
If the predictions are not supported, however, as is the case with much
contemporary work on social dilemmas and settings involving trust and
reciprocity, one has to dig under the surface to begin to understand why.
And if one wants to improve the outcomes achieved over time, one is
faced with the need to understand the deeper structure in the grammar of
institutions discussed in chapter 5 and the types of rules used to create
structure as discussed in chapters 7 and 8. This volume can be viewed as
presenting a series of nested conceptual maps of the explanatory space
that social scientists can use in trying to understand and explain the diversity
of human patterns of behavior. Learning to use a set of conceptual
maps and determining the right amount of detail to use is, however, itself
a skill that takes some time to acquire just as it does with geographic
maps (see Levi 1997b).
Frameworks and Conceptual Maps
For example, if I want to know the quickest route from Providence Bay
to Gore Bay on the Manitoulin Island, where Vincent Ostrom and I spend
summers writing at our cabin on the shores of Lake Huron, I need a very
detailed map of the interior of the island itself. If I want to explain where
the Manitoulin Island is to a colleague—who wants to know where we
spend our summers—I need a less detailed and larger map that shows its
location on the northern shores of Lake Huron, one of the Great Lakes of
the North American continent. If I try to use a map of the entire Western
Hemisphere, however, the Great Lakes are all so small that locating the
Manitoulin Island itself may be a challenge. I may only be able to point
to the Province of Ontario in Canada, where it is located, or to the entire
set of the Great Lakes. The advantage of a good set of geographic maps
is that after centuries of hard work, multiple levels of detailed maps of
most places are available and are nested in a consistent manner within
one another. Most of us recognize that there is not one optimal map
that can be used for all purposes. Each level of detail is useful for different
purposes.
The “map” that I will elucidate in this volume is a conceptual framework
called, as mentioned above, the Institutional Analysis and Development
(IAD) framework. The publication of “The Three Worlds of Action:
A Metatheoretical Synthesis of Institutional Approaches” (Kiser and Ostrom
1982) represented the initial published attempt to describe the IAD
framework. Our goal was to help integrate work undertaken by political
scientists, economists, anthropologists, lawyers, sociologists, psycholo-
  
D IV ER SI TY AN D S TR UC TU RE D I NT ER AC TI ON S 9
gists, and others interested in how institutions affect the incentives confronting
individuals and their resultant behavior.4
During the time since
this publication, the framework has been developed further5
and applied
to analyze a diversity of empirical settings. These include:
• the study of land boards in Botswana (Wynne 1989);
• the impact of institutions on creating effective monitoring and evaluations
in government development projects (Gordillo and Andersson 2004);
• the incentives of operators and state government regarding coal roads in
Kentucky (Oakerson 1981);
• the evolution of coffee cooperatives in Cameroon (Walker 1998);
• the causes and effects of property-right changes among the Maasai of Kenya
(Mwangi 2003);
• the performance of housing condominiums in Korea (J. Choe 1992);
• the regulation of the phone industry in the United States (Schaaf 1989);
• the effect of rules on the outcomes of common-pool resource settings
throughout the world (Oakerson 1992; Blomquist 1992; E. Ostrom 1990,
1992b; Agrawal 1999; Schlager 1994, 2004; Tang 1992; E. Ostrom, Gardner,
and Walker 1994; Lam 1998; de Castro 2000; Dolsˇak 2000; Futemma
2000; Yandle 2001; Gibson, McKean, and Ostrom 2000);
• a comparison of nonprofit, for profit, and government day-care centers (Bushouse
1999);
• the impact of decentralization on forest governance in Bolivia (Andersson
2002, 2004);
• the evolution of banking reform in the United States (Polski 2003); and
• the effect of incentives on donor and recipient behavior related to international
aid (Gibson, Anderson et al. 2005).
Our confidence in the usefulness of the IAD framework has grown
steadily in light of the wide diversity of empirical settings where it has
helped colleagues identify the key variables to undertake a systematic
analysis of the structure of the situations that individuals faced and
how rules, the nature of the events involved, and community affected
these situations over time. What is certainly true is that the number of
specific variables involved in each of these empirical studies is very
large. The specific values of variables involved in any one study (or one
location in a study) differ from the specific values of variables involved in
another study.
The problem of many variables, and potentially few instances of any
one combination of these variables, has been recognized by other scholars
as one of the perplexing problems haunting systematic empirical testing
of social science theories. James Coleman (1964, 516–19) referred to the
development and testing of “sometimes true theories,” by which he meant
that explanations were likely to hold under specific conditions and not
  
C HA PT ER ON E10
under others. If a small number of conditions were identified, sometimes
true theories would not present a major problem for the social sciences.
Rigorous analysis of many important questions, however, does eventually
require examining a large number of variables. Viewing macropolitical
orders in developed Western societies, for example, Fritz Scharpf
(1997, 22) points out that the national institutional settings “known to
affect policy processes can be described as being either unitary or federal,
parliamentary or presidential, have two- or multi-party systems in which
interactions are competitive or consociational, and with pluralist or neocorporatist
systems of interest intermediation.” Each one of the five variables
can exist in one or the other “setting” independently of the other
four variables. And, to make it worse, there may be variables related to
the particular policy area—such as banking, environmental policy, or education—that
may also change. “For comparative policy research, this
means that the potential number of different constellations of situational
and institutional factors will be extremely large—so large, in fact, that it
is rather unlikely that exactly the same factor combination will appear in
many empirical cases” (23). A similar level of complexity exists when
analyzing factors affecting the performance of city-county consolidation
efforts (Carr and Feiock 2004).
Hammond and Butler (2003) have illustrated this problem clearly in
their critique of the work of some institutional theorists who have made
overly strong claims for the overarching differences between parliamentary
and presidential systems. Presidential systems—according to Burns
(1963), Sundquist (1968), and Valenzuela (1993)—are thought to slow, if
not halt, policy change and lead to obstruction, frustration, and deadlock
interspersed with occasional bursts of change when a president faces both
houses of Congress dominated by his own party. Hammond and Butler
carefully analyze the interaction between rules and the preference profiles
that may exist in five variations of institutional rules. They conclude “that
considering institutional rules alone provides an inadequate guide to the
behavior of any system” (Hammond and Butler 2003, 183).
As Marwell and Oliver (1993, 25) put it, the “predictions that we can
validly generate must be complex, interactive and conditional.” And, we
can hope that some changes in a component are neutral—or have no impact
on outcomes—in at least some settings (as biologists are now learning
about in regard to genotypes; see Gavrilets 2003). While verifying
the empirical warrantability of precise predictions has been the guiding
standard for much of the work in political economy, we may have to
be satisfied with an understanding of the complexity of structures and a
capacity to expect a broad pattern of outcomes from a structure rather
than a precise point prediction. An outcome consistent with a pattern
  
D IV ER SI TY AN D S TR UC TU RE D I NT ER AC TI ON S 11
may be the best verification we can achieve in settings of substantial complexity
(Crutchfield and Schuster 2003).
Thus, the many relevant variables, the immense number of combinations
of these variables that exist, and their organization into multiple
levels of analysis make understanding organized social life a complex endeavor.
If every social science discipline or subdiscipline uses a different
language for key terms and focuses on different levels of explanation as
the “proper” way to understand behavior and outcomes, one can understand
why discourse may resemble a Tower of Babel rather than a cumulative
body of knowledge. This book is devoted to the task of building
on the efforts of many scholars to develop a conceptual approach that
hopefully has a higher chance of cumulation than many of the separate
paths currently in vogue in contemporary social sciences.
Holons: Nested Part-Whole Units of Analysis
Like good geographic maps, the IAD framework can be presented at
scales ranging from exceedingly fine-grained to extremely broad-grained.
Human decision making is the result of many layers of internal processing
starting with the biophysical structure, but with layers upon layers of
cognitive structure on top of the biophysical components (Hofstadter
1979). Further, many of the values pursued by individuals are intrinsic
values that may not be represented by external material objects, and their
presence and strength are important parts of the individual to be examined.
Building on top of the single individual are structures composed of
multiple individuals—families, firms, industries, nations, and many other
units—themselves composed of many parts and, in turn, parts of still
larger structures. What is a whole system at one level is a part of a system
at another level.
Arthur Koestler (1973) refers to such nested subassemblies of partwhole
units in complex adaptive systems as holons. “The term holon may
be applied to any stable sub-whole in an organismic or social hierarchy,
which displays rule-governed behaviour and/or structural Gestalt constancy”
(291). Christopher Alexander (1964) earlier conceptualized all
components of social arrangements as having a pattern and being a unit.
Units have subunits and are themselves parts of larger units that fit together
as a pattern. Koestler asserts that a “hierarchically organized whole
cannot be ‘reduced’ to its elementary parts; but it can be ‘dissected’ into
its constituent branches on which the holons represent the nodes of the
tree, and the lines connecting them the channels of communication, control
or transportation, as the case may be” (1973, 291). Thus, much of
the analysis presented in this book will be a form of “dissecting” complex
  
C HA PT ER ON E12
systems into composite holons that are then dissected further. Explanations
occur at multiple levels and different spatial and temporal scales.
Because explanations occur at multiple levels and different spatial and
temporal scales, the relevant theoretical concepts needed to understand
phenomena at one level do not necessarily scale up or down. One of the
core puzzles facing the field of landscape ecology, for example, is the problem
of identifying the scale at which a process or phenomenon occurs.
According to Pickett and Cadenasso (1995, 333), “The basic question
about scale in ecology consists of determining whether a given phenomenon
appears or applies across a broad range of scales, or whether it is
limited to a narrow range of scales” (see also S. Levin 1992).
The parts used to construct a holon are frequently not descriptive of
the holon they have created. A house is constructed out of floor joists,
roof beams, lumber, roofing material, nails, and so forth. When one wants
to talk about the house itself, one usually talks about the number of
rooms, the style of the house, the number of stories, rather than the number
of nails used in construction—even though a contractor and a hardware
salesperson may try to estimate exactly this variable at some point
during construction. When one wants to talk about the street on which
the house is located, one uses terms such as the size of the lots, the width
of the road, the complementarity or lack of complementarity of the building
style, and the like. Descriptions of a neighborhood will use still different
concepts, as will a description of an urban or rural political jurisdiction
in which a neighborhood is located. On the other hand, some
concepts can be used to dissect holons operating at different scales of
analysis.
Consequently, the institutional analyst faces a major challenge in identifying
the appropriate level of analysis relevant to addressing a particular
puzzle and learning an appropriate language for understanding at least
that focal level and one or two levels above and below that focal level. It
is not only social scientists who face this problem. At a meeting of the
global change scientists held in Bonn in March of 2001, Peter Lemke of
the World Climate Research Project indicated that the earlier emphasis in
climate research was all on global weather forecasts. This has proved to
be a myth and a delusion. Now they recognize that to do good weather
forecasting, one has to have detailed local models supplemented by global
weather models. Both local and global are needed. They are complementary
rather than competitive. Physical scientists are trying now in their
global models to integrate some of the more localized conditions, but that
turns out to be very difficult.
Ecologists have struggled with understanding ecological systems composed
of communities, modular units within communities, subunits
within these, and attributes of the species in a community (such as diver-
  
D IV ER SI TY AN D S TR UC TU RE D I NT ER AC TI ON S 13
Exogenous Variables
Action
Arena
Interactions
Outcomes
Evaluative
Criteria
Figure 1.1 The focal level of analysis—an action arena.
sity) or of individual species (such as trophic level) (see Tilman 1999;
Tilman, Lehman, and Bristow 1998). Extensive field research, analytical
modeling, and simulations now enable ecologists to make relatively
strong predictions about some of these interactions. “Increasing species
diversity is likely associated with more complex community structure, as
species with unique ecological roles are added. The introduction of new
ecological roles may be stabilizing or destabilizing, depending on how
species function within the community. For example, the addition of a
third trophic level to an otherwise stable community with only prey and
predators could potentially destabilize the system” (Ives, Klug, and Goss
2000, 409). Social scientists are slowly gaining greater capabilities for
understanding multilevel complex systems, but until we develop the appropriate
theoretical language for analyzing these systems, we will continue
to condemn all complex communities of interacting human organizations
as chaotic, as was the dominant view of urban scholars during the
last half-century (see, for example, Hawley and Zimmer 1970).
Action Arenas as Focal Units of Analysis
The focal level for this book is the holon called an action arena in which
two holons—participants and an action situation—interact as they are
affected by exogenous variables (at least at the time of analysis at this
level) and produce outcomes that in turn affect the participants and the
action situation. Action arenas exist in the home; in the neighborhood;
in local, regional, national, and international councils; in firms and markets;
and in the interactions among all of these arenas with others. The
simplest and most aggregated way of representing any of these arenas
when they are the focal level of analysis is shown in figure 1.1, where
exogenous variables affect the structure of an action arena, generating
interactions that produce outcomes. Evaluative criteria are used to judge
the performance of the system by examining the patterns of interactions
and outcomes.
Outcomes feed back onto the participants and the situation and may
transform both over time. Over time, outcomes may also slowly affect
  
C HA PT ER ON E14
some of the exogenous variables. In undertaking an analysis, however,
one treats the exogenous variables as fixed—at least for the purpose of
the analysis. When the interactions yielding outcomes are productive for
those involved, the participants may increase their commitment to maintaining
the structure of the situation as it is, so as to continue to receive
positive outcomes. When participants view interactions as unfair or otherwise
inappropriate, they may change their strategies even when they are
receiving positive outcomes from the situation (Fehr and Ga¨chter 2000b).
When outcomes are perceived by those involved (or others) as less valued
than other outcomes that might be obtained, some will raise questions
about trying to change the structure of the situations by moving to a
different level and changing the exogenous variables themselves. Or, if the
procedures were viewed as unfair, motivation to change the structure may
exist (Frey, Benz, and Stutzer 2004).
Similar efforts to identify a core unit of analysis, such as the action
arena, that is contained in many diverse environments have a long history.
Core units of analyses identified by other scholars include:
• collective structures (Allport 1962);
• events (Appleyard 1987; Heise 1979);
• frames (Goffman 1974);
• social action and interaction settings (Burns and Flam 1987);
• logic of the situation (Farr 1985; Popper 1961, 1976);
• problematic social situations (Raub and Voss 1986);
• scripts (Schank and Abelson 1977);
• transactions (Commons [1924] 1968); and
• units of meaning (Barwise and Perry 1983; Raiffa 1982).
Because the IAD framework is a multitier conceptual map, the simplest
schematic representation of an action arena shown in figure 1.1 will be
unpacked—and then further unpacked and unpacked throughout the initial
chapters of this book. Action arenas include two holons: an action
situation and the participant in that situation (see figure 1.2). An action
situation can, in turn, be characterized using seven clusters of variables:
(1) participants (who may be either single individuals or corporate
actors), (2) positions, (3) potential outcomes, (4) action-outcome linkages,
(5) the control that participants exercise, (6) types of information
generated, and (7) the costs and benefits assigned to actions and outcomes
(see figure 2.1 in the next chapter). Thus, an action situation refers to the
social space where participants with diverse preferences interact, exchange
goods and services, solve problems, dominate one another, or fight
(among the many things that individuals do in action arenas). In chapter
2, we will zoom in and unpack the action situation as a focal unit of
analysis. We will illustrate the working parts of an action situation in
  
D IV ER SI TY AN D S TR UC TU RE D I NT ER AC TI ON S 15
Exogenous Variables
Action Arena
Interactions
Outcomes
Evaluative
Criteria
Biophysical/
Material Conditions
Attributes of
Community
Rules
Action
Situations
Participants
Figure 1.2 A framework for institutional analysis. Source: Adapted from E. Ostrom,
Gardner, and Walker 1994, 37.
chapter 3, showing how this holon can be operationalized in an experimental
laboratory. In chapter 4, we will zoom in to unpack the concept
of a participant and discuss the puzzles and possibilities available to animate
the actor. But first, let’s use zoom out to examine the variables that
are treated as exogenous when examining an action arena (but may themselves
be an outcome of another action arena). Let’s look at a broader
overview of the IAD conceptual map.
Zooming Out to an Overview of the IAD Framework
An institutional analyst can take two additional steps after an effort is
made to understand the initial structure of an action arena leading to a
particular pattern of interactions and outcomes. One step moves outward
and inquires into the exogenous factors that affect the structure of an
action arena. From this vantage point, any particular action arena is now
viewed as a set of dependent variables. The factors affecting the structure
of an action arena include three clusters of variables: (1) the rules used
by participants to order their relationships, (2) the attributes of the biophysical
world that are acted upon in these arenas, and (3) the structure
of the more general community within which any particular arena is
placed (see Kiser and Ostrom 1982). The next section of this chapter
provides a brief introduction to this first step (see the left side of figure
1.2). How rules influence the action arena will then be discussed in much
more depth in chapters 5, 6, and 7.
The second step also moves outward—but to the “other side” of a particular
action arena—to look at how action arenas are linked together either
sequentially or simultaneously. This step will be discussed in the last section
of chapter 2 after discussion of the components of action situations.
  
C HA PT ER ON E16
Viewing Action Arenas as Dependent Variables
Underlying the way analysts conceptualize action situations and the participants
that interact in them are implicit assumptions about the rules participants
use to order their relationships, about attributes of the biophysical
world, and about the nature of the community within which the arena
occurs. Some analysts are not interested in the role of these underlying
variables and focus only on a particular arena whose structure is given.
On the other hand, institutional analysts may be more interested in one
factor affecting the structure of arenas than they are interested in others.
Anthropologists and sociologists tend to be more interested in how shared
or divisive value systems in a community affect the ways humans organize
their relationships with one another. Environmentalists tend to focus on
various ways that physical and biological systems interact and create opportunities
or constraints on the situations human beings face. Political
scientists tend to focus on how specific combinations of rules affect incentives.
Rules, the biophysical and material world, and the nature of the
community all jointly affect the types of actions that individuals can take,
the benefits and costs of these actions and potential outcomes, and the
likely outcomes achieved.
The Concept of Rules
The concept of rules is central to the analysis of institutions (Hodgson
2004a). The term rules, however, is used by scholars to refer to many
concepts with quite diverse meanings. In an important philosophical
treatment of rules, Max Black (1962) identified four different usages of
the term in everyday conversations. According to Black, the word rule is
used to denote regulations, instructions, precepts, and principles. When
used in its regulation sense, rules refer to something “laid down by an
authority (a legislature, judge, magistrate, board of directors, university
president, parent) as required of certain persons (or, alternatively, forbidden
or permitted)” (115). The example of a rule in the regulation sense
that Black uses is: “The dealer at bridge must bid first.” When using rule
in its regulation sense, one can meaningfully refer to activities such as the
rule “being announced, put into effect, enforced (energetically, strictly,
laxly, invariably, occasionally), disobeyed, broken, rescinded, changed,
revoked, reinstated” (109).
When the term rule is used to denote an instruction, it is closer in meaning
to an effective strategy for how to solve a problem. An example of
this usage is, “In solving quartic equations, first eliminate the cubic term”
(110). When speaking about a rule in this sense, one would not talk about
  
D IV ER SI TY AN D S TR UC TU RE D I NT ER AC TI ON S 17
a rule being enforced, rescinded, reinstated, or any of the other activities
relevant to regulation. When rule denotes a precept, the term is being used
as a maxim for prudential or moral behavior. An example would be: “A
good rule is: to put charity ahead of justice” (111). Again, one would not
speak of enforcing, rescinding, or reinstating a rule in the precept sense.
The fourth sense in which the term rule is used in everyday language is
to describe a law or principle. An example of this usage is: “Cyclones
rotate clockwise, anticyclones anticlockwise” (113). Principles or physical
laws are subject to empirical test, and as such truth values can be
ascribed to them. But physical laws are not put into effect, broken, or
rescinded.
Social scientists employ all four of the uses of the term rule that Black
identifies—and others as well (see discussion in chapter 5). Scholars engaged
in institutional analysis frequently use the term to denote a regulation.
The definition of rules used in this book is close to what Black identified
as the regulation sense. Rules can be thought of as the set of
instructions for creating an action situation in a particular environment.
In some ways, rules have an analogous role to that of genes. Genes combine
to build a phenotype. Rules combine to build the structure of an
action situation. The property rights that participants hold in diverse settings
are a result of the underlying set of rules-in-use (Libecap 1989).
Rules, in the instruction sense, can be thought of as the strategies
adopted by participants within ongoing situations. I will consistently use
the term strategy rather than rule for individual plans of action. Rules
in the precept sense are part of the generally accepted moral fabric of a
community (Allen 2005). We refer to these cultural prescriptions as
norms. Rules in the principle sense are physical laws.
Until recently, rules have not been a central focus of most of the social
sciences. Even in game theory where “the rules of the game” seem to play
an important role, there has not been much interest in examining where
rules come from or how they change. Game-theoretical rules include all
physical laws that constrain a situation as well as rules devised by humans
to structure a situation. The rules of the game—including both physical
and institutional factors—structure the game itself, but have been irrelevant
to many game theorists once a game can be unambiguously represented.
An influential contributor to the development of game theory, Anatol
Rapoport (1966, 18) stated this distinction very clearly: “Rules are
important only to the extent that they allow the outcomes resulting from
the choices of participants to be unambiguously specified. . . . Any other
game with possibly quite different rules but leading to the same relations
among the choices and the outcomes is considered equivalent to the game
in question. In short, game theory is concerned with rules only to the
extent that the rules help define the choice situation and the outcomes
  
C HA PT ER ON E18
associated with the choices. Otherwise the rules of games play no part in
game theory.” So long as the game theorist has adequately represented
this focal level of analysis, the theorist interested primarily in finding the
solution to a game has no need to dig into the rules, attributes of a community,
and physical laws that create the structure of the situation. As institutional
analysts asked to diagnose why perverse outcomes occur and to
propose ways to improve the outcomes of many action situations, on the
other hand, we have to dig below and learn how rules create the set being
analyzed. One cannot improve outcomes without knowing how the structure
is itself produced (Eggertsson 2005).
As will be discussed in more depth in chapter 5, rules as used in this
book are defined to be shared understandings by participants about enforced
prescriptions concerning what actions (or outcomes) are required,
prohibited, or permitted (Ganz 1971; V. Ostrom 1980; Commons 1968).
All rules are the result of implicit or explicit efforts to achieve order and
predictability among humans by creating classes of persons (positions)
who are then required, permitted, or forbidden to take classes of actions
in relation to required, permitted, or forbidden outcomes or face the likelihood
of being monitored and sanctioned in a predictable fashion (V. Ostrom
1991).
Well-understood and enforced rules operate so as to rule out some actions
and to rule in others. In a well-ordered human enterprise, some
behaviors are rarely observed because individuals following rules do not
normally engage in that activity in the given setting. It is rare to observe,
for example, that one driver on a public freeway within the United States
will race another driver on that freeway at a speed exceeding one hundred
miles per hour. State highway patrols invest substantial sums in an attempt
to enforce highway speeding laws and to rule out excessive speeds
on freeways.
At a racing track, however, one can observe speeds of well over one
hundred miles per hour and drivers directly racing one another in a determined
manner. The rules of a racing track rule in some activities that are
ruled out on a freeway. Anyone driving on a freeway will observe a range
of speeds rather than the single maximum speed mentioned in the speed
limit law. Speed limits illustrate rules that authorize a range of activities
rather than requiring one particular action. Further, enforcement patterns
differ regarding the range of speed in excess of the official upper limit that
will be tolerated, once observed, before a sanction is issued.
It is also important to recognize that rules need not be written. Nor do
they need to result from formal legal procedures. Institutional rules are
often self-consciously crafted by individuals to change the structure of
repetitive situations that they themselves face in an attempt to improve
the outcomes that they achieve.
  
D IV ER SI TY AN D S TR UC TU RE D I NT ER AC TI ON S 19
ON THE ORIGIN OF RULES
When one is interested in understanding the processes of governance, one
needs to ask where the rules that individuals use in action situations originate.
In an open and democratic governance system, many sources exist
of the rules used by individuals in everyday life. It is not considered illegal
or improper for individuals to self-organize and craft their own rules if
the activities they engage in are legal. In addition to the legislation and
regulations of a formal central government, there are apt to be laws
passed by regional and local governments. Within private firms and voluntary
associations, individuals are authorized to adopt many different
rules determining who is a member of the firm or association, how profits
(benefits) are to be shared, and how decisions will be made. Each family
constitutes its own rule-making body.
When individuals genuinely participate in the crafting of multiple layers
of rules, some of that crafting will occur using pen and paper. Much of
it, however, will occur as problem-solving individuals interact trying to
figure out how to do a better job in the future than they have done in the
past. Colleagues in a work team are crafting their own rules when they
might say to one another, “How about if you do A in the future, and I
will do B, and before we ever make a decision about C again, we both
discuss it and make a joint decision?” In a democratic society, problemsolving
individuals do this all the time. They also participate in more
structured decision-making arrangements, including elections to select
legislators.
WORKING RULES
Thus, when we do a deeper institutional analysis, we attempt first to understand
the working rules that individuals use in making decisions.
Working rules are the set of rules to which participants would make reference
if asked to explain and justify their actions to fellow participants.
While following a rule may become a “social habit,” it is possible to make
participants consciously aware of the rules they use to order their relationships.
Individuals can consciously decide to adopt a different rule and
change their behavior to conform to such a decision. Over time, behavior
in conformance with a new rule may itself become habitual (see Shimanoff
1980; Toulmin 1974; Harre´ 1974). The capacity of humans to use complex
cognitive systems to order their own behavior at a relatively subconscious
level makes it difficult at times for empirical researchers to ascertain
what the working rules for an ongoing action arena may actually be in
practice. It is the task of an institutional analyst, however, to dig under
surface behavior to obtain a good understanding of what rules participants
in a situation are following.6
  
C HA PT ER ON E20
Once we understand the working rules, then, we attempt to understand
where those rules come from. In an open society governed by a “rule of
law,” the general legal framework in use will have its source in actions
taken in constitutional, legislative, and administrative settings augmented
by rule-making decisions taken by individuals in many different particular
settings. In other words, the rules-in-form are consistent with the rulesin-use
(Sproule-Jones 1993). In a system that is not governed by a “rule
of law,” there may be central laws and considerable effort made to enforce
them, but individuals generally attempt to evade rather than obey the law.
THE PREDICTABILITY OF RULES
Rule following or conforming actions are not as predictable as biological
or physical behavior explained by scientific laws. All rules are formulated
in human language. As such, rules share problems of lack of clarity, misunderstanding,
and change that typify any language-based phenomenon
(V. Ostrom 1980, 1997). Words are always simpler than the phenomenon
to which they refer. In many office jobs, for example, the rules require an
employee to work a specified number of hours per week. If a staff member
is physically at their desk for the required number of hours, is daydreaming
about a future vacation or preparing a grocery list for a shopping
trip on the way home within the rules? Interpreting rules is more
challenging than writing them down.
The stability of rule-ordered actions is dependent upon the shared
meaning assigned to words used to formulate a set of rules. If no shared
meaning exists when a rule is formulated, confusion will exist about what
actions are required, permitted, or forbidden. Regularities in actions cannot
result if those who must repeatedly interpret the meaning of a rule
within action situations arrive at multiple interpretations. Because “rules
are not self-formulating, self-determining, or self-enforcing” (V. Ostrom
1980, 342), it is human agents who formulate them, apply them in particular
situations, and attempt to enforce performance consistent with them.
Even if shared meaning exists at the time of the acceptance of a rule,
transformations in technology, in shared norms, and in circumstances
more generally change the events to which rules apply. “Applying language
to changing configurations of development increases the ambiguities
and threatens the shared criteria of choice with an erosion of their
appropriate meaning” (342).
The stability of rule-ordered relationships is also dependent upon enforcement.
According to Commons ([1924] 1968, 138), rules “simply say
what individuals must, must not, may, can, and cannot do, if the authoritative
agency that decides disputes brings the collective power of the community
to bear upon the said individuals.” Breaking rules is an option that
  
D IV ER SI TY AN D S TR UC TU RE D I NT ER AC TI ON S 21
is always available to participants in an action situation (as contrasted to
players in a formal game-theoretic model), but associated with breaking
rules is a risk of being monitored and sanctioned. If the risk is low, the
predictability and stability of a situation are reduced. And instability can
grow over time. If one person can cheat without fear of being caught,
others can also cheat with impunity. If the risk of exposure and sanctioning
is high, participants can expect that others will make choices from
within the set of permitted and required actions. The acceptance of rules
represents a type of Faustian bargain (V. Ostrom 1996). Someone is given
authority to use coercion to increase benefits for others—hopefully, for
most others.
The simplifying assumption is frequently made in analytical theories
that individuals in an action situation will take only those actions that
are lawful given the rules that apply. For many purposes, this simplifying
assumption helps the analyst proceed to examine important theoretical
questions not related to how well the rules are enforced. Highly complicated
games, such as football, can indeed be explained with more ease
because of the presence of active and aggressive on-site referees who constantly
monitor the behavior of the players and assign penalties for infraction
of rules.7
And these monitors face real incentives for monitoring consistently
and for applying fair and accepted penalties. Both the fans and
the managers of the relevant sports teams pay a lot of attention to what
the monitors are doing and the fairness of their judgments. In settings
where a heavy investment is not made in monitoring the ongoing actions
of participants, however, considerable difference between predicted and
actual behavior can occur as a result of the lack of congruence between
a model of lawful behavior and the illegal actions that individuals frequently
take in such situations.
This is not to imply that the only reason individuals follow rules is
because they are enforced. If individuals voluntarily participate in a situation,
they must share some general sense that most of the rules governing
the situation are appropriate. Otherwise, the cost of enforcement within
voluntary activities becomes high enough that it is difficult, if not impossible,
to maintain predictability in an ongoing voluntary activity. (One can
expect that it is always difficult to maintain predictability in an ongoing
activity where participants do not have the freedom to enter and leave the
situation.)
WHAT ARE THE IMPORTANT RULES?
What rules are important for institutional analysis? For some institutional
scholars, the important difference among rules has to do with the system
of property rights in use. At a very general level, it is sometimes useful to
  
C HA PT ER ON E22
know whether the rules related to a situation can be broadly classed as
government property, private property, community property, or no property
which is an open-access setting (Bromley et al. 1992). Scholars in the
legal pluralist tradition have strongly criticized these categories as not
being precise enough to understand the incentives facing participants and
thus are inadequate as a foundation for public policy (Benda-Beckmann
2001). They argue that an analyst needs to learn more about particular
property rights that specify particular bundles of rights (such as the right
to enter a state park versus the right to hunt deer in the same park) in
much more detail than those broad categories of rights (Benda-Beckmann
1995, 1997).
A myriad of specific rules are used in structuring complex action arenas.
Scholars have been trapped into endless cataloging of rules not related to
a method of classification most useful for theoretical explanations. But
classification is a necessary step in developing a science. Anyone attempting
to define a useful typology of rules must be concerned that the
classification is more than a method for imposing superficial order onto
an extremely large set of seemingly disparate rules. The way we have
tackled this problem using the IAD framework is to classify rules according
to their direct impact on the working parts of an action situation
(as will be discussed in chapters 6 and 7).
Biophysical and Material Conditions
While a rule configuration affects all of the elements of an action situation,
some of the variables of an action situation (and thus the overall set
of incentives facing individuals in a situation) are also affected by attributes
of the biophysical and material world being acted upon or transformed.
What actions are physically possible, what outcomes can be produced,
how actions are linked to outcomes, and what is contained in the
actors’ information sets are affected by the world being acted upon in a
situation. The same set of rules may yield entirely different types of action
situations depending upon the types of events in the world being acted
upon by participants. These “events” are frequently referred to by political
economists as the “goods and services” being produced, consumed,
and allocated in a situation as well as the technology available for these
processes.
The attributes of the biophysical and material conditions and their
transformation are explicitly examined when the analyst self-consciously
asks a series of questions about how the world being acted upon in a
situation affects the outcome, action sets, action-outcome linkages, and
information sets in that situation. The relative importance of the rule configuration
and biophysical conditions structuring an action situation var-
  
D IV ER SI TY AN D S TR UC TU RE D I NT ER AC TI ON S 23
ies dramatically across different types of settings. The rule configuration
almost totally constitutes some games, like chess, where physical attributes
are relatively unimportant. The relative importance of working
rules to biophysical attributes also varies dramatically within action situations
considered to be part of the public sector. Rules define and constrain
voting behavior inside a legislature more than attributes of the biophysical
world. Voting can be accomplished by raising hands, by paper ballots, by
calling for the ayes and nays, by marching before an official counter, or
by installing computer terminals for each legislator on which votes are
registered. In regard to communication within a legislature, however, attributes
of the biophysical world strongly affect the available options. The
principle that only one person can be heard and understood at a time in
any one forum strongly affects the capacity of legislators to communicate
effectively with one another (see V. Ostrom 1987).
Considerable academic literature has focused on the effect of attributes
of goods on the results obtained within action situations. A key assumption
made in the analysis of a competitive market is that the outcomes of
an exchange are highly excludable, easily divisible and transferable, and
internalized by those who participate in the exchange. Markets are predicted
to fail as effective decision mechanisms when they are the only
arena available for producing, consuming, or allocating a wide variety
of goods that do not meet the criteria of excludability, divisibility, and
transferability. Market failure means that the incentives facing individuals
in a situation, where the rules are those of a competitive market but the
goods do not have the characteristics of “private goods,” are insufficient
to motivate individuals to produce, allocate, and consume these goods at
an optimal level.
Let us briefly consider two attributes that are frequently used to distinguish
among four basic goods and services: exclusion and subtractability
of use. Exclusion relates to the difficulty of restricting those who benefit
from the provision of a good or a service. Subtractability refers to the
extent to which one individual’s use subtracts from the availability of a
good or service for consumption by others. Both of these two attributes
can range from low to high. When these attributes are dichotomized and
arrayed as shown in figure 1.3, they can be used as the defining attributes
of four basic types of goods: toll goods (sometimes referred to as club
goods), private goods, public goods, and common-pool resources. Goods
that are generally considered to be “public goods” yield nonsubtractive
benefits that can be enjoyed jointly by many people who are hard to exclude
from obtaining these benefits. Peace is a public good, as my enjoyment
of peace does not subtract from the enjoyment of others. Commonpool
resources yield benefits where beneficiaries are hard to exclude but
each person’s use of a resource system subtracts units of that resource
  
C HA PT ER ON E24
Subtractability of use
Difficulty of
excluding
potential
beneficiaries
Private goods
Common-pool resources
Toll goods
Public goods
Low
High
Low High
Figure 1.3 Four basic types of goods. Source: Adapted from V. Ostrom and
E. Ostrom 1977, 12.
from a finite total amount available for harvesting (E. Ostrom, Gardner,
and Walker 1994; Aggarwal and Dupont 1999). When a fisher harvests
a ton of fish, those fish are not available to any other fisher.
EXCLUDABILITY AND THE FREE-RIDER PROBLEM
When the benefits of a good are available to a group, whether or not
members of the group contribute to the provision of the good, that good
is characterized by problems of excludability. Where exclusion is costly,
those wishing to provide a good or service face a potential free-rider or
collective-action problem (Olson 1965). Individuals who gain from the
maintenance of an irrigation system, for example, may not wish to contribute
labor or taxes to maintenance activities, hoping that others will
bear the burden. This is not to say that all individuals will free-ride whenever
they can. A strong incentive exists to be a free-rider in all situations
where potential beneficiaries cannot easily be excluded for failing to contribute
to the provision of a good or service.
When it is costly to exclude individuals from enjoying benefits from an
investment, private, profit-seeking entrepreneurs, who must recoup their
investments through quid pro quo exchanges, have few incentives to provide
such services on their own initiative. Excludability problems can thus
lead to the problem of free-riding, which in turn leads to underinvestment
in capital and its maintenance.
Public sector provision of common-pool resources or infrastructure
facilities raises additional problems in determining preferences and organizing
finances. When exclusion is low cost to the supplier, preferences
are revealed as a result of many quid pro quo transactions. Producers
learn about preferences through the consumers’ willingness to pay for
various goods offered for sale. Where exclusion is difficult, designing
mechanisms that honestly reflect beneficiaries’ preferences and their will-
  
D IV ER SI TY AN D S TR UC TU RE D I NT ER AC TI ON S 25
ingness to pay is challenging, regardless of whether the providing unit is
organized in the public or the private sphere. In very small groups, those
affected are usually able to discuss their preferences and constraints on a
face-to-face basis and to reach a rough consensus. In larger groups, decisions
about infrastructure are apt to be made through mechanisms such
as voting or the delegation of authority to public officials. The extensive
literature on voting systems demonstrates how difficult it is to translate
individual preferences into collective choices that adequately reflect individual
views (Arrow 1951; Monroe forthcoming).
Another attribute of some goods with excludability problems is that
once they are provided, consumers may have no choice whatsoever as to
whether they will consume. An example is the public spraying of insects.
If an individual does not want this public service to be provided, there
are even stronger incentives not to comply with a general tax levy. Thus,
compliance with a broad financing instrument may, in turn, depend upon
the legitimacy of the public-choice mechanism used to make provision
decisions.
SUBTRACTABILITY
Goods and facilities can generate a flow of services that range from being
fully subtractable upon consumption by one user to another extreme
where consumption by one does not subtract from the flow of services
available to others. The withdrawal of a quantity of water from an irrigation
canal by one farmer means that there is that much less water for
anyone else to use. Most agricultural uses of water are fully subtractive,
whereas many other uses of water—such as for power generation or navigation—are
not. Most of the water that passes through a turbine to generate
power, for instance, can be used again downstream. When the use of
a flow of services by one individual subtracts from what is available to
others and when the flow is scarce relative to demand, users will be
tempted to try to obtain as much as they can of the flow for fear that it
will not be available later.
Effective rules are required if scarce, fully subtractive service flows are
to be allocated in a productive way (E. Ostrom, Gardner, and Walker
1994). Charging prices for subtractive services obviously constitutes one
such allocation mechanism. Sometimes, however, it is not feasible to price
services. In these instances, some individuals will be able to grab considerably
more of the subtractive services than others, thereby leading to noneconomic
uses of the flow and high levels of conflict among users.
Allocation rules also affect the incentives of users to maintain a system.
Farmers located at the tail end of an irrigation system that lacks effective
allocation rules have little motivation to contribute to the maintenance
  
C HA PT ER ON E26
of that system because they only occasionally receive their share of water.
Similarly, farmers located at the head end of such a system are not motivated
to provide maintenance services voluntarily because they will receive
disproportionate shares of the water whether or not the system is
well-maintained (E. Ostrom 1996).
ADDITIONAL ATTRIBUTES
In addition to exclusion and subtractability, the structure of action situations
is also affected by a diversity of other attributes that affect how rules
combine with physical and material conditions to generate positive or
negative incentives. The number of attributes that may affect the structure
of a situation is extraordinarily large, and I do not want even to start a
list in this volume. The crucial point for the institutional analyst is that
rules that help produce incentives leading to productive outcomes in one
setting may fail drastically when the biophysical world differs. As our
extensive studies of common-pool resources have shown, for example,
effective rules depend on the size of the resource; the mobility of its resource
units (e.g., water, wildlife, or trees); the presence of storage in the
system; the amount and distribution of rainfall, soils, slope, and elevation;
and many other factors (see E. Ostrom, Gardner, and Walker 1994).
These additional attributes are slowly being integrated into a body of
coherent theory about the impact of physical and material conditions on
the structure of the situations that individuals face and their resulting
incentives and behavior. Analysts diagnosing policy problems need to be
sensitive to the very large difference among settings and the need to tailor
rules to diverse combinations of attributes rather than some assumed uniformity
across all situations in a particular sector within a country.
Attributes of the Community
A third set of variables that affect the structure of an action arena relate
to the concept of the community within which any focal action arena is
located. The concept of community is again one that has many definitions
and meanings across and within the social sciences. Given the breadth of
what I already plan to tackle, I do not plan to focus in detail on how
various attributes of community affect the structure of situations within
a community (see Agrawal and Gibson 2001 for an excellent overall review
of the concept of community). The attributes of a community that
are important in affecting action arenas include: the values of behavior
generally accepted in the community; the level of common understanding
that potential participants share (or do not share) about the structure of
  
D IV ER SI TY AN D S TR UC TU RE D I NT ER AC TI ON S 27
particular types of action arenas; the extent of homogeneity in the preferences
of those living in a community; the size and composition of the
relevant community; and the extent of inequality of basic assets among
those affected.
The term culture is frequently applied to the values shared within a
community. Culture affects the mental models that participants in a situation
may share. Cultures evolve over time faster than our underlying
genetic endowment can evolve. Cultures have in turn affected how
the human brain itself has evolved (Boyd and Richerson 1985; Richerson
and Boyd 2002). The history of experience with governance institutions
at multiple levels affects the way local participants understand, implement,
modify, or ignore rules written by external officials (Medard and
Geheb 2001).
For example, when all participants share a common set of values and
interact with one another in a multiplex set of arrangements within a
small community, the probabilities of their developing adequate rules and
norms to govern repetitive relationships are much greater (Taylor 1987).
The importance of building a reputation for keeping one’s word is important
in such a community, and the cost of developing monitoring and
sanctioning mechanisms is relatively low. If the participants in a situation
come from many different cultures, speak different languages, and are
distrustful of one another, the costs of devising and sustaining effective
rules are substantially increased.
Whether individuals use a written vernacular language to express their
ideas, develop common understanding, share learning, and explain the
foundation of their social order is also a crucial variable of relevance to
institutional analysis (V. Ostrom 1997). Without a written vernacular language,
individuals face considerably more difficulties in accumulating
their own learning in a usable form to transmit from one generation to
the next.
Institutional Frameworks, Theories, and Models
So far in this chapter, I have provided a brief overview of the IAD framework
without telling the reader what I mean by a framework. The terms—
framework, theory, and model—are all used almost interchangeably by
diverse social scientists. This leads to considerable confusion as to what
they mean.8
Frequently, what one scholar calls a framework others call a
model or a theory.9
In this book, I will use these concepts to mean a nested
set of theoretical concepts—which range from the most general to the
most detailed types of assumptions made by the analyst. Analyses con-
  
C HA PT ER ON E28
ducted at each level provide different degrees of specificity related to a
particular problem (Schlager 1999).
The development and use of a general framework helps to identify the
elements (and the relationships among these elements) that one needs to
consider for institutional analysis. Frameworks organize diagnostic and
prescriptive inquiry. They provide the most general set of variables that
should be used to analyze all types of settings relevant for the framework.
Frameworks provide a metatheoretic language that is necessary to talk
about theories and that can be used to compare theories. They attempt
to identify the universal elements that any relevant theory would need to
include. Many differences in surface reality can result from the way these
variables combine with or interact with one another. Thus, the elements
contained in a framework help the analyst generate the questions that
need to be addressed when first conducting an analysis.
The development and use of theories enable the analyst to specify which
components of a framework are relevant for certain kinds of questions
and to make broad working assumptions about these elements. Thus,
theories focus on parts of a framework and make specific assumptions
that are necessary for an analyst to diagnose a phenomenon, explain its
processes, and predict outcomes. To conduct empirical research, a scholar
needs to select one or more theories to use in generating predictions about
expected patterns of relationships. Several theories are usually compatible
with any framework. Empirical research should narrow the range of applicable
theories over time by showing the superiority of the remaining
theories to explain data. Microeconomic theory, game theory, transaction
cost theory, social choice theory, public choice, constitutional and covenantal
theory, and theories of public goods and common-pool resources
are all compatible with the IAD framework discussed in this book.
The development and use of models make precise assumptions about a
limited set of parameters and variables. Logic, mathematics, game theory
models, experimentation and simulation, and other means are used to
explore the consequences of these assumptions systematically on a limited
set of outcomes. Multiple models are compatible with most theories. In
an effort to understand the strategic structure of the games that irrigators
play in differently organized irrigation systems, for example, Weissing
and Ostrom (1991a, 1991b) developed four families of models to explore
the likely consequences of different institutional and physical combinations
relevant to understanding how successful farmer organizations arranged
for monitoring and sanctioning activities. These models enabled
us to analyze in a precise manner a subpart of the theory of commonpool
resources and thus also one combination of the components of the
IAD framework. Models are extensively used in contemporary policy
  
D IV ER SI TY AN D S TR UC TU RE D I NT ER AC TI ON S 29
analysis by officials working with the World Bank, national governments,
as well as state and local governments.
For policy makers and scholars interested in issues related to how different
governance systems enable individuals to solve problems democratically
by modifying rules at various levels, the IAD framework helps to
organize diagnostic, analytical, and prescriptive capabilities. It is similar
in structure and intent to the “Actor-Centered Institutionalism” framework
developed by Renate Mayntz and Fritz Scharpf (1995) and applied
to several national policy settings by Fritz Scharpf (1997). It also aids in
the accumulation of knowledge from empirical studies and in the assessment
of past efforts at reforms.
Without the capacity to undertake systematic, comparative institutional
assessments, recommendations of reform may be based on naive
ideas about which kinds of institutions are “good” or “bad” and not
on an analysis of performance. Some policy analysts tend to recommend
private property as a way of solving any and all problems involving overuse
of a resource. While private property works effectively in some environments,
it is naive to presume it will work well in all (see Tietenberg
2002; Rose 2002). One needs a common framework and family of theories
in order to address questions of reforms and transitions. Particular
models then help the analyst to deduce specific predictions about likely
outcomes of highly simplified structures. Models are useful in policy analysis
when they are well-tailored to the particular problem at hand. Models
are used inappropriately when applied to the study of problematic situations
that do not closely fit the assumptions of the model (see E. Ostrom
1990 for a critique of the overreliance on open access models of commonpool
resources regardless of whether users had created their own rules to
cope with overharvesting or not).
The Limited Frame of This Book
Several times in the past, I have participated with colleagues in efforts to
outline a book that examined how rule configurations, attributes of
goods, and attributes of the community all affected the structure of action
situations, individual choices, outcomes, and the evaluation of out-
comes.10
Each time, the projected volume mushroomed in size and overcame
our capacity to organize it. Thus in this book, I have tried to focus
primarily on how rules affect the structure of action situations instead of
trying to work out the details of the entire framework. The focus on the
components of institutions in this volume should not be interpreted to
mean that I feel that institutions are the only factor affecting outcomes in
all action situations.
  
C HA PT ER ON E30
Institutions are only one of a large number of elements that affect behavior
in any particular situation at a particular time and place. No single
cause exists for human behavior. To live, one needs oxygen, water, and
nutrition. All are key parts of the explanation of life. Life itself operates
at multiple levels. Genes underlie phenotypic structures in a manner that
is broadly analogous to the way that rules underlie action situations. But
neither genes nor rules fully determine behavior of the phenotypes that
they help to create. Selection processes on genes operate largely at the
individual level, but rules—as well as other cultural “memes”—are likely
to be selected at multiple levels (see Hammerstein 2003). When one steps
back, however, for all of the complexity and multiple levels, there is a
large amount of similarity of underlying factors. In the biological world,
it is somewhat amazing that there is only a small proportion of the genes
that differ between an elephant and a mouse. As we develop the logic of
institutions further, we will see that many situations that have the surface
appearance of being vastly different have similar underlying parts. Thus,
our task is to identify the working parts, the grammar, the alphabet of
the phenotype of human social behavior as well as the underlying factors
of rules, biophysical laws, and community.
Thus, the focus of this book reflects my sense that the concept of institutions,
the diversity of institutions and their resilience, and the question of
how institutions structure action situations require major attention. This
volume is, thus, an effort to take an in-depth look at one major part of
what is needed to develop fuller theories of social organization. In this
volume, I will try to articulate in more detail than has been possible before
what I think the components of institutions are and how they can be used
to generate explanations of human behavior in diverse situations. The
focus on institutions should not be interpreted, however, as a position
that rules are always the most important factor affecting interactions and
outcomes. In the midst of a hurricane, rules may diminish greatly in their
importance in affecting individual behavior.
This volume should be thought of as part of a general effort to understand
institutions so as to provide a better formulation for improving their
performance. Our book Rules, Games, and Common-Pool Resources
(E. Ostrom, Gardner, and Walker 1994) provides one focused application
of the IAD framework. Michael McGinnis has recently edited three volumes
(1999a, 1999b, 2000) that present earlier elucidations of the IAD
framework and empirical studies that are closely related. Kenneth Bickers
and John Williams (2001) and Michael McGinnis and John Williams
(2001) clearly elucidate important aspects of the general approach.
I am writing this book from the perspective of a policy analyst. Without
the careful development of a rigorous and empirically verifiable set of
theories of social organization, we cannot do a very good job of fixing
  
D IV ER SI TY AN D S TR UC TU RE D I NT ER AC TI ON S 31
problems through institutional change. And, if we cannot link the theoretical
results into a coherent overall approach, we cannot cumulate knowledge.
All too often, major policy initiatives lead to counterintentional results.
We need to understand institutions in order to improve their
performance over time (North 2005).
As I demonstrate in chapter 8, however, the option of optimal design
is not available to mere mortals. The number of combinations of specific
rules that are used to create action situations is far larger than any set
that analysts could ever analyze even with space-age computer assistance.
This impossibility does not, however, leave me discouraged or hopeless.
It does, however, lead me to have great respect for robust institutions that
have generated substantial benefits over long periods of time (see Shepsle
1989; E. Ostrom 1990). None have been designed in one single step.
Rather, accrued learning and knowledge have led those with good information
about participants, strategies, ecological conditions, and changes
in technology and economic relationships over time to craft sustainable
institutions, even though no one will ever know if they are optimal. Thus,
in chapter 9, I dig into the process of learning, adaptation, and evolution
as processes that enable polycentric institutional arrangements to utilize
very general design principles in the dynamic processes of trying to improve
human welfare over time. It is also necessary to discuss the threats
that can destroy the resilience of complex social systems.
  
Two
Zooming In and Linking Action Situations
An Action Situation as a Focal Unit of Analysis
Whenever two or more individuals are faced with a set of potential actions
that jointly produce outcomes, these individuals can be said to be “in”
an action situation. Typical action situations include:
• buyers and sellers exchanging goods in a market;
• legislators making legislative decisions about future laws;
• powerful politicians bargaining over the allocation of public support;
• users of a common-pool resource withdrawing resource units (such as fish,
water, or timber);
• heads of state negotiating an international treaty.
The structure of all of these situations—and many more—can be described
and analyzed by using a common set of variables. These are: (1)
the set of participants, (2) the positions to be filled by participants, (3) the
potential outcomes, (4) the set of allowable actions and the function that
maps actions into realized outcomes, (5) the control that an individual
has in regard to this function, (6) the information available to participants
about actions and outcomes and their linkages, and (7) the costs and benefits—which
serve as incentives and deterrents—assigned to actions and
outcomes. The internal structure of an action situation can be represented
as shown in figure 2.1. In addition to the internal structure, whether a
situation will occur once, a known and finite number of times, or indefinitely
affects the strategies of individuals.
The number of participants and positions in a situation may vary, but
there must always be participants in positions for one to talk about an
action situation. Similarly, there must be potential actions that participants
can take. The set of available actions represents the means that
participants have to achieve particular outcomes in that situation. Information
about the situation may vary, but all participants must have access
to some common information about the situation for an analyst to even
state that the participants are in the same situation. The costs and benefits
assigned to actions and outcomes can be thought of as the external incentives
and deterrents in a situation. How these affect the choice of participants
regarding specific actions, and thus the cumulation of results, de-
  
L IN KI NG AC TI ON SI TU AT IO NS 33
Exogenous Variables
PARTICIPANTS
assigned to
POSITIONS
assigned to
ACTIONS
INFORMATION
about
CONTROL
over
Linked to
NET COSTS
AND BENEFITS
assigned to
POTENTIAL
OUTCOMES
Given implicit models used by actors
Observed Interactions and Outcomes
Evaluative Criteria
Empirical Analysis
Given the theory used by an analyst
Predicted Interactions and Outcomes
Evaluative Criteria
Theoretical Analysis
Figure 2.1 The internal structure of an action situation.
pends also on the initial resources and valuation patterns held by the
participants. In some action situations, the standing of participants is
grossly unequal, allowing some to have substantial power over others and
the relative benefits they can achieve.
When doing analysis at a focal level of analysis, these working parts
are the “givens” of a situation. For the purposes of analyzing the likely
human behavior and outcomes within a particular structure, one assumes
that the structure of the situation is fixed for the short run. Within a
particular situation, individuals can attempt to choose only in light of
their beliefs about the opportunities and constraints of that situation. In
an open society, individuals may be able eventually to affect the structure
of action situations in which they repeatedly find themselves by changing
the rule configurations affecting the structure of these situations. To do so,
they move to deeper analytical levels (collective-choice or constitutionalchoice
action situations) where the outcomes generated are changes in
  
C HA PT ER TW O34
the rules of other action situations. In a closed society, individuals at an
operational level may have little opportunity to change rules at any level
and may find themselves in highly exploitative situations. We discuss shifting
to a higher level in the last half of this chapter.
The working parts of an action situation are both necessary and sufficient
to describe the structure of an action situation. They are similar to
the elements identified by game theorists to construct formal game models
(see Gardner 2003; Gintis 2000b). A formal description of a game is thus
one way of describing a subset of all action situations and will soon be
used to illustrate the concept of a simple action situation.1
The mathematical tools of game theory are powerful and enlightening.
They can only be used, however, to elucidate the structure of relatively
simple action situations. The full game tree for a highly structured game,
such as chess, cannot be fully articulated. In chess, for example, the first
player can open with any of twenty actions and the second player can
respond with twenty as well. Thus, after the first two moves, there are
already four hundred branches to specify if one wanted to try to represent
chess as a formal game. “It has been estimated that the total number of
possible moves in chess is on the order of 10120
, or a ‘one’ with 120 zeros
after it. . . . A supercomputer a thousand times faster than your PC, making
a billion calculations a second, would need approximately 3 × 10103
years to check out all of these moves” (Dixit and Skeath 1999, 66).
In addition to using the elements of an action situation to analyze a
formal game, the elements can also be used to describe the structure of
more complex action situations in a nonmathematical form. One strategy
is to represent key parts of complex chains of action situations as a game
without trying to represent the entire structure as one game (McGinnis
forthcoming). The basic elements of an action situation have also been
used in empirical studies of complex action situations in the field.
Scholars associated with the Workshop in Political Theory and Policy
Analysis have developed three large databases that measured as many
aspects as we could regarding the structure of action situations facing
appropriators from common-pool resources: inshore fisheries (see
Schlager 1990, 1994, 2004; Schlager and Ostrom 1992, 1993); irrigation
systems (see Tang 1992; Lam 1998); and forests (see Gibson, McKean,
and Ostrom 2000; Moran and Ostrom 2005; Poteete and Ostrom 2004).
These efforts to measure the structure and the realized outcomes of various
property regimes as they relate to diverse common-pool resources led
us to realize both how difficult it is to measure these concepts in field
settings as well as the substantial scientific benefits achieved by so doing.
Specific sets of questions on our coding forms were our way of operationalizing
the basic working parts of operational-level action situations.
  
L IN KI NG AC TI ON SI TU AT IO NS 35
A carefully crafted case study is another method for analyzing more
complex action situations and their linkages (Yandle 2001). Analytical
narratives are an important technique for examining the structure of complex
action-outcome linkages and their consequences (Bates et al. 1998,
2000). So long as theorists use a consistent language to describe their
structure, much can be learned from single case studies over time or comparative
case studies of action situations that are not presented in formal
language (see Theesfeld 2004; Yandle and Dewees 2003).
Computer programs have been written as ways of representing the actions
of players. An IBM team, for example, put years of effort into a
chess-playing program—Deep Blue—to try to beat a world champion
chess player. The IBM team succeeded in beating Gary Kasparov, the
then world champion chess player in 1997, only after many years of try-
ing.2
Agent-based computational models are currently used by some analysts
to examine a variety of action situations that are too complex to
be analyzed completely using mathematical models (see Axelrod 1997;
Tesfatsion 2002; Janssen 2003; Janssen and Ostrom forthcoming a; Parker
et al. 2003; Hodgson and Knudsen 2004). The method of institutional
analysis described here can thus be implemented using a wide variety of
analytical tools.
Example of a Simple Action Situation
Before turning to a discussion of the individual working parts of a situation,
let us illustrate the concept of an action situation using the tools of
game theory to examine a simple game—the Snatch Game.3
The Snatch
Game characterizes situations where the individuals involved do not share
norms or rules. In other words, it is a game in a “state of nature.” In this
normless and ruleless environment, let us assume there are two farming
households who are fully self-reliant. Each produces a different commodity—say,
chickens and potatoes. Meals composed of all chicken or all
potatoes are not as flavorful or healthy as meals composed of both
chicken and potatoes. Both households would thus benefit from finding
a way of exchanging some of the chickens and potatoes they produce.
The structure of a Snatch Game in a state of nature is shown in figure
2.2. In this game, Household 1 and Household 2 both have ten comparable
units of the commodity they produce. Household 1 has ten chickens
and Household 2 has ten sacks of potatoes (each sack of potatoes is considered
comparable to one chicken). Both would be better off if they could
exchange half of their own commodities for half of those grown by the
other farming household. Say, each household valued their own commodity
as one. If they could exchange five units of their own production for
  
C HA PT ER TW O36
(10, 10)
(5, 20)(15, 15)
Snatch (10, 10)
Do Not Offer
Refuse Offer
Offer
Household 2
Exchange
Household 1
Figure 2.2 The Snatch Game. Source: Based on STEAL in Plott and Meyer
1975, 70.
five of the other farming household, the marginal value of the second
commodity would give them twice as much value—or a total of fifteen
value units each. The problem is how to accomplish an exchange when
there are no well-defined property rights.
Household 1 could either offer or not offer to exchange five units with
Household 2. If Household 2 gets an offer, then it has three choices.
The first is to refuse to exchange. The outcome here would be for both
households to retain their original ten units. Second, Household 2 can
agree to an exchange and both households would be better off with fifteen
value units. A third alternative—especially if Household 2 had several
young, strong sons—would be to snatch the five commodities offered
by Household 1 and keep their own. This leaves Household 1 with five
units. Household 2 then has fifteen commodity units valued at twenty
units of value.
Let us assume that both households value only the goods they finally
receive. If Household 2 has the physical capability of snatching the goods
once they were brought out in this “lawless” situation, no exchange
would take place. Given that Household 2 has not developed strong
norms against snatching, which affect the value of the outcomes and are
also known to Household 1, Household 1 would predict that Household
2 will snatch the goods. And without internal norms against using physical
force to take possession of goods, Household 2 would definitely snatch
Household 1’s goods, if given the opportunity. Given this certainty,
  
L IN KI NG AC TI ON SI TU AT IO NS 37
Household 1 would never offer an exchange in the first place. The predicted
equilibrium of this game is an inefficient outcome—no exchange.
Readers familiar with game theory will recognize that the Snatch Game
is one example of a very broad class of games that has the structure of a
one-shot, sequential social dilemma. Social dilemmas are ubiquitous in
economic, political, and social life. They occur whenever the private returns
to each participant are greater than their share of a joint return no
matter what other participants do. If the structure of a one-shot social
dilemma game is not changed, and individuals pursue their own immediate,
objective outcomes as the only values taken into account, individuals
will not achieve outcomes that could leave everyone better off. Asymmetric
social dilemmas, similar to the Snatch Game, are sometimes called: the
“Trust Game,” or the “Investment Game,” or even the “Peasant Dictator
Game” (Berg, Dickhaut, and McCabe 1995). We will describe findings
from a series of experiments on these games in chapter 3.
The pervasiveness of social dilemmas has repeatedly been recognized
in the great books of political philosophy. Hobbes described such a setting
as a “war of all against all.” Rousseau used a stag hunt to illustrate the
problem of a group needing to all work together to hunt a large animal
but facing the temptation to break up into separate groups when small
animals appeared on the scene that were easy to catch. A small group
could catch a rabbit, but ruined the chance for the group to obtain a
large animal. Many important books of the last several decades have been
devoted to an analysis of simultaneous or sequential social dilemmas
(Barry and Hardin 1982; Taylor 1987; Schelling 1978). We shall often
use social dilemma situations as examples throughout this book, since
understanding how individuals act within social dilemma situations constitutes
one of the major puzzles facing all contemporary social science
disciplines.
We shall return to a discussion of the Snatch Game at different junctures
throughout the book—not because we think that all of the interesting
action situations are illustrated by simple, two-player games. Rather,
simple situations can help us understand the concept of an action situation
itself. And these simple situations also illustrate that a large body of
important situations—social dilemmas—are at the heart of our discourse
about institutions and the diversity of institutions.
The Basic Working Parts of Action Situations
Let us now turn to the elements of an action situation so we can begin to
understand these fundamental elements of all interactive situations.
  
C HA PT ER TW O38
Participants
Participants in an action situation are decision-making entities assigned
to a position and capable of selecting actions from a set of alternatives
made available at nodes in a decision process. The participants in action
situations can also be corporate actors—nations, states in a federal system,
private corporations, NGOs, and so forth. Whenever participants
are organizations, one treats them “in” the situation as if they were a
single individual but one that is linked to a series of additional situations
within their own organization. For some purposes, one may ignore the
linked situations—especially when the interests of the organization, and
thus the strategy it will follow, are very clear and unlikely to change due
to an internal challenge. Alternatively, one may self-consciously examine
the linked structure (see McGinnis and Williams 1989). Several attributes
of participants are relevant when representing and analyzing specific situations.
These include (1) the number of participants, (2) their status as
individuals or as a team or composite actor, (3) and various individual
attributes, such as age, education, gender, and experience.
THE NUMBER OF PARTICIPANTS
Interdependent action situations require at least two participants—such
as the two households in the Snatch Game—where the actions of each
affect the outcomes for both. Adding a third participant changes the structure
of the situation substantially.4
The dividing line between major types
of games, for example, is between two-person and N-person games,
where N is defined as any number greater than two. The specific number
of participants is frequently overtly specified in real-world settings (or in
formal theories about these settings) such as legislatures, juries, and most
sports. Some descriptions of a situation, however, specify the number of
participants in a looser fashion such as a small or large group, or faceto-face
relationships versus impersonal relationships. Since many other
components of an action situation are affected by the number of participants,
this is a particularly important variable in the analysis of any action
situation.
THE INDIVIDUAL OR TEAM STATUS OF PARTICIPANTS
Participants in many action situations are individual persons, or they may
represent a team or composite actor, such as the households in the Snatch
Game. Under specific conditions, a group of individuals may be considered
as one participant (a team) in a particular action situation. Let us
address the conditions that are necessary to treat a group as a participant
when they participate individually in at least one other linked situation.
  
L IN KI NG AC TI ON SI TU AT IO NS 39
Drawing on the work of Fritz Scharpf (1997), we need first to distinguish
between sets of individuals who share many similar characteristics
and whose aggregate behavior may be predictable from knowledge acquired
about a sample of individuals. In electoral politics, one can discuss
“the urban voter,” or “veterans,” or “the labor vote.” These are shorthand
terms for a class of individuals who share characteristics that are
perceived to be very important in affecting their individual preferences
related to some events. These are meaningful concepts. They help the analyst
make sense out of many events when a large number of individuals
act independently, and it is useful to gain a realistic expectation about
their likely actions. They are not, however, acting as a team.
As Scharpf, and Granovetter (1978) before him acknowledge, when
individuals who share some key attributes also differ in the strength of
some other key variable, aggregate behavior may be characterized by frequency-dependent
behavior with strong threshold effects. This has been
used in the explanations of street protests or revolutionary actions (Kuran
1989). For example, those who feel the most strongly against a policy
may be willing to stage protest marches even if others do not join them.
If the number of individuals who have strong views is not sufficiently
large, however, other individuals may not be willing to engage in protest
actions. On the other hand, depending on the distribution of preferences,
it is possible for a strong initial showing to trigger more individuals,
whose actions in turn trigger still others, to participate. Such processes
can still be explained primarily at the individual level.
For an institutional analyst to consider a set of individuals to be a “composite”
actor, one must assume that the individuals intend to participate
in a collective action. One needs to assume that “the individuals involved
intend to create a joint product or to achieve a common purpose”
(Scharpf 1997, 54). Such composite actors—such as a household—will,
of course, vary in regard to the type of internal decision-making mechanisms
that they will use. Some will depend on very widespread preferences
of their members—which Scharpf calls “collective actors.” “Corporate
actors,” on the other hand, are not so dependent on the preferences of
their members and beneficiaries, and activities “are carried out by staff
members whose own private preferences are supposed to be neutralized
by employment contracts” (54).
A fully organized market with well-defined property rights, for example,
may include buyers and sellers who are organized as firms as well as
individual participants. Firms are composed of many individuals. Each
firm in a market is treated as if it were a single participant, but this is a
“shorthand” way of viewing the lattice of internal action situations within
a particular firm that leads to external decisions to be taken in market
settings.
  
C HA PT ER TW O40
Whether an analyst treats separate individuals as participants or as
members of a collective or corporate organization depends upon the concerns
and questions of the analyst. The game of bridge, for example, may
be represented as having either four participants or two teams, each composed
of two individuals. If the analyst is interested in examining communication
behavior during bidding, bridge is best represented as involving
four participants. For the purpose of explaining how moves lead to a final
score, the game may best be represented as involving two teams.
ATTRIBUTES OF PARTICIPANTS
Participants bring a diversity of ascribed or acquired characteristics to any
situation. These characteristics may not influence their actions in some
situations, while having a major impact on others. Whether drivers passing
one another on a busy highway are both of the same social or ethnic
background is unlikely to affect their actions. A young, inexperienced
driver of any background, however, may approach passing another car
with great hesitation and not pass when most adult drivers would do so.
The same two adults might hesitate a long time before extending trust
and reciprocity to each other if facing an opportunity to enter a long-term
contract in a community that has recently faced racial or ethnic conflict.
Even the simple Snatch Game is likely to have different outcomes depending
on the attributes of Households 1 and 2. The outcomes of many
situations depend on the knowledge and skills of the parties. Two chess
masters will play a chess tournament differently than two young children
just learning the game.
Attributes of participants are also affected by the rules structuring an
action situation. The A in the ADICO syntax presented in chapter 5 identifies
the attributes of participants as an important element of all rules.
Individuals with some ascribed or acquired attributes, such as ethnic
background, gender, or education may be enjoined from participation.
Positions
Another element of an action situation is the set of positions or “anonymous
slots” into and out of which participants move. Examples of positions
include players, voters, judges, buyers, sellers, legislators, police officers,
and so forth. Positions and participants are separate elements in a
situation even though they may not be clearly so identified in practice. In
many formal games, the distinction between a participant and the position
that a participant holds is blurred. In the Snatch Game, for example,
there are two participants and two positions. The positions are simply
labeled as Household 1 and Household 2. The household in the first posi-
  
L IN KI NG AC TI ON SI TU AT IO NS 41
tion has the opportunity to make the first move. While in this instance,
given the prediction of no exchange, holding the first position does not
give a special advantage. Frequently, however, first mover advantages
exist and may give one participant substantial power over the other.
The number of positions is frequently fewer than the number of participants.
In legislative committee situations, for example, there are frequently
only two positions—chair and member—while there may be five
to twenty-five or more participants. When game theorists identify the elements
of a game, they rarely mention positions in their definition (see
Rapoport 1966; Gintis 2000b). They then tend to name positions in the
text describing a game, as I did above, using terms such as Player 1 and
Player 2, Principal and Agent, or Agenda Setter and Member.
Depending on the structure of the situation, a participant may simultaneously
occupy more than one position. All participants will occupy
whatever is the most inclusive position in a situation—member, citizen,
employee, and the like. In a private firm, additional positions such as
foreman, division manager, or president will be occupied by some participants
while they continue to occupy the most inclusive position—that
of employee.
Positions are thus the connecting link between participants and actions.
In some situations, any participant in any position may be authorized
to take any of the allowable actions in that situation. However, in most
“organized” situations, the capability to take particular actions is assigned
to a specific position and not to all positions. The nature of a position
assigned to participants in an action situation defines the “standing”
of the participant in that situation. The standing of a position is the set
of authorized actions and limits on actions that the holder of the position
can take at particular choice sets in the situation. Those who hold the
position of a member of a legislative committee are authorized to debate
issues and vote on them. The member who holds the position of chair can
usually develop the agenda for the order of how issues will be brought
before the committee or even whether a proposal will even be discussed.
Determining this agenda frequently determines which issue will win in a
final vote (Plott and Levine 1978; Weingast 1989).
Action situations vary substantially in the degree to which participants
control their own entry into or exit from a position. A defendant in a
criminal trial does not control his or her movement into or out of this
position. Participants in many formal social dilemma games are treated
as if they have no choice regarding entry and exits. When exit is feasible,
however, the outcomes of social dilemma situations are likely to be different
(Orbell and Dawes 1993). Orbell, Schwarz-Shea, and Simmons
(1984) have shown that cooperators, when given the opportunity to exit
after playing with a defector, choose the exit option more than defectors.
  
C HA PT ER TW O42
A participant in a legislature could not be a member of the legislature
against his or her will, but could lose this position involuntarily. In some
hierarchical situations covered by civil service systems, individuals have
to compete vigorously for positions by passing examinations, but once
appointed, they may hold their positions for life, subject to their taking
legal actions.
Potential Outcomes
In the analysis of formal games, the standard practice is to report the
outcomes of joint decisions as the analyst assumes they are ranked in
utility to the participants in particular positions. In the Snatch Game, the
value of ten is assigned to each of the households for keeping their own
chickens and potatoes. A utility value of fifteen is designated for a successful
exchange of half of the chickens for half of the potatoes, and the utility
value of twenty to Household 2 for keeping its own commodities and
snatching five from Household 1. When the purpose of analyzing a situation
is focused entirely on understanding the result of a particular structure
and the analyst is certain about the ordinal ranking of participants’
values over outcomes, then this abbreviated process of representing outcomes
in utility space is an effective means of analyzing a situation.
When the analyst wants to understand how rules or attributes of the
biophysical world or the community change an action situation, greater
precision is obtained by separately analyzing the biophysical outcomes
and then the value assigned by participants to outcomes. In the Snatch
Game, the physical outcomes are the actual bundles of commodities that
exist at each end point. If no offer is made or an offer is rejected, both
households end up with ten units of commodities. If Household 2 successfully
snatches Household 1’s commodities offered for sale, it ends up with
fifteen commodity units and Household 1 is left with five units of their
own commodity. If they reach a successful exchange, each household now
has five sacks of potatoes and five chickens, which is assigned a utility
valuation of fifteen units.
As discussed below, external benefits and costs are frequently assigned
to outcomes by payoff rules. If there were a market where commodities
were exchanged at known prices, one could assign a monetary value to
the commodities. If there were taxes imposed on the exchange of commodities
(a sales tax), one could represent the outcomes in a monetary
unit representing the market price minus the tax. If one wanted to examine
the profitability of growing chickens as contrasted to potatoes or other
crops, one would represent the outcomes in terms of the monetary value
of the realized sales value minus the monetary value of the inputs (land,
labor, and other variable inputs).
  
L IN KI NG AC TI ON SI TU AT IO NS 43
Combining biophysical outcomes, external payoffs, and participants’
internal valuation into one measure is useful for making decisions in a
static setting. It does not, however, enable an analyst to identify how specific
rules might affect the structure of a situation. The set of physically
possible actions and resulting transformations remains the same if payoff
rules are the only rules to be changed affecting a situation. What is affected
by a change in payoff rules is the net level of benefits or costs assigned
to a particular path of actions and physical results. Thus, to examine
the effect of rules in a careful and systematic manner, one needs to
consider the underlying physical transformations separately from the material
rewards assigned a chain of actions and results.
One example of the essential difference is between the amount of goods
produced during a particular time period and the net financial sums assigned
to participants in positions (workers, managers, owners, etc.) for
that time period. It is also important to consider the internal valuation
placed by a participant on the rewards and costs assigned to physical
results. Thus, in the approach taken in this volume, there are three components
to what individuals value as outcomes: (1) the physical results obtained
as a result of a chain of actions by participants, (2) the material
rewards or costs assigned to actions and results by payoff rules, and (3)
the valuation placed on the combination of the first and second components
by the participants. The valuation placed on exchanging potatoes
for chickens would be quite different—perhaps even negative—for a vegetarian
household, as contrasted to one that enjoyed eating chicken.
The state variables affected by a situation, however measured, are what
the participants in an action situation are thought of as affecting. A frequent
assumption is made that the participants self-consciously decide to
affect particular results or to leave them in as they were in the status
quo. Analysts can also include “unintended outcomes” within the set of
potential outcomes included in an action situation. Polluters, for example,
frequently do not know the full range of physical results generated as a
result of their actions.
THE OPPORTUNITY INVOLVED IN AN ACTION SITUATION
The “opportunity” involved in an action situation can be defined as the
range of the values of the outcome variables potentially affected in a situation
(von Wright 1966, 124; Commons [1924] 1968, 67). If one variable
is affected, the opportunity existing in a situation is the difference between
the highest and lowest achievable value on that outcome variable. When
that variable is dichotomous, its presence or absence is the full range of
variation. The range on a quantified variable may be small or great. The
level of opportunity in a market exchange, for example, is the difference
  
C HA PT ER TW O44
between the lowest price offered by a potential buyer and the highest price
offered by a potential seller. When that range is severely constrained, the
market does not offer much opportunity for affecting the price that a
buyer obtains or a seller must pay. When the range is large, the market
offers considerable opportunity for affecting the price (Commons [1924]
1968, 66–67). When multiple state variables are included as potential
outcomes, the concept of opportunity becomes more difficult to measure
but involves a cumulative score of the range of all the state variables affected
using concepts such as net benefits or utility.
THE STATUS QUO OUTCOME
The status quo is a concept used to describe the relationship between the
values of the end-state variables compared to the values of the initial-state
variables. If none of the values on any of the state variables has changed,
one concludes that the outcome remains the “status quo.” In the Snatch
Game, the status quo exists under two conditions: (1) if the first player
does not make an offer or (2) if the first player does make an offer and
the second player refuses it. When one conceptualizes the outcomes of a
legislative session as the values of a set of policy variables, the status quo
is always included in the set of potential outcomes. Unless there is a minimum
winning coalition within those authorized to change particular policy
variables, the status quo policy will continue in effect.
As discussed in some detail in the last half of this chapter, the outcome
of one action situation may include the possibility or necessity of proceeding
to another situation. Thus, the outcome of a committee hearing in
the U.S. House of Representatives may be the successful reporting out of
proposed legislation. Combined with this success, however, is the necessity
to take action on the floor of the House, in a committee of the Senate,
on the floor of the Senate, and potentially in a combined Senate-House
committee that reports back to both houses that again reconsiders the
legislation one more time (Shepsle and Weingast 1984).
If one action situation is not reached until after particular outcomes are
achieved in other action situations, the first situation can be considered a
“contingent action situation” (Coleman 1973, 64). The internal behavior
of participants in most organizations can be analyzed as a complex series
of differentially structured action situations linked by a set of procedural
rules. When analyzing the “actions” of one organization in a multiorganizational
setting, many of these internal situations will not be separately
analyzed and some overall method for determining the most likely action
to be taken by a particular type of organization in a specific type of action
situation will be substituted for the full chain of linked situations. In work
  
L IN KI NG AC TI ON SI TU AT IO NS 45
on two-level games, McGinnis and Williams (1989) analyze international
bargaining situations where they examine both the interaction of nations
in a bargaining situation and then the internal structure of relationships
within the respective governments.5
Actions
Participants assigned to a position in an action situation must choose
from a set of actions at any particular stage in a decision process. An
action can be thought of as a selection of a setting or a value on a control
variable (e.g., a dial or switch) that a participant hopes will affect an
outcome variable. In game theory, the set of actions available to a participant
at a specific sequence in a game—a decision point—is the available
moves. The specific action selected by a participant from the set of authorized
actions is called a choice. In the Snatch Game, the first player has a
choice between two moves (offer or not offer) and the second player a
choice among three moves (refuse, snatch, or exchange). A complete specification
of the moves to be taken in all possible contingencies in a oneshot
or repeated game is called a strategy.
The types of variables included within the concept of a set of authorized
actions are vast. Examples of typical action sets available to participants
who hold positions in different types of situations are arrayed in table
2.1. The term “action” thus includes both overt acts as well as the choice
not to act in some situations or “forbearances” (von Wright 1966). Both
voting for one or another candidate and abstaining from voting are included
in a voter’s action set, for example. Both can be conceptualized as
a setting on a control variable.
Action-Outcome Linkages
A setting on a control variable is considered “linked” to a state variable
when it is possible to use that setting to cause the state variable (1) to come
into being, (2) to disappear, or (3) to change in degree. A light switch, for
example, is a control variable with two positions—on and off. It is linked
to some source of light that shines or does not shine. By turning the switch
to off, one can make the light disappear. By turning the switch to on, one
can produce light. In this situation, the control variable does not change
unless a human actor takes a positive action to change it. Since the state
variable may also change as a result of some physical process (such as the
light bulb burning out), the actor can be said to have some effect on the
outcome by knowing the linkage and choosing whether to change the
  
C HA PT ER TW O46
TABLE 2.1
Action sets assigned to positions in several “typical” action situations
Position to which action set is assigned Authorized actions included in action set
A voter in an election Vote for Candidate i, Candidate j, or
abstain from voting
A seller in a competitive market Sell up to X quantity of good Y at the
going price; do not sell any of good Y at
the going price
A member of a legislature Submit an amendment to add provisions
to an already submitted bill; submit
an amendment to delete or change
provisions to an already submitted
bill; submit no amendments
A chair of a committee Order the votes on an original bill (B), an
amendment (A), and the status quo (Q)
in any of the three following ways:
(1) A vs. B and then the winner vs. Q
(2) A vs. Q and then the winner vs. B
(3) B vs. Q and then the winner vs. A
setting (take a positive action) or to retain the old setting (to forbear from
taking a positive action).
A state variable may be linked to many control variables. One might
think of a situation in which three switches jointly control a light—at
least two of them must be in the “on” position for the light to appear. A
person authorized to set one of the switches to on or off can potentially
affect whether the light is on or off, but cannot totally control the presence
or absence of light. If only one other switch is turned on, a person assigned
to one of the other two switches can either make the light appear (by
turning their switch on) or can produce darkness (by turning their switch
off). If one person’s switch is already off, the other person can allow darkness
to continue by refraining from changing his or her control switch. It
is in this sense that a “nonaction” may affect an outcome variable. Since
all three light switches are linked to the light, choices to change the switch
position or leave it the same are both considered to be an action.
To give some political content to these concepts of action and transformation
functions, let us analyze the situation facing the chair of a fiveperson
county council located in a state that authorizes local options regarding
the legality of alcohol being sold within the county boundaries.
The other four members of the council are divided equally on the issue.
On a tie vote, a chair can vote in favor of a law, vote against a law, or
abstain. What effect his or her positive action (voting for or against) or
  
L IN KI NG AC TI ON SI TU AT IO NS 47
TABLE 2.2
Results if collective-choice rules specify that a tie vote retains the status quo
Selling liquor in Legislation
the county is proposes Result with tie Act of chair Final outcome
Illegal Illegal Illegal Abstain Illegal
(Forbearing to let a state remain absent)
Illegal Legal Illegal Vote Yes Legal
(Acting to make a state come into being)
Legal Illegal Legal Vote Yes Illegal
(Acting to make a state disappear)
Legal Legal Legal Abstain Legal
(Forbearing to let a state continue)
TABLE 2.3
Results if collective-choice rules specify that a tie vote leads to a change
Selling liquor in Legislation
the county is proposes Result with tie Act of chair Final outcome
Illegal Legal Legal Vote No Illegal
(Acting to keep a state from occurring)
Illegal Legal Legal Abstain Legal
(Forbearing to let a state come into being)
Legal Illegal Illegal Abstain Illegal
(Forbearing to let a state disappear)
Legal Illegal Illegal Vote No Legal
(Acting to prevent a state from disappearing)
forbearance (abstaining) has depends on the rules governing what happens
if there is a tie vote, as shown in tables 2.2 and 2.3.
Any specific situation would not have all eight elementary modes of
action and forbearance present at once. Which subsets of the eight would
be included in an action situation obviously depend on the collectivechoice
rules of a particular situation regarding what happens with a tie
vote. Actions designed to produce a certain state or an outcome are frequently
called production functions. Various combinations of materials,
time, and effort are used to produce a commodity. Transformation functions
are not all as determinate as those presented above. Stochastic transformation
functions relate actions to outcomes in a probabilistic manner.
Thus, the concepts of certainty, risk, and uncertainty are relevant to considerations
of these linkages.
  
C HA PT ER TW O48
CERTAINTY, RISK, AND UNCERTAINTY
In action situations that are formally analyzed, the linkage of actions to
outcomes is usually presumed to be knowable. In such situations, one
could represent the chain of actions and results as a tree, as in the Snatch
Game. When the full tree or matrix is presumed to be known to the participants,
analysts refer to three types of linkage between actions and outcomes:
certainty, risk, and uncertainty.
In a certain linkage, every available action is linked directly with one
and only one outcome. The perfectly competitive market, as represented
in neoclassical economic theory, links actions to outcomes in a certain
manner. Neither sellers nor buyers in a perfectly competitive market have
any control over price. The price at which alternative quantities of a product
can be bought and sold is determined by market forces, and thus certain,
in the short run. Offers to buy or sell at the market price lead to one
and only one outcome.
The analytical world of certainty is vast and includes much of the formal
theory in economics and management science. Efforts to find maxima
and minima of functions, feasible regions, production possibility frontiers,
and so forth, are all modes of analysis appropriate for situations
characterized by certainty. Outside of formally organized large-scale markets,
few interactive situations are likely to have one-to-one relationships
between actions and outcomes.
Action-outcome linkages that are considered to be risky or uncertain
involve one-to-many relationships between actions and outcomes, like
both games presented above. In a situation characterized by risk, the objective
probabilities’ relationships between each action and set of outcomes
can be known (F. Knight 1921; Cashdan 1990). A classic situation
of this type is the urn filled with a known number of red and black balls
where the individual must decide upon actions that can lead to one or
another outcome depending upon the color of the next ball drawn from
the urn. A roulette wheel is an instrument of risk. Insurance firms face a
world characterized by risk when they calculate premiums to be charged
for different types of insurance policies. Once data about the distribution
of past events are available, probabilities can be assigned to different distributions
of future events. When the probability that particular actions
will lead to particular outcomes is known, a probability matrix can be
developed that indicates the probability of each outcome occurring given
particular combinations of states of affairs and actions.
An essential indeterminacy of social interactions remains in the context
of many action situations. This indeterminacy is called uncertainty (F.
Knight 1921). When institutional arrangements leave open wide avenues
for choice, and each individual’s outcome is dependent upon the actions
  
L IN KI NG AC TI ON SI TU AT IO NS 49
taken by others, uncertainty will characterize the resulting decision situation.
Uncertainty characterizes a situation in which the probabilities of
specific actions leading to outcomes are unknowable. The set of actions
and the set of outcomes are still assumed to be finite and knowable. The
linkages between actions and outcomes are also presumed to be knowable.
Most formal games are characterized by uncertainty, as are most
situations existing in the political realm. The potential decision of the
individual is taken into account in the decision making of the “other.”
Each individual must decide in the presence of at least one “strategic
other” whose decision to act will be contingent upon expectations about
the individual’s action (Coleman 1973, 42). For situations that can be
represented as formal games, the Nash equilibrium and its refinements
are a method for predicting the likely action-outcome linkages assuming
that all participants view the situation with common knowledge about its
structure and the rationality of the other players. Anthropologists and
ecologists have delved into a number of fascinating questions related to
the presence of substantial risk in the ecology in which humans find themselves
(see Kaplan, Hill, and Hurtado 1990; Low 1990). Hawkes, O’Connell,
and Blurton Jones (2001), for example, explain the frequently observed
relationship that hunted meat from large animals is widely shared,
while food obtained from subsistence agriculture is not as related to the
difference in the risk of obtaining food from these two activities (see also
Kameda, Takezawa, and Hastie 2003 for an excellent synthesis of the
debate over explanation).
The distinction presented here among certainty, risk, and uncertainty
focuses entirely on structural aspects of the situation (linkage patterns)
and not on the level of information that an individual has about the situation.
Theorists sometimes blur this distinction. This blurring leads to considerable
confusion when they treat the uncertainty in the situation and
the lack of information by an individual as the same phenomena. The
separation of situational and cognitive aspects of uncertainty draws heavily
on the earlier work of F. Knight (1921) and Buchanan and di Pierro
(1980), who stress the need to distinguish between the structure of the
situation and the cognition of the individual. H. Simon (1955) and Williamson
(1975) also distinguish between situational complexity and uncertainty
on the one hand and cognitive limitations on the other.
Control
The extent of control over the linkage of the action to outcomes that a
participant has varies from absolute to almost none. An individual has
total control (omnipotence) over an outcome variable, which we may call
oi in a situation if for each value oi potentially affected within that situa-
  
C HA PT ER TW O50
tion there is an action ai , such that the conditional probability of oi given
ai equals one. For two-dimensional outcomes, an individual has total control
if for each combination of outcome variable values oioj there is an ai,
such that the conditional probability of oioj given ai equals one (Coleman
1973, 61).
An individual has partial control over a state variable if the conditional
probability of a change in the value oi of the state variable given an action
ai available to the individual is greater than zero and less than one. Partial
control can, thus, vary from an extremely small chance of affecting an
outcome to a high probability of affecting the outcome. A participant can
be said to be impotent with respect to an outcome when he or she has no
control over the values of a state variable (see von Wright 1966, 129–31
for a discussion of control that is similar to that of Coleman).
The “power” of an individual in a situation is the value of the opportunity
(the range in the outcomes afforded by the situation) times the extent
of control. Thus, an individual can have a small degree of power, even
though the individual has absolute control if the amount of opportunity
in a situation is small. The amount of power may also be small when the
opportunity is large, but the individual has only a small degree of control.
Action situations may involve differential distributions of control and opportunity
to different individuals in the situation. Consequently, individuals
may differ in the amount of power they have in the situation. Concepts
of opportunity, control, and power are thus defined as situation-dependent.
A single individual may have dramatically differing levels of control,
opportunity, and power in the different action situations in which they
participate. An executive may have more power in regard to those issues
where the executive must initiate action than when the executive can only
stop action.
Information about the Action Situation
Participants in an action situation may have access to complete or incomplete
information. Almost all formal representations of action situations
assume that participants have access to complete information by which
is meant that each participant could know the number of other participants,
the positions, the outcomes, the actions available, how actions are
linked to outcomes (and thus the certainty, risk, or uncertainty of the
linkage), the information available to other players, and the payoffs of
the same. In other words, complete information is an assumption that
each participant could know the full structure of an action situation as
defined here.
When participants are assumed to have access to complete information,
a further distinction is made in formal theory between perfect and imper-
  
L IN KI NG AC TI ON SI TU AT IO NS 51
fect information. When a participant has access to perfect information,
they cannot only know all of their own past actions, they can also know
the actions of all other players before they make any move. In other
words, they can know the exact decision node at which they are making
a choice. Household 2 in the Snatch Game has perfect information. Under
imperfect but complete information, the individual is assumed to have
access to knowledge of the full structure of the situation, but may not
have access to knowledge about all the moves that other participants have
taken prior to a particular move. The participants could know all the
possible nodes at which he or she could be, but is unable to distinguish
the exact node for the current move.
When information is less than complete, the question of who knows
what at what juncture becomes very important. With incomplete information,
how much any one individual contributes to a joint undertaking is
often difficult for others to judge. When joint outcomes depend on multiple
actors contributing inputs that are costly and difficult to measure,
incentives exist for individuals to behave opportunistically (Williamson
1975). Opportunism—deceitful behavior intended to improve one’s own
welfare at the expense of others—may take many forms, from inconsequential,
perhaps unconscious, shirking to a carefully calculated effort to
defraud others with whom one is engaged in ongoing relationships. The
opportunism of individuals who may say one thing and do something else
further compounds the problem of incomplete information.
In many work situations, a boss cannot know exactly what employees
are doing and how this adds to or takes away from the accomplishment
of a joint output. Developing a contract that motivates them to be highly
productive has long been recognized as a challenge (Barnard 1938). In
the contemporary economics literature, the problem of a principal negotiating
with an agent—when the agent’s actions cannot be fully known
to the principal—is known as the Principal-Agent problem (Fudenberg,
Holmstro¨m, and Milgrom 1990; Laffont and Martimort 2002). Asymmetric
information problems also occur when individuals try to develop
a contract to share risk. An insurance firm can never know the intentions
and behavior of those it insures. Thus, all insurance situations involve a
moral hazard problem (Arrow 1963).
When action situations are being analyzed in a less formal manner,
analysts are more apt to assume that participants have access to incomplete
information about the full set of actions available to them, the full
set of outcomes, and about how actions are linked to outcomes. This
assumption is frequently linked to fundamental assumptions made about
the limitations of human cognitive abilities in a model of the individual.
However, the usefulness of an assumption of complete or incomplete information
also relates to the complexity of the action situation being mod-
  
C HA PT ER TW O52
eled and the language structure used, the channels of communication, and
the type of communication normally assumed to be present in the action
situation.
Costs and Benefits
The discussion of outcomes, and of action-outcome linkages presented
above, relies on the relationships between control variables and state variables.
In addition to the physical actions and outcomes that are involved
in a situation, rewards and/or sanctions may be distributed to participants
in positions dependent upon the path taken to achieve a particular outcome.
Costs and benefits are cumulative. External rewards or sanctions
assigned to outcomes include the rewards (financial returns) or sanctions
(taxes or fines) assigned to actions taken along a path to an outcome.
For simplicity, it is frequently assumed in theoretical models that acts
are costly and outcomes are beneficial. Participants are then viewed as
weighing the costs of an action against the benefits of an outcome. Actions
may, however, have associated benefits, and outcomes may be “bads”
instead of “goods.” In the Snatch Game, for example, one of the outcomes
is very bad for Household 1 and very good for Household 2. All actions
are represented in figure 2.2 as having zero costs. If the place where
Households 1 and 2 meet is located some distance away from each of
their farms, the cost of traveling to this location could be included in a
payoff table. Thus, the final external value assigned to an outcome for a
participant in a position is the net value of the path taken to an outcome
given the choices made by the participant and those of others. The external
value to a seller in a fully developed market setting is the profit resulting
from the sale of a quantity of goods minus the cost of buying or
producing those goods.
In conducting an institutional analysis, a distinction needs to be made
between a physical outcome, an external reward or sanction, and the valuation
that a participant assigns to the physical transformation and external
rewards and/or sanctions. So long as the physical linkage between
actions and outcomes remains the same, the outcome remains the same.
A firm following a routine production process generates, on average, the
same quantity of a product per unit of time. If the price of the commodity
increases or decreases substantially, or a new tax is imposed on the sale
of the commodity, the net profits obtained differ substantially. The external
or extrinsic values are the set of “reward” variables affected by the
path of actions and outcomes. Examples of extrinsic rewards include the
financial returns assigned a worker in a Principal-Agent contract. The
reward may be assigned strictly on action variables (e.g., how many hours
the worker clocks in), strictly on outcome variables (e.g., how much of a
  
L IN KI NG AC TI ON SI TU AT IO NS 53
particular final or intermediate product is produced), or on some combination
of action and outcome variables (a wage plus a production bonus).
Extrinsic benefits and costs are frequently assigned through the operation
of a rule system and thus do not rely only on biophysical processes.
If the primary interest is to predict what will happen in a particular
situation, and not how to change the situation, the only value that a researcher
needs to use is the net value assigned by participants to the
achievement of an outcome. This value is referred to as utility in economics
and game theory. Utility is a summary measure of all the net values to
the individuals of all the benefits and costs of arriving at a particular
outcome. Game theorists use utility values in representing an outcome
and only rarely decompose this into component costs and benefits. Some
individuals may pay attention only to the objective, extrinsic valuation
assigned to an action-outcome link. For simplicity, many analysts assume
that subjective utility is positively associated with the net value of the
external rewards. In economics, theorists normally assume that utility is
monotonically associated with profits, for example. As discussed extensively
in chapter 4, this assumption is reasonable to make in many but not
all situations. Individuals may assign either a positive or negative intrinsic
value to actions or outcomes. Participants in the same situation may not
even assign the same internal valuation to the objective outcome (see
Schiemann 2002).
The intrinsic valuation attached to an external reward or sanction is
the internal value that individuals associate with the components of the
objective transformations and rewards (Frey 1997a; Deci 1975). Joy,
shame, regret, and guilt are all forms of intrinsic values. If the person
evaluates an action as being improper, they may assign a negative intrinsic
value. If the person is proud of an action, they may assign it a positive
intrinsic value. Measuring intrinsic valuation is extremely challenging. In
fact, dissecting the holon referred to as “the participant” in this framework
is among the more challenging tasks an institutional analyst faces.
Thus, all of chapter 4 is devoted to a beginning elucidation of this process.
The Number of Times the Action Situation Will Be Repeated
In addition to the internal components of an action situation, it is also
important to know whether the situation is a one-shot or a repeated interaction.
Analysts usually agree, for example, that the outcomes that individuals
will obtain in a social dilemma game like a Prisoner’s Dilemma
or the Snatch Game will depend, among other factors, on whether the
participants are engaged in a one-time encounter or over an indefinitely
long sequence of plays. The predicted equilibrium in a single-play Snatch
  
C HA PT ER TW O54
Game (assuming that both players do not know one another and value
only the objective outcomes) is that neither participant would trust the
other when in the first position. Thus, both players are worse off than
they could be if they exchanged commodities. If this game were repeated
a finite number of times, most game theorists would predict the same
equilibrium as for a single-round game due to backward induction. The
last play of this game is similar to the single-round game. Given this,
analysts predict that participants will not cooperate on this last round.
Once the participants have decided not to cooperate on the last round,
they would also decide not to cooperate on the second-to-the-last round,
and so forth, back to the first round.
When participants in a social dilemma game are placed into an indefinite
series of rounds, however, the disadvantage of continued lack of cooperation
can lead them to adopt a conditional cooperative strategy so
long as other participants also cooperate. The well-known folk theorem
of game theory establishes that full cooperation is one of the feasible equilibria
that participants in an infinitely repeated (or even an indefinitely
repeated) situation may achieve if they use one of several conditional cooperative
strategies (Kreps et al. 1982). It is, however, only one of many
equilibria. Thus, participants face a challenging coordination problem in
reaching this outcome.
One of the most famous and useful conditional strategies is “Tit for
Tat,” where players in a repeated, symmetric social dilemma cooperate
on their first round and then take whatever action their counterpart took
in the last round (Axelrod 1984). If there are only two participants in a
repeated, symmetric social dilemma situation, they can monitor and punish
one another through future actions. Thus, whenever one of the participants
tries to take advantage of the other participants, the second participant
can directly punish the first participant on the next round. Once a
group is larger than two, simple strategies or heuristics like Tit for Tat are
much harder to implement since an error by one person may set the entire
group into a never-ending series of retributions.
The possibility of participating in a repeated symmetric situation, then,
may be enough to lead small groups of individuals to seek out mutually
advantageous strategies when they might not otherwise do so. For asymmetric
social dilemmas, however, such as the Snatch Game, there is no
single, simple heuristic—like Tit for Tat—that can be successfully employed
to reach the higher joint outcome. This is due to the fact that the
first player can simply stay with the status quo, and what the second
player would do is irrelevant. The second player does not have an easy
method for “punishing” the first player in later rounds as there may be
no later rounds. Scholars are working on a set of heuristics that are pos-
  
L IN KI NG AC TI ON SI TU AT IO NS 55
ited to help individuals cope with asymmetric, social dilemma situations
where trust and reciprocity are so very important (see Rieskamp and
Gigerenzer 2003).
Linking Action Arenas
In addition to analysis that digs deeper into the exogenous factors affecting
action arenas, an important development in institutional analysis is
the examination of linked arenas. Whereas the concept of a “single” arena
may include large numbers of participants and complex chains of action,
most of social reality is composed of multiple arenas linked sequentially
or simultaneously (see Shubik 1986). Rarely do action situations exist
entirely independently of other situations. Political scientists frequently
study processes within a legislature where decisions are first made in a
committee, then in one house, then in a second house, then in conference
committees, then by the chief executive. Decisions inside one unit of a
private firm must also go through a sequence of action situations before
they are final. Given the importance of repeated interactions to the development
of a reputation for reciprocity and the importance of reciprocity
for achieving higher levels of outcome over time, individuals have a strong
motivation to link situations so as to utilize the capabilities achieved
through gaining a reputation in one situation that helps to provide credibility
in others.
Action situations are linked in two different and important ways. The
first kind is an organizational linkage. Within larger organizations, what
happens in the purchasing department affects what happens in the production
and sales departments and vice versa. For some purposes, it is
useful to isolate a single situation in an organization to try to understand
why particular kinds of outcomes are achieved in that situation without
embedding the situation of interest in a larger lattice of related situations.
Further, interesting and important institutional arrangements for coordinating
complex chains of actions among large numbers of actors involve
multiple organizations competing with one another according to a set of
rules. Markets are the most frequently studied institutional arrangements
that achieve coordination by relying primarily on rule-governed, competitive
relationships among organizations. Rule-governed competition
among two or more political parties is considered by many analysts to be
an important requisite for a democratic polity. Less studied, but potentially
as important a means for achieving responsiveness and efficiency in
producing public goods and services are arrangements that allow rule-
  
C HA PT ER TW O56
ordered competition among two or more potential producers of public
goods and services (see McGinnis 1999b).
A more fundamental form of linking is shifting levels of analysis from
one situation to a deeper rule-changing situation. All action situations
where individuals engage in the provision, production, distribution, appropriation,
assignment, or consumption of goods and services are classified
as operational situations. Operational situations are themselves affected
by the biophysical world, the rules affecting the structure of the
operational situation, and the attributes of the community in which the
situation was embedded. If operational situations were all that is analyzed
in this book, we would not need to go further than the presentation of
the Institutional Analysis and Development (IAD) framework for a single
level of analysis. In this case, the rules analyzed are presumed to be fixed
and the linkage between a situation at one level (where rules are assumed
to be fixed and unchanging) and another level (where rules are themselves
the outcomes produced) is unnecessary.
We do, however, want to understand how operational situations are
linked to a deeper situation where the rules of the operational situation
are made or modified. Not only are there two levels of analysis; one can
dig deeper into where the rules come from in a collective-choice situation
to examine constitutional-choice situations. For practical reasons, we
stop the digging at a metaconstitutional level that is relevant when individuals
are in the process of constituting or reconstituting ongoing relationships.
A metaconstitutional level occurs when participants examine
the consequences of diverse rules for who is to represent them in a constitutional
process.
In this chapter, we will focus first on linked action situations where
the outcomes from one situation are eventually fed into a series of other
situations, but where rules themselves remain relatively constant. Then
we will address the deeper question of how to examine the linkage between
one situation and its immediate deeper-level situation where the
rules of the first situation may be changed.
Organizational Linkages
Behavior within any particular situation may depend upon expected outcomes
in another situation. As mentioned above, a potential outcome of
one situation may be to enter a second situation. Some social processes
may need to be thought of as composed of a series of linked situations.
The outcomes of any one situation become inputs into the next situation.
The intermediate outcomes of an early situation may not have much value
unless the full series of linked situations is completed. Thus, getting a bill
  
L IN KI NG AC TI ON SI TU AT IO NS 57
passed by the House of Representatives can be viewed as an intermediate
step in finally getting a bill passed into law. The full set of situations within
the Senate and the signature by the president are necessary before a bill
becomes a law. The intermediate outcome is a necessary step for the final
outcome, but may not have much assigned worth.
An organization is composed of one or more (usually more) action situations
linked together by prescriptions specifying how outcomes from one
situation become inputs into others. Organizations may be thought of as
a tree or a lattice with situations at each node. A particular set of rules
structures the situation at each node. A general set of rules partially structures
all internal situations and specifies the paths that may be chosen
from one situation to the next. Thus, a tournament is one form of organization
that prescribes how players will proceed through the tournament
tree. Many bureaucratic organizations have a similar general structure to
that of tournaments.
Where one draws the boundaries on the analysis of linked situations
depends on the questions of interest to the analyst. In regard to collectivechoice
action situations, one may want to understand why some kinds of
candidates are selected in elections and not others. If this is the only question,
one can focus on primary elections and the relationships among
diverse supporters and the policy position of candidates in a “candidate
selection game” (see Downs 1957 for the classic analysis of this situation).
Other situations are only potentially linked together. Thus, many commercial
transactions occur within organizations or across market situations.
Most of these occur without any reference to a court. However,
invoking court proceedings is a very important method available to participants
in these transactions to constrain the action set of other participants
to those that have been agreed to. If it is extremely costly to use
court proceedings, then one would predict that the action sets of participants
would be more likely to include some actions that are illegal or are
outside the bounds of a contract. Thus, the relative ease and availability
of such action situations as courts to serve as monitors, sanctioners, and
to provide remedies is extremely important in understanding the behavior
of actors within any particular situation.
Actions and outcomes that occur in one location may also stimulate
reactions by others located in another situation located elsewhere. Firms
that are highly profitable are frequently used as models by others as to
how to organize themselves for success. Farm households who innovate
and are successful or common-property arrangements that increase their
joint yield are frequently copied by others. These connections are like a
ripple across the landscape rather than strongly linked situations.
  
C HA PT ER TW O58
Multiple Levels of Analysis
Besides multiple and nested action arenas at any one level of analysis,
nesting of arenas also occurs across several levels of analysis. All rules are
nested in another set of rules that define how the first set of rules can be
changed. The nesting of rules within rules at several levels is similar to
the nesting of computer languages at several levels. What can be done at
a higher level will depend on the capabilities and limits of the rules at
that level and at a deeper level. Whenever one addresses questions about
institutional change, as contrasted to ongoing actions within institutional
constraints, it is necessary to recognize that:
1. Changes in the rules used to order action at one level occur within a currently
“fixed” set of rules at a deeper level.
2. Changes in deeper-level rules usually are more difficult and more costly
to accomplish, thus increasing the stability of mutual expectations among individuals
interacting according to the deeper set of rules.
As mentioned earlier in the chapter, it is useful to distinguish three levels
of rules that cumulatively affect the actions taken and outcomes obtained
in any setting (Kiser and Ostrom 1982).6
Operational rules directly affect
day-to-day decisions made by the participants in any setting. These can
change relatively rapidly—from day to day. Collective-choice rules affect
operational activities and results through their effects in determining who
is eligible to be a participant and the specific rules to be used in changing
operational rules. These change at a much slower pace. Constitutionalchoice
rules first affect collective-choice activities by determining who is
eligible to be a participant and the rules to be used in crafting the set
of collective-choice rules that, in turn, affect the set of operational rules
(Brennan and Buchanan 1985; Buchanan and Tullock 1962). Constitutional-choice
rules change at the slowest pace. One can even think about a
“metaconstitutional” level underlying all the others that is not frequently
analyzed. One can think of the linkages among these rules and related
level of analysis as shown in figure 2.3.7
For most practical applications, three levels are enough.8
No theoretical
justification exists for three and only three levels (Diermeier and Krehbiel
2003). For the purposes of formal theory, we may need to assume as long
a series of layers as is needed until we hit rock bottom—the biophysical
world. Very deep layering—even infinite layering—turns out to be needed
in many aspects of formal theory. Game theorists, for example, have had
to assume that the common understanding needed for one to assume that
there is a game is nested infinitely. “Information is common knowledge if
it is known to all players, each player knows that all of them know it,
and each of them knows that all of them know that all of them know it,
  
L IN KI NG AC TI ON SI TU AT IO NS 59
CommunityOperational
Rules-in-Use
Biophysical World
CommunityCollective Choice
Rules-in-Use
Biophysical World
CommunityConstitutional
Rules-in-Use
Biophysical World
CommunityBiophysical World
Individuals' Actions Taken that Directly Affect
Rules that Affect Constitutional Situations
METACONSTITUTIONAL SITUATIONS
(Prescribing, Invoking, Monitoring, Applying, Enforcing)
Individuals' Actions Taken that Directly Affect
Rules that Affect Collective-Choice Situations
CONSTITUTIONAL SITUATIONS
(Prescribing, Invoking, Monitoring, Applying, Enforcing)
Individuals' Actions Taken that Directly Affect
Rules that Affect Operational Situations
COLLECTIVE-CHOICE SITUATIONS
(Prescribing, Invoking, Monitoring, Applying, Enforcing)
Individuals' Actions Taken that Directly Affect
State Variables in the World
OPERATIONAL SITUATIONS
(Provision, Production, Distribution, Appropriation, Assignment, Consumption)
Figure 2.3 Levels of analysis and outcomes. Source: E. Ostrom 1999, 60.
  
C HA PT ER TW O60
and so forth ad infinitum” (Rasmusen 1989, 50). Thus, one can always
assume that there are even more primitive rules underlying those that
one is analyzing at any one level—thus our positing of a very general
metaconstitutional level—until one gets to the constraints of a biophysical
world. At that point, the only distinction that can be made among actors
is their strictly physical strength, as Hobbes long ago recognized. For most
institutional analyses, three nested levels are sufficient.
At an operational tier, participants interact in light of the internal and
external incentives they face to generate outcomes directly in the world.
Examples of operational-level situations include:
• Families deciding to move or not move into an urban neighborhood depending
on the proportion of neighbors who share attributes (such as wealth,
race, ethnicity) with them in the potential future neighborhood as compared
to their present neighborhood (see H. Young 1998, 62–65; Schelling 1978).
• Workers hired by a boss to undertake a complex task who have private information
about their level of effort and face incentives to keep effort at a minimum
(for examples from a developed economy perspective, see Miller 2001;
and from a developing economy perspective, see Ensminger 2001).
• Harvesters from a common-pool resource deciding how much, when, and
with what technology to appropriate resource units (Gibson, McKean, and
Ostrom 2000).
The participants in collective-choice games may be the same participants
as in linked operational-choice games (as when all firms in an industry
agree upon a particular industry standard that they will all use in
manufacturing goods). Or, participants in collective-choice games may
differ from those in operational games. They may, for example, be legislative
representatives selected in electoral games (themselves part of the collective-choice
level of action) to be the agents of a set of principals—
the citizens engaged in a wide diversity of operational games that will be
affected over time by legislative policies. Examples of collective-choice
(policy choice) situations include:
• devising allocation schemes that can be used by NASA in allocating uses of
space stations (Plott and Porter 1996);
• elected representatives in diversely structured legislatures devising policies
based on popular preference distributions (Hammond and Butler
2003);
• citizens in diversely structured European countries making electoral decisions
in light of the information generated by their political systems (Benz
and Stutzer 2004).
Policy decisions affect the structure of arenas where individuals are making
operational decisions and thus impacting directly on a biophysical
world. The situation could as well be at a constitutional tier, where deci-
  
L IN KI NG AC TI ON SI TU AT IO NS 61
sions are made about who is eligible to participate in policymaking and
the rules that will be used to undertake policymaking.
Participants in the third level can again either be participants in the
other two levels or not. And, participants in constitutional choices may
not recognize that they are making a constitutional rule—they may be
simply trying to fix a problem with the way that they have been making
policy choices over the last several years. Examples of constitutionalchoice
arenas are:
• a group of groundwater producers meeting to discuss the formation of a
public district in order to develop regulations on the amount of water withdrawn
from the basin (Blomquist 1992);
• representatives from regions within a country debating the design or modification
of a national constitution (V. Ostrom 1987; Reynolds 2002; Dougherty
2001);
• representatives from countries in regions that are developing strategies to
create new multicountry unions for some purposes and possibly even to
change their borders (O’Leary, Lustick, and Callaghy 2001).
At each level of analysis there may be one or more arenas in which the
types of decisions made at that level will occur. In the collective-choice,
constitutional, and metaconstitutional situations, activities involve prescribing,
invoking, monitoring, applying, and enforcing rules (Lasswell
and Kaplan 1950; Oakerson 1994).
The concept of an “arena” as described earlier does not necessarily
imply a formal setting, but can include such formal settings as legislatures
and courts. Policy making (or governance) regarding the rules that will
be used to regulate operational-level choices is usually carried out in one
or more collective-choice arenas, as shown in figure 2.4. Constitutional
arenas are frequently formal arenas, but these may occur within specially
called conventions, within formal courts, or within a legislature that
changes a basic rule about making collective-choice rules.9
When a theorist chooses to analyze a situation at any particular level,
the theorist must assume that the institutional rules at that level are temporarily
fixed for the purpose of analysis. These rules form a part of the structure
of the situation rather than the solution to the game created by that
structure. When the purpose of analysis is to understand the origin of the
rules at one level, knowing the structure of the situation at the next level is
essential for that enterprise. The equilibria achieved at one level are thus
supported by equilibria that have been achieved at deeper levels.10
Understanding
the role of these nested levels does not, however, require that the
analyst specify the full supporting infrastructure in elucidating how individuals
are expected to behave at one level.11
Thus, assuming that there are
multiple levels where decisions are made that affect actions at other levels
actually greatly simplifies analysis rather than complicating it.
  
C HA PT ER TW O62
National, regional, and/or local
formal collective-choice arenas
Legislatures
Regulatory agencies
Courts
Informal gatherings
Appropriation teams
Private associations
Self-organized collective-choice
arenas
Operational
rules-in-use
Formal third-party monitoring
and enforcement activities
Informal third-party monitoring
and enforcement activities
Figure 2.4 Relationships of formal and informal collective-choice arenas. Source:
Adapted from E. Ostrom 1990, 53.
Level-Shifting Strategies
An individual engages in “level shifting” whenever he or she begins to
contemplate how to change any of the constraints on an operational situation
(or, on a collective-choice situation) that are potentially under the
control of the participants in that situation. A groundwater pumper, for
example, attempts to shift levels when he or she states to another pumper,
“Hey, this race of ours to withdraw water is going to destroy the basin
and leave us all worse off than we could be. Why don’t we change our
rules so as to avoid such a disaster?”
Any participant can shift levels of analysis in his or her own thinking
at any point in time while engaged in action at a different level. Before a
resource user suggests a rule to another user, the first has already analyzed
some of the consequences of adopting a revised set of rules. Shifting levels
of analysis can be accomplished by any actor operating independently of
others. That actor simply contemplates the opportunities and constraints
that might be available at a different level for solving some of the problems
occurring at a current level.
When the individual estimates that substantial benefits are likely to
occur if others agree to an actual shift of levels of action and to change
rules, the individual may then be willing to invest resources to try to convince
others that they should agree on a shift and consider the constraints
that are currently in effect. Shifting levels of action does not mean deciding
to change the rules by making the shift. Shifting enables only those
who shift to contemplate overtly a different set of rules (or other constraints
that may potentially be under their control, such as making a
capital investment in their joint enterprise.) One possible result of a shift
of levels may be to keep the status quo rules.
  
L IN KI NG AC TI ON SI TU AT IO NS 63
In field settings, it is hard to tell where one situation starts and another
stops. Life continues in what appears to be a seamless web as individuals
move from home to market to work (action situations typically characterized
by reciprocity, by exchange, or by team problem solving or command).
Further, within arenas, choices of actions within a set of rules as
contrasted to choices among future rules are frequently made without
recognizing that the level of action has shifted. So, when a “boss” says to
an “employee,” “How about changing the way we do X?” and the two
discuss options and jointly agree upon a better way, they have shifted
from taking actions within previously established rules to making decisions
about the rules structuring future actions. In other words, using IAD
language, they have shifted to a collective-choice arena.
The costs of shifting levels of decision making vary dramatically from
one setting to another. In some settings the same individuals are involved
in constitutional, collective-choice, and operational situations. Shifting
levels may be accomplished at low costs. A group of resource users discussing
common problems at an operational level may somewhat naturally
turn to a discussion of what could be done to solve any of their
current problems potentially including overuse, underinvestment, lack of
information, and the like. Part of this discourse may be focused on enforcing
their current rules more effectively. Another part of their discussion
may include reference to alternative rules that might reassign their own
rights and duties. After such a discussion, they may agree to stay with
their current rules or seriously evaluate the likely effect of changing rules
and agreeing on a change.
Formal procedures—including petitions, court hearings, legislation, and/
or referenda—may be required to shift levels of action in many situations.
Bureaucratic officials may control access to an arena in which rules and
other constraints could be changed. Judicial procedures may be used for
this purpose. The transaction costs of shifting levels and transforming an
ongoing situation may be very high. In such cases, participants at one level
may continue to rely for long periods on rules that produce suboptimal
outcomes at that level because the expected costs of changing rules are
higher than the benefits they could derive from a better set of rules. Alternatively,
they may devise their own de facto rules that they enforce themselves
since they cannot turn to authorities to enforce them. Developing de facto
rules outside formal channels may be less costly than trying to use the formal
channels available to participants in some political systems.
It is through shifting levels of action that participants may be able to
self-consciously design rules in their efforts to change patterns of undesirable
interactions and outcomes at operational or collective-choice levels.
The lack of a self-conscious examination of this strategy leads some
scholars to presume that individuals facing social dilemma situations have
no way out but continuing suboptimal outcomes or relying on “external
  
C HA PT ER TW O64
actors” to change the constraints they are facing. Garrett Hardin (1968),
for example, asks readers to picture a set of herders trapped in a “tragedy
of the commons” that they could not themselves overcome. In an open,
self-organizing society with considerable opportunities for individuals
to devise their own rules for the public and private enterprises they establish,
shifting levels of analysis is, however, something that participants do
frequently.
Once this possibility is seriously considered, however, it raises the problem
of how to analyze rules themselves rather than simply taking them as
unchangeable constraints. This is the major challenge that will be addressed
in chapters 5 to 9 of this volume. If rules are to be used as the
tools of a self-governing society, the diversity of tools themselves must be
examined. This is a nontrivial problem, especially given the importance
of the biophysical world and the broader communities in which rules are
only one factor affecting structure.
Predicting Outcomes
Depending upon the analytical structure of a situation and the particular
assumptions about the actor used, the analyst makes strong or weak inferences
about results. In tightly constrained, one-shot action situations
under conditions of complete information, where participants are motivated
to select particular strategies or chains of actions that jointly lead
to stable equilibria, an analyst can frequently make strong inferences and
specific predictions about likely patterns of behavior and outcomes. Game
theorists draw on solution concepts, such as the Nash equilibrium, to
predict outcomes.
Many field situations, however, do not facilitate making clear predictions.
Instead of completely independent decision making, individuals
may be embedded in communities where unobserved norms of fairness
and conservation may change the structure of a situation dramatically.
Within these situations, participants may adopt a broader range of strategies
than expected using an assumption of a narrow self-interest. Further,
individuals may change their strategies over time as they learn about the
results of past actions. The institutional analyst examining these more
open, less-constrained situations must make weaker inferences and predict
the patterns of outcomes that are relatively more or less likely to
result from a particular type of situation. In an experimental laboratory,
for example, giving subjects in many types of social dilemmas an opportunity
to communicate on a face-to-face basis is likely to substantially
change the strategies chosen by subjects to ones that are not predicted by
noncooperative theory (see E. Ostrom, Gardner, and Walker 1994 and
cites contained therein). This is not, however, a determinate process. The
  
L IN KI NG AC TI ON SI TU AT IO NS 65
variation in outcomes achieved is relatively large in groups that engage in
face-to-face communication, as discussed in chapter 3.
In field settings, one can assume that providing arenas where at least
some individuals engage in face-to-face discussions will usually change
the outcomes achieved. The important role of communication is achieving
a common understanding of the problems jointly faced. Discourse frequently
generates ideas concerning various ways of coping more effectively
with these problems. Further, communication helps participants to
learn what norms individuals share or do not share and whether sufficient
trust exists that individuals can adopt plans of actions that depend on
trustworthy behavior (see E. Ostrom 1998, 2001). Even large communities
whose forests are degraded can potentially turn a bad situation
around through their own efforts. Varughese (2000) documents how a
community of over twenty-five hundred people organized its own forest
association and created subcommittees so as to involve more members in
face-to-face discussion and decisions. This community has devised many
innovative strategies for improving forest conditions while trying to keep
the costs on community members relatively low. Many historical factors,
as well as the current structure of the situation, affect the likelihood of
organization and communication. Thus, no determinate predictions can
be made.
Some situations within any one of these levels may be simple enough
that one can generate a clear and empirically supported prediction about
likely behavior and outcomes—as one can, for example, in a highly competitive
market producing goods characterized by low costs of exclusion
and subtractability. Here, one can rely on well-tested results from prior
theoretical and empirical work. It is usually much more difficult to predict
results when one is analyzing a collective-choice or constitutional-choice
level situation as it impacts on operational-level settings. The process of
changing the property rights of inshore fisheries in New Zealand by the
national government involved substantial conflict and reorganization
throughout its early history (Yandle and Dewees 2003). From an effort
to create strictly private transferable rights, the system has evolved into
a complex but workable system. Comanagement institutions have been
crafted to complement individual property rights—something not predicted
by anyone when the initial institutional change was initiated (ibid.).
The problem of predicting outcomes is especially challenging when new
and unfamiliar collective- or constitutional-choice rules are selected.12
When new and unanalyzed situations are created by the process of changing
parts of a rule configuration, institutional analysis needs to proceed
to undertake a deeper analysis of how participants view the new rules,
how they come to understand them, how they will be monitored and
enforced, and what types of individual actions and collective outcomes
are produced. This is frequently a difficult and complex theoretical and
  
C HA PT ER TW O66
empirical task. As Mantzavinos, North, and Shariq (2004, 75) stress, the
“greatest challenge for the social sciences is to explain change—or more
specifically, social, political, economic, and organizational change.”
Evaluating Outcomes
In addition to predicting outcomes, the institutional analyst may also evaluate
the outcomes that are being achieved as well as the likely set of outcomes
that could be achieved under alternative institutional arrangements.
Participants in action situations and those observing these
situations use evaluative criteria that are applied to the outcomes as well
as the processes of achieving outcomes. The number of potential evaluative
criteria is large. In this chapter, we can only briefly focus on (1) economic
efficiency; (2) equity; (3) adaptability, resilience, and robustness;
(4) accountability; and (5) conformance to general morality.
Economic Efficiency
Economic efficiency is measured by the magnitude of the change in the
flow of net benefits associated with an allocation or reallocation of resources.
The concept of efficiency plays a central role in studies estimating
the benefits and costs or rates of return to investments that are often used
to determine the economic feasibility or desirability of public policies.
When considering alternative institutional arrangements, therefore, it is
important to consider how revisions in the rules affecting participants will
alter behavior and, hence, the allocation of resources.
Equity
Assessing equity is undertaken in two ways: (1) on the basis of the equality
between individuals’ contributions to an effort and the benefits they derive
and (2) on the basis of differential abilities to pay. The concept of
equity that underlies an exchange economy holds that those who benefit
from a service should bear the burden of financing that service. This is
called fiscal equivalence. Perceptions of fiscal equivalence or a lack thereof
can affect the willingness of individuals to contribute toward the development
and maintenance of public facilities.
Outcomes that tend to redistribute resources to poorer individuals are
considered to improve equity from a redistributional equity perspective.
Thus, in some cases, efficiency criteria would urge that scarce resources be
used where they produce the greatest net benefit, while equity goals would
urge an effort to benefit particularly needy groups. Likewise, redistributional
criteria may differ in rankings from those of achieving fiscal equivalence.
  
L IN KI NG AC TI ON SI TU AT IO NS 67
Adaptability, Resilience, and Robustness
Another criterion that can be applied to repeated situations is how behavior
changes over time. Do individuals learn from experience within an
action situation? Do they adapt to new circumstances as they arise or
become rigid in their responses over time? The concept of resilience that
originally was developed in ecology has now been applied to social systems.
Resilience is defined as the amount of disruption needed to transform
a system from stability domain (characterized by a configuration of
mutually reinforcing processes and structures) to another (Holling 1973;
Gunderson and Holling 2001). The concept of robustness tends to be
used more in engineering, while resilience was initially defined related to
ecological systems. Robustness refers to the maintenance of a system’s
performance even when it is subject to external, unpredictable disturbances
(Carlson and Doyle 2002; Anderies, Janssen, and Ostrom 2004).
Accountability
When evaluating collective-choice or constitutional-choice levels, one can
ask whether officials are accountable to citizens concerning the policies
and rules chosen. Without accountability, actors can engage in the various
opportunistic, strategic behaviors. Concern for accountability may not
conflict with efficiency and equity goals. Indeed, achieving efficiency requires
that information about the preferences of citizens be available to
decision makers, as does achieving accountability. Institutional arrangements
that effectively aggregate this information assist in realizing efficiency
at the same time that they serve to increase accountability and to
promote the achievement of redistributional objectives.
Conformance to General Morality
In addition to accountability, one may wish to evaluate the level of general
morality fostered by a particular set of institutional arrangements. Are
those who are able to cheat and go undetected able to obtain very high
payoffs? Are those who keep promises more likely to be rewarded and
advanced in their careers? How do those who repeatedly interact within
a set of institutional arrangements learn to relate to one another over the
long term? Are the procedures fair? (Frey, Benz, and Stutzer 2004)
The Need for Trade-offs
Trade-offs are often necessary in using performance criteria as a basis
for selecting from alternative institutional arrangements. Some criteria
are more important when evaluating patterns of actions and outcomes at
  
C HA PT ER TW O68
a collective-choice or constitutional-choice level than at an operational
level. The trade-off issue arises most explicitly in considering alternative
methods of funding public projects at a collective-choice level. Evaluating
how institutional arrangements compare across overall criteria is
quite a challenge. Analytical examination of the likely trade-offs between
intermediate costs is valuable in attempting to understand comparative
institutional performance (see E. Ostrom, Schroeder, and Wynne 1993,
chap. 5).
In this chapter, we have focused first on a core analytical unit of the IAD
framework—the action situation. This is where the institutional analyst
focuses on explaining results that affect the daily lives of participants as
well as the resources affected by participants. Fortunately, the same components—participants,
positions, actions linked to outcomes, control, information,
and costs and benefits—characterize collective-choice as well
as constitutional-choice arenas.
Given the importance of this theoretical concept, we will devote one
more chapter to illustrate the working parts and outcomes of differently
structured action situations. Chapter 3 moves into the experimental lab,
where the researcher must self-consciously create an action situation designed
to enable the testing of theoretical predictions. Chapter 3 will help
the reader really understand the working components of this core unit of
analysis.
Chapter 4 will then be devoted to the other component of an action
arena—the participants. The reader will be well aware by the end of chapter
3 that the primary model of individual behavior used extensively in
game theory, economics, and institutional analysis yields predictions that
are not supported in many experiments discussed in chapter 3 and the
literature cited therein. Thus, we face a substantial challenge in animating
institutional analyses.
Chapters 5 through 9 then turn to the question of how we can analyze
the rules used as tools to change the structure of action situations. Unfortunately,
the language used by social scientists to discuss the rules, norms,
and strategies used by participants in situations is extremely confusing.
Thus, considerable effort has to be expended in chapters 5, 6, and 7 to
develop a consistent, theoretical language to define and then use these
terms in a coherent and cumulative manner. Once this is accomplished,
we will illustrate the usefulness of the concepts that we have carefully
defined with extensive examples from empirical research.
  
Three
Studying Action Situations in the Lab
READERS OF PAST descriptions of the working components of the Institutional
Analysis and Development (IAD) framework have frequently urged
me to provide some examples of an action situation to help them make
this abstract concept more meaningful. Studying action situations in an
experimental laboratory turns out to be an excellent way to understand
how the components of an action situation and changes in these components,
even small changes, can make a difference in behavior and outcomes.
Experiments provide very good examples of how action situations
work. In the laboratory, the researcher carefully establishes the specific
components of an action situation to be studied.
While it is never possible to establish a perfect experiment, extensive
methodological work has been undertaken since the pathbreaking work
of Vernon Smith (1982). Smith challenged social scientists to exhibit great
care in the design of experiments by having the payoffs closely tied to the
incentives of the decision environment. He further urged researchers to
ensure that the assumed structure was clear to the subjects by pretesting
instructions extensively. Smith helped to establish a tradition of paying
subjects well and closely related to the theoretical payoff structure being
investigated, instead of having the subjects participate in experiments as
part of required entry-level courses. Further, researchers have now developed
a real commitment to share experimental instructions and their data
with one another and to encourage replications in order to ensure that
there was nothing “unique” about one implementation of an experiment.
Still further, experimental studies are now being used to “test bed” competing
strategies for implementing new public policies (see Plott and Porter
1996).
In this chapter, I will provide a brief overview of two action situations
that have been studied extensively in experimental laboratories in many
countries. The first is the Trust Game, similar to the Snatch Game discussed
in chapter 2. The second is a Commons Dilemma, which we will
discuss again in later chapters. I do so for three reasons: (1) to allow the
reader who is not familiar with the concept of an action situation or a
game to see how this concept has been operationalized; (2) to illustrate
how small changes in the structure of an action situation can lead to big
differences in outcomes; and (3) to illustrate how the findings from experi-
  
C HA PT ER TH RE E70
ments challenge the presumption that all participants use the same internal
form of rationality to make decisions in all settings. In light of these
findings, in chapter 4 we will examine the puzzle facing institutional analysts
of how to model individual behavior.
Both of the experiments discussed in this chapter are social dilemmas.
In all social dilemmas, individuals face short-term incentives that, if followed,
would lead them all to be worse off than feasibly attainable outcomes.
The first action situation to be discussed is the Trust Game. Subjects
in a laboratory experiment are assigned cash or tokens that have
varying values to be converted into cash immediately after an experiment
is concluded. The Trust Game has been studied extensively (E. Ostrom
and Walker 2003). Behavior in the laboratory is not fully consistent with
what is predicted by noncooperative game theory when monetary payoffs
are assumed to be monotonically related to the internal utility values that
a subject assigns to outcomes.
The second type of experiment is of appropriation (harvesting) from a
common-pool resource—a Commons Dilemma. Many of the experiments
described in the second section of this chapter have been conducted by
colleagues at Indiana University as well as being replicated by scholars at
other universities. Common-pool resource situations will help the reader
think about more complex N-person situations that are closely related to
specific field settings including inshore fisheries, irrigation systems, forest
resources, and groundwater basins. While behavior in the laboratory is
broadly consistent with noncooperative, game-theoretic predictions in a
baseline common-pool resource experiment, interactions and outcomes
are not consistent with this prediction when face-to-face communication
is allowed or when subjects can use costly punishment against one another.
There is, thus, a lot to learn about operationalizing action situations
from experimental work. Given the multiplicity of findings that are not
consistent with conventional game-theoretical analysis, we will be setting
the stage for an extended discussion of “animating institutional analysis”
in chapter 4.
The Trust Game in the Experimental Laboratory
In an elegantly simple, two-person game, Berg, Dickhaut, and McCabe
(1995) designed a game that enables researchers to focus in clearly on
factors that affect the likelihood that an individual will take a costly action
because she places trust in a second individual. As we discussed in
chapter 2, unless the first household in the Snatch Game trusted that the
second household would exchange goods rather than snatch Household
1’s goods, no positive transaction would occur. There are many such situ-
  
S TU DY IN G A CT IO N S IT UATI ON S I N T HE LA B 71
Figure 3.1 The Trust Game.
ations that individuals face in all aspects of life. The most simple baseline
game is composed of the following elements (see figure 3.1):
1. Participants. Two subjects play the game.
2. Positions. The two positions are the Investor and the Trustee.
3. Actions. The Investor is given X at the beginning of the game. The Investor
then decides to keep all of X, to give T to the Trustee and to keep X − T, or
to give all of it (T = X) to the Trustee. The Trustee then decides how much of
the funds—Y—to return to the Investor.
4. Outcomes. The outcomes are the size of the funds allocated to the two
players in light of the decisions they have made.
5. Action-outcome linkages. The amount invested in the Trustee earns a rate
of return (supplied by the experimenter) of 1 + r.
6. Information. Both players are told the full tree of possibilities and that
their own identity will remain anonymous to the other player and to the experimenter
(double-blind).
7. Potential payoffs. The payoffs are affected by the rate of return (1 + r). r
has frequently been operationalized as 2, so that the amount that the Investor
sends to the Trustee is tripled. The payoff to the Investor is (X − T) + Y. The
payoff for the Trustee is (1 + r) T − Y, assuming the Investor sent something in
the first place or zero otherwise.1
T can then be used as a measure of trust and
(1 + r) T − Y for T > 0 as a measure of trustworthiness. The original payoff rates
of Berg and colleagues were X = $10.00, r = 2. Thus, if the Investor allocated all
of the funds to the Trustee, the Trustee had $30.00 to keep or share. Beyond
the objective pecuniary payoffs, it is also possible that the individual assigns a
utility to the objective, external payoffs that increases or decreases the value of
the objective payoff to the individual or assigns utility to the objective values
that the other player receives.
A self-interested Trustee wanting to maximize pecuniary returns would
return zero to the Investor. If the Investor expects this, no funds will be
  
C HA PT ER TH RE E72
invested in the first place. Similar to the predicted outcome for the Snatch
Game, the predicted outcome using noncooperative game theory and assuming
that individuals attempt to maximize monetary returns is zero
invested. The baseline game is barren of many of the social factors that
are thought to affect trust. The players are strangers and do not even see
one another. There is no way that they can establish a link to one another
through promise-giving or the like. The Trust Game is similar in structure
to a sequential Prisoner’s Dilemma. The baseline game represents a situation
requiring trust in about as pristine a form as one can imagine.
Using a one-shot decision setting with double-blind experimental procedures
to ensure complete anonymity, Berg, Dickhaut, and McCabe
found that 30 of 32 subjects in the position of Investor sent money to the
Trustee ($5.16 on average). Of the 30 subjects in the position of Trustee,
18 returned more than $1.00 ($4.66 on average) and one-third of them
sent more funds to the Investor than they received. On average, those
Investors who sent $5.00 or more received an average return in excess of
the amount they sent. It was those Investors sending less than $5.00 who
received a negative net-average return. In other words, on average, those
Investors who trusted their counterpart the most were the ones who left
the game with more wealth than those who were less trusting.
This experimental design of the Trust Game is simple, but captures the
essence of trusting and reciprocal behavior so effectively that it has been
replicated and extended in many diverse settings and countries (see Cook
and Cooper 2003). The initial Berg, Dickhaut, and McCabe experiment
can be thought of as examining the level of trust and reciprocity among
relatively equal players. All of the subjects were students at the University
of Minnesota. Which subjects were chosen as the Investor or the Trustee
was determined randomly.
Ernst Fehr and colleagues at the University of Zurich have undertaken
a related set of experiments that shifted the framing and structure of the
situation from one involving participants who related to one another in
a horizontal manner to one that self-consciously involved hierarchy involving
either “employers or buyers” and “workers or sellers” in a vertical
relationship (see Fehr, Kirchsteiger, and Riedl 1993; Fehr and Falk
1999; Fehr and Ga¨chter 1998). In one of the Swiss experiments (Fehr,
Kirchsteiger, and Riedl 1993), the experimenters randomly assigned subjects
to two groups placed in separate rooms without capacity to see or
talk directly with one another (as did Berg and colleagues). For all rounds
of the experiment, one set of subjects was told they were “Workers” and
would be paid a wage if employed, and another set of subjects was told
they were “Employers.” There were always more Workers than Employers
so that not all of them would be “hired.” All were informed about the
specific wages that could be offered and the range of effort that a Worker
  
S TU DY IN G A CT IO N S IT UATI ON S I N T HE LA B 73
could expend at a monetary cost. The Employer’s income depended on
the effort expended by their Workers.
Employers made the first move in each of the twelve periods of an experimental
session by making contract offers of wage rates. The size of
the payoff to the Employer depended on the effort that a Worker expended
once he was hired with a specific wage. Since there were more
Workers than Employers, however, several Workers would remain unemployed
and receive only a set fee for showing up at the experiment. Thus
the Workers were at an initial disadvantage and might expect to receive
a minimal wage offer.
In the second stage, the experimental Workers who had accepted a
wage offer in a period made a decision regarding the level of effort they
would return to their Employers, ranging from a minimum of 0.1 to a
maximum of 1.0 with a reduction in their payoffs dependent on the level
of effort chosen. Any level of effort above 0.1 would be a “return gift”
back to the Employer, at a cost to the Employer that was much less than
the benefit that the Employer would receive. For a wage of 60, for example,
increasing effort by 0.1 would increase the cost to the Worker by 2
units at most, but would increase the return to the Employer by 6 units.
The anonymity of all subjects was assured by the experimenters.
The game-theoretical prediction for this experiment is similar to that
for the Berg, Dickhaut, and McCabe setting. The Employer should offer
a minimum wage since the person in this position can expect that the
Worker will return minimal effort (Fehr, Kirchsteiger, and Riedl 1993,
443). On the other hand, subjects in both positions lose potential income
if they follow the theoretically “rational” thing to do (as is the case for
Investors and Trustees in the Berg, Dickhaut, and McCabe experiment).
Fehr and colleagues repeatedly found, however, that subjects in the position
of Employer offered substantially more than the theoretical minimum
expected and that Workers reciprocated by allocating higher than minimal
costly effort. In this and a replication in a more challenging experiment,
where a double auction determined wages (Fehr and Falk 1999),
average effort was a positive function of the level of wage offered and far
above the predicted level.
Kenneth Koford (2003) has replicated both the original Berg, Dickhaut,
and McCabe horizontal experiments as well as some of the vertical (Employer-Worker)
experiments run by Fehr and colleagues, with students
recruited from Sofia University and Varna Economics University, both in
Bulgaria. This enabled Koford to directly examine the impact of the relative
horizontal or vertical standing of the two positions in the experiment.
In the horizontal experiment, 44 out of 47 of the subjects in the position
of Investor sent some money, which was quite similar to the 30 out of 32
  
C HA PT ER TH RE E74
Investors in Berg, Dickhaut, and McCabe’s experiment. The distribution
of amounts sent was also quite similar.
In the vertical experiment, Koford found a different pattern to that of
Fehr and Falk (1999). The effort levels of the subjects in the position of
Worker were much lower: “16 of the 41 effort levels were the minimum,
.1. The mean effort is just .368” (Koford 2003, 17). Further, the wage
level was not associated with the level of effort. Thus, in the horizontal
design, Koford found the same level of trust and reciprocity as the U.S.
and Swiss experiments, but not in the vertical designs.
Koford explains the intriguing difference in findings by drawing on the
business and cultural traditions and history in Bulgaria and the Balkan
countries more generally. Given the long history of exploitative Ottoman
and then communist rulers, government provided no basic security or
trust for a people. Folks had to trust others in similar circumstances.
Those in business in Bulgaria report extensive fraud and difficulties with
trusting others, however (see Koford and Miller 1995). Students are more
likely to engage in a general trust relationship with one another.
It seems that under specific conditions Bulgarian students trust as much as
American students, and very likely more. This may be due to the focus upon
“students,” who feel that they should share a principle of solidarity in Bulgaria.
Then, when they are divided into Workers and Employers, this solidarity disappears.
Discussion with students suggested that if the horizontal experiment
were run with the other group being significantly different—say, ordinary
Workers, or a different ethnic group . . . the level of trust would be considerably
lower. (Koford 2003, 21)
The horizontal Trust Game has also been replicated by Buchan, Croson,
and Johnson (1999) in another cross-country design intended to assess
the impact of the socioeconomic setting of participants on behavior.
They conducted the Trust Game with 188 subjects from China, Japan,
Korea, and the United States. They find no pure country effects in terms
of the amount sent by subjects in the role of the Investor (on average, 67
percent of the endowment) or in the amount returned (on average, 31
percent of the amount received).
In addition to examining country effects, the researchers examined several
other contextual variables including: the cultural beliefs of the participant
(based on questionnaire data where questions relate to an individual’s
attitude toward group versus individual outcomes), social distance
(manipulated experimentally), and communication (manipulated experimentally).
Subjects who showed a greater orientation toward group outcomes
sent more funds as both an Investor and as Trustee. The opportunity
to communicate information about one’s self and learn something
about the other person with whom a subject was paired also had a positive
  
S TU DY IN G A CT IO N S IT UATI ON S I N T HE LA B 75
effect on amounts sent and returned. Buchan and colleagues found that
“trusters prosper.” In other words, “subjects who sent above average
amounts to their partners, took home greater wealth than did subjects
who sent only average amounts or less” (Buchan, Croson, and Johnson
1999, 22).2
In a follow-up study to examine the impact of enabling the Trustee to
build a reputation, Dickhaut et al. (1997) added a publicly announced
second round to the basic structure of the base experiment. Both participants
played in both rounds. According to standard, noncooperative
game theory using backward induction, this second round should make
no difference to the behavior of individuals in the first round. Dickaut
and colleagues wanted to ascertain if Trustees would act even more trustworthy
than those who had participated in the first study in order to
assure Investors that they could be trusted.
In the first round, they found that 10 of the 23 Investors sent the maximum
sum ($10.00) to the Trustee and that none of them sent zero. Trustees
exhibited higher levels of reciprocity in the first round than had been
exhibited in the baseline study. Twenty of them returned more than their
counterpart had sent to them, leading to a positive-sum outcome for all
involved. The findings for the second and final round, however, followed
a different pattern. Nineteen out of 20 of the Investors who had received
positive returns in round one made a positive investment again in round
two, and all three of the Investors who had received a negative return in
the first round sent zero to the Trustee in the second round. The big difference
was that only 7 of the 19 Investors received a positive net return in
round two. The reciprocity that had been exhibited in round one was
substantially reduced in the second, and known-to-be final, round.
Gu¨th, Ockenfels, and Wendel (1997) simplified the experimental conditions
so that Investors had similar dichotomous options as Household 1
in the Snatch Game—to trust or not trust the Trustee. The Trustee could
then either “exploit” the Investor by choosing an option that paid the
Trustee a high payoff and gave the Investor very little, or the Trustee could
reciprocate the trust by choosing an option that led to lower but equal
payoffs for both. Gu¨th and colleagues found that the majority of the subjects
in the first position (21 out of 28) extended trust, but that their
trusting action was not reciprocated by a majority of those in the second
position. On the other hand, in a replication of this experimental structure,
Kirchler, Fehr, and Evans (1996) had results very close to those of
Berg, Dickhaut, and McCabe with high levels of trust and reciprocity
exhibited (see also Dickhaut et al. 1997).3
Gu¨th, Ockenfels, and Wendel
(1997) also experimented with a further modification of the Trust Game.
In this experiment, they had subjects experience the game once and receive
the payoffs from that round. Then they asked the subjects to bid for the
  
C HA PT ER TH RE E76
role of Investor and Trustee. Very low levels of trust were extended when
played by those who bid for the Investor position (and low levels of reciprocity
were extended by those who were trusted).
In an experiment that explores the impact of changes in the information
exchanged as well as threatened sanctions on outcomes of a Trust Game,
Fehr and Rockenbach (2003) obtained some fascinating results (see also
Fehr, Ga¨chter, and Kirchsteiger 1996 and Fehr and Ga¨chter forthcoming
for related experiments). In one series of experiments—which they call
the “trust condition”—both the Investor and the Trustee receive ten
money units (MUs). As in the above games, the Investor may either send
zero or some number between 1 and 10 MUs to the Trustee. Whatever
amount is sent is tripled. At the time of sending funds, the Investor has
to indicate a “desired back-transfer” that can range from zero to the full
amount of the tripled transfer. The information about the amount sent
and the desired back-transfer is given to the Trustee, who then decides
how much, if any, to send back. The only difference between this experiment
and the Berg, Dickhaut, and McCabe experiment is the information
given to the Trustee about the amount the investor expects back.
In a second condition—which they call an “incentive condition”—the
Investor is given a choice between indicating whether he or she has chosen
a fining option that will deduct four MUs from the Trustee if the Investor
does not receive at least as much as the specified desired level. The Investor
can overtly waive this opportunity to have these funds deducted if the
returned level is not as desired. If chosen by the Investor, the deduction is
subtracted from the Trustee’s payoff, and the Investor receives whatever
funds the Trustee decides to send back without contributing to the deduction.
Fehr and Rockenbach explored this incentive option in order to examine
the proposition that a threatened sanction would increase the level
of funds returned by the Trustee.
They found that Trustees sent substantial returns in all conditions. Like
Berg and colleagues, they found that the amount of funds returned were
positively associated with the amount of money invested in the first place.
Nineteen of the 24 Trustees in the trust condition (79 percent) paid back
more than zero, and 19 of the 45 subjects in the Trustee position in the
incentive condition (42 percent) paid back more than zero (Fehr and
Rockenbach 2003, 138). As shown in figure 3.2, the highest return occurred
when the Investor refrained from imposing a fine in the incentive
condition. The lowest level of return occurred when the Investor indicated
a high desired back-transfer and a fine for not meeting this level. Fehr and
Rockenbach explained both findings as the result of what they refer to as
“strong reciprocity.” “First, refraining from the threat of fining, although
the threat is available, could itself be perceived as a fair act, which induces
  
S TU DY IN G A CT IO N S IT UATI ON S I N T HE LA B 77
Figure 3.2 Trust and reciprocity under three experimental conditions. Source:
Fehr and Rockenbach 2003, 138, and data provided by Ernst Fehr.
the trustees to increase their cooperation. Second, attempts to use the
sanction to enforce an unfair distribution of income may be perceived as
hostile acts, inducing the trustees to reduce cooperation” (2003, 139–
40). Thus, the threat of a fine, which is frequently thought to be a major
technique for solving problems related to trust, backfired. Instead of enhancing
the level of reciprocity shown by Trustees, an externally established
fine that is endogenously used as a threat by the Investor reduces
reciprocity rather than increasing it. This is consistent with the theoretical
and experimental work of Bruno Frey (1994, 1997a) on external sanctions
crowding out reciprocity. We will return to the question of how
sanctions are arrived at, and their impact, when we discuss sanctions
again in the context of the common-pool resource experiments as well as
in later sections of this book.
The findings from these rich experimental studies of the Trust Game
are consistent with empirical studies of trust in other settings (Gambetta
1988; Rothstein 2005; Delhey and Newton 2003). They also provide important
insights for an institutional analyst. First of all, it is relatively
easy to see how Berg, Dickhaut, and McCabe initially created an action
situation in an experimental lab. We can also observe how small changes
  
C HA PT ER TH RE E78
in these working parts—including: (1) the relative standing of the participants,
(2) the number of rounds, (3) the range of actions authorized, (4)
the benefits and costs of actions and outcomes, (5) the diverse ways of
choosing subjects to hold a position (e.g., random assignment or an auction),
or (6) the diverse sanctioning mechanisms—affected interactions
and outcomes. Hopefully, the concept of an action situation is now a little
clearer for the reader.
Many readers, however, will be surprised at the high levels of trust
exhibited in the first experiment and its replications in multiple cultures.
The game-theoretic prediction for this game is very stark and clear. Because
the Trustee is expected by a rational player not to return anything
to an Investor, the Investor is not expected to trust an unknown stranger
by sending any funds. Like all social dilemmas, this prediction leaves all
participants worse off than they could have achieved. The multiple
replications in which a substantial level of trust is exhibited raises serious
questions about the universal validity of relying entirely on the rational
choice model of the individual (a topic to be explored in some depth in
chapter 4).4
Still further questions are raised about the capability of the classic
model of self-interested rationality to explain the adverse impact of externally
imposed threats of sanctions on the willingness of Trustees to reciprocate
Trust. When faced with a threat of sanction, if the Trustee did not
return the level of funds specified by the Investor, many Trustees did not
return the funds specified—thus, levels of reciprocity were reduced by the
externally imposed threat of a sanction. This empirical result challenges
not only the model of the individual that is widely used but also the standard
recommendation that external sanctions are the best way to solve
social dilemmas that lie at the heart of many types of basic economic,
social, and political problems. We will discuss the core difference in outcomes
achieved between externally imposed sanctions and sanctions
agreed upon by those involved in the next section of this chapter.
A Commons Dilemma in the Experimental Laboratory
The Trust Game characterizes a wide diversity of settings where one person
has to trust someone else and that person has to reciprocate in order
for both to be better off. This game seems too simple for some scholars.
“All one needs to do is create a form of contract law,” they say, “in order
for the Investor to bind the Trustee to a contract so as to return the investment.”
Then they argue, you would not have to depend on reciprocity
but rather on the legal system. One of the expected benefits of creating
a relevant legal structure is to use sanctions against those who are not
  
S TU DY IN G A CT IO N S IT UATI ON S I N T HE LA B 79
trustworthy. As we have seen, however, from the experiments using sanctions,
the way that legal structures are implemented can make a big difference.
Externally imposed sanctions can lead to resentment rather than
guilt and adversely affect the willingness to reciprocate trust.
Creating contract law or property rights is more difficult than it appears
on the surface (as we will discuss in chapter 8, and as many advisers
have learned in Eastern Europe and the former Soviet Union—see, for
example, de Soto 2000). Many contracts in Western economies are undertaken
based on trust and reciprocity (Arrow 1974) rather than drawing
on the legal system. Further, the Fehr and Rockenbach experiments demonstrate
that a sanctioning system imposed without agreement of the participants
may reduce reciprocity rather than increase it. Hopefully, the
reader can by now appreciate both how to operationalize an action situation
as well as learning about the substantial findings of experimental
research on trust. Further theoretical developments on trust have now
been achieved as a result of this extensive research (see Bacharach and
Gambetta 2001; Ahn et al. 2003; McCabe and Smith 2003; E. Ostrom
and Walker 2003).
We will now look at a second illustration of an action situation in the
laboratory that is inherently far more complex than the Trust Game—
a common-pool resource situation. When the users of a common-pool
resource are tempted to overuse the resource, the structure of the situation
is a social dilemma. Social dilemmas are pervasive in social life, and
proposed solutions to these dilemmas have occupied all great political
philosophers including Aristotle, Hobbes, and Hume. There are many
differently structured social dilemmas, but they all are characterized
by a situation where everyone is tempted to take one action but all will
be better off if all (or most of them) take another action. Studying how
subjects behave in this type of social dilemma helps us understand more
general questions of relevance across the social sciences than simply the
study of natural resources. We will also return to discuss the commonpool
resource setting in chapters 8 and 9 to examine how actual users of
these kinds of resources create their own rules to deal with the problems
of overuse.
The Definition of a Common-Pool Resource
A common-pool resource, such as a lake, an ocean, an irrigation system,
a fishing ground, a forest, the Internet, or the stratosphere, is a natural or
man-made resource from which it is difficult to exclude or limit users once
the resource is provided by nature or produced by humans (E. Ostrom,
Gardner, and Walker 1994). One person’s consumption of resource units,
such as water, fish, or trees, removes those units from what is available
  
C HA PT ER TH RE E80
to others. Thus, the trees or fish harvested by one user are no longer available
for others. As discussed in chapter 1, the difficulty of excluding beneficiaries
is a characteristic that common-pool resources share with public
goods. Further, the subtractability of resource units from a common-pool
resource is an attribute shared with private goods. In the rest of this chapter,
I will focus primarily on renewable natural resources as exemplars of
common-pool resources, but the theoretical arguments are relevant to
man-made common-pool resources, such as the Internet, as well.
When the resource units produced by a common-pool resource have a
high value and institutional rules do not restrict the way resource units
are appropriated (an open-access situation), individuals face strong incentives
to appropriate more and more resource units leading eventually to
congestion, overuse, and even the destruction of the resource itself. Because
of the difficulty of excluding beneficiaries, the free-rider problem is
a potential threat to efforts to reduce appropriation and improve the longterm
outcomes achieved from the use of a common-pool resource. If some
individuals reduce their appropriation levels, the benefits they generate
are shared with others whether the others also cut back on their appropriation
or not. Some individuals may free-ride on the costly actions of others
unless ways are found to reduce free-riding as an attractive strategy. When
free-riding is a major problem, those who would willingly reduce their
own appropriations if others did are unwilling to make a sacrifice for the
benefit of a large number of free riders.
Consequently, one of the important problems facing the joint users of
a common-pool resource is known as the “Commons Dilemma,” given
the potential incentives in all jointly used common-pool resources for individuals
to appropriate more resource units when acting independently
than they would if they could find some way of coordinating their appropriation
activities. Joint users of a common-pool resource often face many
other problems including assignment problems, technological externality
problems, provision problems, and maintenance problems (E. Ostrom,
Gardner, and Walker 1994). And, the specific character of each of these
problems differs substantially from one resource to the next. In this
chapter, I focus on the problem of controlling appropriation since this is
what most policy analysts associate with “the tragedy of the commons”
(G. Hardin 1968; Dietz, Ostrom, and Stern 2003).
A Baseline Appropriation Situation
We need to start with a static, baseline situation that is as simple as
we can specify it without losing crucial aspects of the problems that
real appropriators face in the field. This static, baseline situation is composed
of:
  
S TU DY IN G A CT IO N S IT UATI ON S I N T HE LA B 81
1. Participants. A set of n symmetric subjects who do not have any outside
relationships with one another.
2. Positions. No differentiation exists in the positions these subjects hold
relevant to the common-pool resource. In other words, there is only one position
of appropriator.
3. Actions. Appropriators must decide how to allocate tokens assigned to
them in each time period. Basically, one can think of these appropriators as
being “endowed” with a total set of assets, e, which they are free to allocate in
any proportion during each time period to two activities. In a field setting one
can think that every day, each appropriator must decide between spending time
trying to harvest resource units from the common-pool resource or using their
time to earn money in an outside option, such as working in a local factory. To
simplify the problem, we posit that all appropriators have the same endowment
(just as we all have only twenty-four hours per day), and face the same outside
opportunity. Thus, they have to decide how much of their endowment to devote
each round to appropriation from the common-pool resource or in gaining
returns from an outside option.
4. Outcomes. The actions they take affect the amount of resource units that
can be appropriated from the common-pool resource or returns earned from
the outside option.
5. Action-outcome linkages. The function maps the actions of all of the appropriators
given the biophysical structure of the resource itself onto outcomes.
While these functions are frequently stochastic in field settings and affected by
many variables in addition to the actions of individuals, we will consider only
determinant functions of appropriation actions in the baseline setting. The
wage function simply multiplies the amount of time allocated to it by whatever
is the standard wage. The resource function is a concave function, F, which
depends on the number of assets, xi, which is a fraction of e allocated to appropriation.
Initially, the sum of all of the individuals’ actions, ∑xi, generates better
outcomes than the safe investment in wage labor. If the appropriators decide
to allocate a sufficiently large number of their available assets, the outcome
they receive is less than their best alternative. In other words, allocating too
many assets to the common-pool resource is counterproductive (see Gordon
1954; Scott 1955).
6. Information. As an initial information condition (because of the instructions
carefully given to all subjects), we assume that appropriators know the
shape of the function linking actions to outcomes and know that they are symmetric
in assets and opportunities. Information about outcomes is generated
after each decision round is completed. Appropriators may not communicate
with one another. It is assumed that each appropriator will assume that all other
appropriators are rational actors and will adopt the “best response” to their
own actions. The best response function should lead all appropriators to overharvest
from the resource.
  
C HA PT ER TH RE E82
7. Potential payoffs. Payoff functions specify the value of the wage rate and
the value of the resource units obtained from the common-pool resource. Specifically,
the payoff to an appropriator is given by:
we if xi = 0
w (e − xi ) + (xi / ∑xi ) F (∑xi ) if xi > 0. (1)
Basically, if appropriators put all of the assets into their outside option,
they receive a certain monetary return equal to the amount of their endowment
times an unchanging rate of return (w).5
If appropriators put
some of their endowed assets into the outside option and some into the
common-pool resource, they get part of their return from the outside option
and the rest from their proportional investment in the common-pool
resource times the total output of the common-pool resource as determined
by function F.
Predicted Outcomes for a Commons Dilemma
in the Laboratory
In a series of laboratory experiments conducted at Indiana University, we
thought it crucial to examine behavior in an appropriation situation with
a nonlinear transformation (action-outcome) function and a sufficient
number of players that knowledge of outcomes did not automatically provide
information about each player’s actions. In this chapter, I can only
briefly discuss the results of these experiments. All procedures and specifications
are thoroughly documented in E. Ostrom, Gardner, and Walker
1994 and in journal articles cited therein.
In the baseline experiments, we utilized the following equation for the
transformation function, F, measured in units of output (outcome units):
23 (∑xi ) − 25 (∑xi )2
. (2)
Instead of asking subjects to pretend they were fishing or harvesting timber,
the situation was described as involving a choice between investing
in either of two markets having the structure as specified above. All eight
subjects were assigned either 10 or 25 tokens as their endowment in each
round of play. Their outside opportunity was valued at $.05 per token.
They earned $.01 on each outcome unit they received from investing tokens
in the common-pool resource. Subjects were informed that they
would participate in an experiment that would last no more than two
hours. The number of rounds in each experiment varied between twenty
and thirty rounds. In addition to being told the payoff function specifi-
  
S TU DY IN G A CT IO N S IT UATI ON S I N T HE LA B 83
cally, subjects were provided with look-up tables that eased their task of
determining outcomes depending on their own and others’ decisions.
With these specifications, the predicted outcome for a finitely repeated
game where subjects are not discounting the future and each subject is
assumed to be maximizing monetary returns is for each subject to invest
8 tokens in the common-pool resource for a total of 64 tokens. By design,
the prediction is the same for both endowment levels. At this level of
investment, they would each earn $.66 per round in the 10-token experiments
and $.70 per round in the 25-token experiments (players were paid
one-half of their computer returns in the 25-token experiments to keep
the payoffs roughly similar). The players could, however, earn considerably
more if the total number of tokens invested was 36 tokens (rather
than 64 tokens) in the common-pool resource. This optimal level of investment
would earn each subject $.91 per round in the 10-token experiment
and $.83 per round in the 25-token experiment. The baseline experiment
is an example of a commons dilemma in which the game-theoretic,
predicted outcome involves substantial overuse of a common-pool resource,
while a much better outcome could be reached if subjects were to
lower their joint use.
Behavior in a Sparse Experimental N-Person, Repeated
Commons Dilemma
Subjects interacting in baseline experiments substantially overinvested as
predicted. Subjects in the 10-token experiments achieved, on average, 37
percent of the maximum earnings from the common-pool experiment
available to them, while subjects in the 25-token experiments received −3
percent (E. Ostrom, Gardner, and Walker 1994, 116). At the individual
level, however, subjects rarely invested 8 tokens, which is the predicted
level of investment at equilibrium. Instead, all experiments generated an
unpredicted and strong pulsing pattern in which individuals appear to
increase their investments in the common-pool resource until there is a
strong reduction in yield, at which time they tend to reduce their investments
leading to an increase in yields. The pattern is repeated over time.
At an aggregate level, behavior approximates the predicted Nash equilibrium
in the 10-token experiments. Outcomes are far lower than predicted
in the early rounds of the 25-token experiment and only begin to approach
the predicted level in later rounds. No game-theoretical explanation
yet exists for the pulsing pattern or the substantial difference between
the 10-token and the 25-token experiments.
In response to postexperiment questions, subjects explained that they
were using several heuristics. One of the heuristics they used was to invest
more in the common-pool resource whenever the rate of return on the
  
C HA PT ER TH RE E84
previous round was above $.05 (what they could earn in their next best
alternative) and less if the return was below $.05. In the 10-token experiment,
some players invested all 10 of their tokens whenever the rate of
return in the prior round was above $.05. With such heuristic strategies,
equilibrium is never reached at the individual level. Each player is revising
his or her response to the results obtained in the most recent round, creating
considerable turbulence in the outcomes jointly reached.
Replication in Agent-based Models
An extremely interesting follow-up study was undertaken by Peter Deadman
(1997) in which artificial agents were programmed to use a variety
of heuristics similar to those used by the human subjects in these commonpool
resource (CPR) experiments and to interact in a simulated environment
that exactly replicated the baseline experiments. Deadman found
that the specific results obtained in any series of runs depended on the
particular heuristic (or mix of heuristics) programmed, but the artificial
agents did consistently produce the same kind of pulsing returns and the
consistent difference between 10-token and 25-token environments was
also observed. Deadman describes his results:
As in CPR experiments, the group performance for the simulation follows an
oscillating pattern in which high performance leads to over investment in the
CPR and the resultant drop in performance causes a reduction in group-wide
investment in the CPR. . . . Still more interesting is the observation that the
simulations perform similarly to subjects in laboratory experiments in terms of
average performance over time. At the ten token endowment, the simulations
perform near the Nash equilibria over time. At the 25 token endowment, the
simulations perform near zero percent of optimum over time. (Deadman 1997,
175–76)
Jager and Janssen (2003) also developed a multiagent model using the
consumat framework derived from social psychology (Jager, Janssen, and
Viek 2001). They thought that they could replicate the data from the
baseline commons dilemma experiments described above with an assumption
that individual subjects differed in regard to their Social Value Orientations
(SVOs). In their first series of simulations, they were indeed able,
as Deadman had done independently, to explain the aggregated pattern
of appropriation behavior. Jager and Janssen were not as successful in
replicating individual appropriation decisions (from the experimental
data) as they had been in simulating aggregate outcomes. Jager and Janssen
(2003) suggested that the cognitive processes that subjects use appear
to be important in behavior. Those with low aspiration levels may lock
  
S TU DY IN G A CT IO N S IT UATI ON S I N T HE LA B 85
into a habitual response too soon. Further, those agents “conforming to
the Homo psychologius have a better performance than the Homo economicus
in approximating the empirical data” (98).
Structural Changes in the Laboratory
In addition to the baseline experiments, we and many other researchers
have explored how changes in the structure of the action situation affect
outcomes. These changes are operationalized in the set of instructions
given to subjects and in the procedures adopted within the experiment.
The first structural change is related to the information component of
an action situation. Instead of forbidding communication among subjects,
as in the baseline experiments, subjects are now authorized to communicate
with one another on a face-to-face basis in a group setting before
returning to their own enclosed terminals to make their private decisions.
This introduction of an opportunity for “cheap talk” in a social dilemma,
where agreements are not enforced by an external authority, is viewed
within the context of noncooperative game theory as irrelevant. The same
outcome is predicted as in the baseline experiment.
In a second set of experiments, colleagues also explored whether cheap
talk would enable individuals who acquired heterogeneous assets to
achieve better outcomes than predicted by theory. In a third series of experiments,
the payoff component was changed to allow subjects to sanction
one another at a cost to themselves. Since using this option produces
a benefit for all at a cost to the individual, the game-theoretic prediction
is that no one will choose the costly sanctioning option. Fourth, the authority
rule was changed to allow subjects to covenant with one another
to determine their investment levels and to adopt a sanctioning system
if they wished. Again, the predicted outcome is the same. In all four of
these appropriation experiments, however, subjects demonstrate their
willingness and ability to search out and adopt better outcomes than those
predicted.
Face-to-Face Communication
In the repeated communication experiments, subjects first made ten
rounds of decisions in the context of the baseline appropriation situation
described above. After the tenth round, subjects listened to an announcement
that told them they would have an open group discussion before
each of the next rounds of the experiment. The subjects left their terminals
  
C HA PT ER TH RE E86
and sat in a circle facing one another. After each discussion period, they
returned to their terminals to enter their anonymous decisions. Subjects
used face-to-face communication to discuss together what strategy would
gain them the best outcomes and to agree on what everyone should invest
in the subsequent rounds. After each decision round, they learned what
their aggregate investments had been, but not the decisions of individual
players. Thus, they learned whether total investments were greater than
the total investments they had earlier agreed upon. While in many rounds,
subjects did exactly as they had promised one another they would do,
some defections did occur. If promises were not kept, subjects used this
information about the aggregate investment levels to castigate the unknown
participant who had not kept to the agreement.
This opportunity for repeated face-to-face communication was extremely
successful in increasing joint returns. In the 10-token experiments,
subjects obtained close to 100 percent of the maximum available
returns. There were only 19 instances out of 368 total opportunities (5
percent) where a subject invested more in the common-pool resource
than agreed upon (E. Ostrom, Gardner, and Walker 1994, 154). In the
25-token experiments, subjects also improved their overall performance.
The temptation to defect, however, was greater in the 25-token experiments.
Subjects in the 25-token baseline experiments had received total
returns that were slightly below zero, while in the communication experiments,
they obtained on average 62 percent of the maximum available
returns (with considerable variation across experiments). The defection
rate was 13 percent. Our conclusion in completing an analysis of these
experiments was:
Communication discussions went well beyond discovering what investments
would generate maximum yields. A striking aspect of the discussion rounds
was how rapidly subjects, who had not had an opportunity to establish a welldefined
community with strong internal norms, were able to devise their own
agreements and verbal punishments for those who broke those agreements. . . .
In many cases, statements like “some scumbucket is investing more than we
agreed upon” were a sufficient reproach to change defectors’ behavior. (E. Ostrom,
Gardner, and Walker 1994, 160)
That subjects had internalized norms regarding the importance of keeping
promises is evidenced by several of their behaviors. Simply promising to
cut back on their investments in the common-pool resource led most subjects
to change their investment pattern. Secondly, subjects were indignant
about evidence of investment levels higher than that promised and expressed
their anger openly. Third, those who broke their promise tended
to revert to the promised level after hearing the verbal tongue-lashing of
  
S TU DY IN G A CT IO N S IT UATI ON S I N T HE LA B 87
their colleagues (see A. Simon and Gorgura 2003 for an intensive analysis
of the recorded transcripts). The findings from these initial communication
experiments are consistent with a large number of studies of the impact
of face-to-face communication on the capacity of subjects to solve a
variety of social dilemma problems (see Bohnet and Frey 1998; E. Ostrom
and Walker 1991; Sally 1995 and literature cited therein).
Communication among Heterogeneous Players
Steven Hackett, Edella Schlager, and James Walker (1994) conducted a
series of commons dilemma experiments where they explored whether
communication could ameliorate the problems identified in field and experimental
settings related to heterogeneity among appropriators (R. Hardin
1982; R. Johnson and Libecap 1982; Libecap and Wiggins 1984; Isaac
and Walker 1988a, 1988b; Wiggins and Libecap 1987; Hackett 1992).
The task of agreeing to and sustaining agreements for efficient appropriation
from a common-pool resource is more difficult for heterogeneous
appropriators because of the distributional conflict associated with alternative
sharing rules. In heterogeneous settings, all appropriators may be
made better off by adopting a new rule; some will benefit more than others,
depending upon the sharing rule chosen. Consequently, appropriators
may fail to cooperate on the adoption of a sharing rule because they cannot
agree upon what would constitute a fair distribution of benefits produced
by cooperating.
In order to address appropriator heterogeneity, the Hackett, Schlager,
and Walker experimental design allows for two levels of input endowments.
One subset of appropriators has large endowments of tokens (24);
the other appropriators have small token endowments (8). Group allocations
to invest in the commons at the asymmetric Nash equilibrium are
greater than optimal, but not all potential returns dissipated.
In order for communication to enhance joint payoffs to a heterogeneous
set of subjects, they must agree on (1) the target level of group
allocations to the common-pool resource, and (2) a rule for allocating the
target input allocation across appropriators, and they must create (3) the
necessary “social capital” to attenuate cheating, since agreements are
nonbinding. The existence of heterogeneity in endowments and in historic
allocation levels has no effect on the first problem, but is likely to elicit
disagreement over the second problem. Disagreements then impair the
building of social capital—the third problem.
Subjects knew with certainty the total number of decision makers in
the group, their own token endowment and that of the others, the total
number of tokens in the group, the transformation function, and the num-
  
C HA PT ER TH RE E88
ber of decision rounds in the current treatment condition. After each
round, subjects were shown a display that recorded their payoffs in each
market for that round, total group token investments, and a total of their
cumulative profits for the experiment, but not the allocations made by
specific other players.
Subjects participated in two (consecutive) ten-round sequences of the
asymmetric game. In the first ten rounds, subjects were not allowed to
communicate, but face-to-face communication was allowed during
the second set of ten interactions. Prior to each ten-round treatment sequence,
four subjects were assigned the “large” token endowment, while
the other four subjects were assigned the “small” endowment. Two different
mechanisms were used for assigning these endowments: random and
auction (based on Gu¨th 1988). In the first method, large endowments
were assigned randomly prior to the ten decision rounds without communication,
and again prior to the ten decision rounds with communication.
A multiple-unit ascending price auction was used as the alternative mechanism
for assigning endowment positions because of its demand-revealing
characteristics. In particular, the price paid for the large endowment position
should theoretically correspond with the maximum value placed on
this position by the subject with the fourth highest valuation.
The opportunity to communicate led to a noticeable change in the
pattern of allocations. With the allocation rules agreed upon in communication
rounds, subjects concentrated near the optimal allocation of 56
tokens in total. In the random-assignment and communication condition,
individual allocations of 8 tokens represented the modal response (67
percent). In the auction and communication condition, however, Hackett,
Schlager, and Walker (1994) observed a spread of allocations clustered
between 6 to 10 tokens. For both treatment conditions with no communication,
they observe a level of rent accrual relatively close to that predicted
by the Nash equilibrium (49 percent). Thus, even in an environment of
extreme heterogeneity in subject endowments, communication remains a
powerful mechanism for promoting coordination, resulting in rents very
close to those observed in the homogeneous decision setting discussed
above.
Three follow-up experimental designs were conducted by Pamela
Schmitt, Kurtis Swope, and James Walker (2000). They used the same
baseline design as first described above with three variations in regard to
the information component of the situation. In all of their protocols for
the communication aspects of the experiment (rounds eleven through
twenty-five), six out of the eight players were invited to communicate with
one another in one location. In their first protocol, two of the remaining
“players” were computerized decision makers whose decisions were each
  
S TU DY IN G A CT IO N S IT UATI ON S I N T HE LA B 89
the result of a random draw of a number between four and twelve. The
other six, who could communicate with one another, were informed
about the constraints on the random draw for “Players 7 and 8.” In their
second protocol, two of the players were real subjects who had been separated
from the other six players. These real players did not face a constraint
on their decisions. In the third protocol, the two separated players
were constrained to invest between 4 and 12 tokens in the common-pool
resource. As in the baseline experiments, subjects were informed about
the aggregate investment of all eight players after each round.
In all three protocols, subjects substantially improved their overall efficiency
in the communication rounds as contrasted to the noncommunication
rounds. Limiting who could communicate to six out of the eight
players did, however, make a difference. The six communicating subjects
were never certain if the announced aggregated investment level reflected
higher investments by the two “noncommunicating” players or whether
some of the communicating groups did not follow their agreement. This
uncertainty affected the capacity of the communicating group to keep to
their own agreements. They could always blame the outsiders for any
major overinvestment.
Major differences in interactions and outcomes occurred across the
three protocols. The outsiders in Protocol 2 were the least constrained in
their decisions. The members of the communicating group had the most
difficulty in reaching agreements and following them in Protocol 2. The
six communicating subjects in Protocol 2 had a “scapegoat” they could
blame for high investment levels in rounds eleven through twenty-five. As
shown in table 3.1, the subjects in Protocol 2 were less likely to come to
an agreement in the first place, had a much higher deviation rate and size
of deviation when they did agree, and obtained lower payoffs than in the
earlier Ostrom, Gardner, and Walker experiments (discussed above and
labeled as OGW in table 3.1) or in Protocols 1 or 3. The problem of
imperfect monitoring was less severe in Protocols 1 and 3. Some subjects
among the communicating group were able to deviate without raising
suspicion of cheating in these protocols, but they made only small increases
over what the group had promised each other.
What this series of experiments found does have considerable implications
for those trying to achieve an agreement in the field not to overharvest
from a common-pool resource. The results provide evidence that communication
is less likely to be effective in preventing overharvesting in
Commons Dilemmas “in which a subset of appropriators either cannot
or will not participate in collective action” (Schmitt, Swope, and Walker
2000, 852). The lack of commitment by an outside group is not only a
source of additional investment but also gives “insiders” a scapegoat to
  
TABLE3.1
Summaryresultsacrossenvironments
Decisionenvironment
25tokensProtocol1Protocol2Protocol3
Repeatedcommunication(5experiments)(3experiments)(3experiments)
OGWGroupRandomGroupOutsideGroupOutside
Numberofsubjects8626262
Averagerents−2%−9.4%−12.3%28.5%
Rounds1–10
Averageinvestmentinmarket28.719.088.848.769.508.407.23
Averagepayoffs1.241.111.211.201.171.371.36
Averagerents73%68.5%6.1%80.4%
Averageinvestmentinmarket25.655.958.139.009.905.618.28
Averagepayoffs1.541.571.711.271.301.541.70
Rounds11–25
Agreementrate95%64%31.1%97.8%
Deviationrate13%15.3%50%15.9%
Averagedeviationsize3.32.757.832.52
Source:Schmitt,Swope,andWalker2000,849.
  
S TU DY IN G A CT IO N S IT UATI ON S I N T HE LA B 91
blame if their own harvests are higher than agreed upon. “The problem
becomes more severe when outsiders have less constraints on their overall
appropriation behavior and their ability to behave strategically” (852).
Sanctioning Experiments
Participants in field settings are frequently able to communicate with one
another on a face-to-face basis, at least from time to time, either in formally
constituted meetings or at social gatherings. In many field settings,
where the resource has been sustained over a long time, participants have
also devised a variety of formal or informal ways of monitoring and sanctioning
one another if rules are broken (discussed in chapters 8 and 9).
Engaging in costly monitoring and sanctioning behavior is, however, not
consistent with the theory of norm-free, complete rationality (Elster
1989a, 40–41). Thus, it was important to ascertain whether subjects in a
controlled setting would actually pay funds from their own earnings in
order to sanction the less cooperative behavior of other participants. The
short answer to this question is yes, they will.
All sanctioning experiments used the 25-token design since appropriation
levels had been much higher in this design. Subjects played ten rounds
of the baseline game modified so that the individual contributions in
each round were reported as well as the total outcomes. Subjects were
then told that in the subsequent rounds they would have an opportunity
to pay a fee in order to impose a fine on the payoffs received by another
player. The fees ranged in diverse experiments from $.05 to $.20 and
the fines from $.10 to $.80. In brief, the finding from this series of experiments
was that much more voluntary sanctioning occurs than the zero
level predicted.
Subjects react both to the initial cost of sanctioning and to the fee-tofine
relationships. They sanction more when the cost of sanctioning is less
and when the ratio of the fine to the fee is higher. Sanctioning is primarily
directed at those who invested more in the common-pool resource. A few
sanctions, however, appear to be a form of “blind revenge.” These were
fines made by subjects who had themselves been fined by unknown others
for their high levels of investment. In these few cases, the sanctioners
picked on those whose investments were lower than others, and thus were
suspected of being the ones who had previously sanctioned them (E. Ostrom,
Walker, and Gardner 1992).
In this set of experiments, subjects were able to increase their returns
from the common-pool resource modestly to 39 percent of maximum,
but when the costs of fees and fines were subtracted from the total, these
gains were wiped out. When subjects were given a single opportunity to
  
C HA PT ER TH RE E92
communicate prior to the implementation of sanctioning capabilities,
they were able to gain an average of 85 percent of the maximum payoffs
(69 percent when the costs of the fees and fines were subtracted).
Covenanting Experiments
In self-organized field settings, participants rarely impose sanctions on
one another that have been devised exogenously as was done in the above
sanctioning experiments and in the trust experiments, where the experimenters
assigned one position the authority to sanction the holder of the
other position. In the field, sanctions are much more likely to emerge
from an endogenous process of crafting their own rules, including the
punishments that should be imposed if these rules are broken. Spending
time and effort in a linked collective-choice situation designing rules creates
a public good for all of those involved. Crafting rules for an operational
situation is thus a second-level dilemma that theorists have argued
is no more likely to be solved than the original commons dilemma.
Noncooperative game theory predicts that participants will not undertake
such efforts. This is the foundation for the repeated recommendation
that rules must be imposed on participants by external authorities who
then assume official responsibility for monitoring and enforcing these
rules and are paid a salary for their work. Since self-organized rules are
found in many local common-pool resource situations, it appears that
participants frequently do design their own rules contrary to the theoretical
prediction. Few scholars are able to witness these processes, however,
in the field.
In order to observe what happens in these settings, subjects experienced
with baseline and sanctioning experiments were recalled and given an
opportunity to have a “convention” in the laboratory. In a face-to-face
discussion, subjects could decide whether or not they would like to have
access to a sanctioning mechanism like the one described above, how
much the fines and fees should be, and on the joint investment strategy
that they would like to adopt. All of the subjects in these groups were
endowed with 25 tokens in every round. Four out of six experimental
groups adopted a covenant in which they specified the number of tokens
they would invest and the level of fines to be imposed. The fines determined
by the participants ranged in size from $.10 to $1.00 (E. Ostrom,
Walker, and Gardner 1992).
The groups that crafted their own agreements were able to achieve an
average of 93 percent of the maximum in the periods after their
agreement. And, the defection rate for these experiments was only 4 percent.
The two groups that did not agree to their own covenant did not
fare as well. They averaged 56 percent of the maximum available returns
  
S TU DY IN G A CT IO N S IT UATI ON S I N T HE LA B 93
and faced a defection rate of 42 percent (E. Ostrom, Walker, and Gardner
1992). Consequently, those subjects who used an opportunity to covenant
with one another to agree on a joint strategy and choose their own level
of fines received very close to optimal results based entirely on their own
promises and their own willingness to monitor and sanction one another
when it was occasionally necessary (see Frohlich, Oppenheimer, and
Eavey 1987 for similar findings).
Replications and Extensions of Commons Dilemma Experiments
One of the great advantages of laboratory experiments, as we saw with
the Trust experiments in the first part of this chapter, is that they can be
replicated as well as modified by other researchers so that one can gain
ever greater confidence in the findings. One of the first replications was
conducted by Rocco and Warglien (1995), who found very similar outcomes
in the baseline, no-communication situation as well as in the faceto-face
communication settings. They were interested, in addition, in the
question of whether similar results would be obtained in an experimental
design where communication was not organized on a direct face-to-face
basis. They used identical structural variables of our earlier Commons
Dilemma design with and without face-to-face communication. Then,
they added a design with limited communication in a form of computerized
exchange. They replicated the substantial improvement in efficiency
that was earlier obtained in face-to-face communication. When the communication
was conducted electronically without a face-to-face discussion,
subjects did not increase their cooperation levels to the same extent.
Another very interesting series of replications and extensions was conducted
by Juan-Camilo Cardenas (2000, 2003) using field laboratories
set up in school buildings in rural Colombia rather than a computer-based
laboratory on a university campus. Cardenas initially invited over two
hundred villagers to participate in a series of common-pool resource experiments.
Several closely paralleled the ones conducted at Indiana University
and discussed above. Others extended the questions that could be
addressed. The villagers who Cardenas invited were actual users of local
forests for the extraction of firewood, natural fibers, and log timber as
well as local water resources. One of the basic questions he wanted to
pursue was whether experienced villagers who were heavily dependent
on local forests for wood products would behave in a manner broadly
consistent with that of undergraduate students at an American university.
The answer to this first question turned out to be positive. He wrote
his instructions in Spanish and in a manner that would be easily understood
by villagers. Instead of tokens—which are an easy medium for un-
  
C HA PT ER TH RE E94
dergraduates to understand—he asked villagers to decide on how many
months a year they would spend in the forest gathering wood products
as contrasted to using their time otherwise. Each villager had a copy of a
payoff table, which was the same as that of the other seven participants,
showing that as the number of months that each individual would spend
in the forest increased, she would gain more returns, but that the return
to all of them depended on their own keeping the harvesting time to a
very low level.
In the baseline, no-communication experiments, Cardenas found a similar
pattern as we had found with subjects from Indiana University. Villagers
substantially overinvested in the resource. While there was considerable
variation among groups, villagers on average achieved 57.7 percent
of their optimal return in the last three rounds of the baseline experiments
(Cardenas 2000, 316). The daily wage that most of the villagers could
earn at the time of the experiment was around 7,000 pesos (or around
U.S. $5.40 at the time). If they all invested at an optimal level, they would
earn around 12,900 pesos. They did earn around 7,884 pesos in these
experiments for the two or three hours they were involved in initial practice
sessions and the actual experiments themselves.
Face-to-face communication enabled them to increase efficiency on average
to 76.1 percent of optimal. Considerable variation among groups
existed, which Cardenas was able to explain using information about the
participants filled in after that experiment was completed. He found, for
example, that when most members of the group were already familiar
with common-pool resources such as the collective use of a mangrove,
they used the communication rounds more effectively than when most
members of the group were dependent primarily on their own assets. Cardenas
also found that “social distance and group inequality based on the
economic wealth of the people in the group seemed to constrain the effectiveness
of communication for this same sample of groups” (2000, 317;
see also Cardenas 2003).
Cardenas, Stranlund, and Willis (2000) report another fascinating extension.
In five experiments, the villagers were given a chance to communicate
after the initial ten rounds of the baseline condition. In five other
experiments, they were also told that a new regulation would go into
force that mandated that they should spend no more than the optimal
level of time in the forest each round (which in this case was one month
per villager). They were also told that there would be a 50 percent chance
that conformance to the rule would be monitored each round. The experimenter
rolled a die in front of the subjects each round to determine
whether an inspection would take place. If an even number showed up,
there would be an inspection. The experimenter then drew a number from
chits numbered between one and eight placed in a hat to determine who
would be inspected. Thus, the probability that anyone would be inspected
  
S TU DY IN G A CT IO N S IT UATI ON S I N T HE LA B 95
was 1/16 per round—a low but very realistic probability for monitoring
forest harvesting in rural areas. The monitor checked the investment of
the person whose time had come without revealing the result to others. If
the person was over the limit imposed, a penalty was subtracted from
the payoff to that person. No statement was made to others whether the
appropriator was complying with regulations or not.
The subjects in this experimental condition actually increased their harvesting
levels in this externally imposed sanctioning experiment in contrast
to behavior when no rule at all was imposed, and the subjects could communicate
on a face-to-face basis. What was remarkable about these outcomes
was that subjects, who were simply allowed to communicate with
one another on a face-to-face basis, were able to achieve a higher joint
return than the subjects who had an optimal but imperfectly enforced external
rule imposed on them. These experiments provide further support
for Bruno Frey’s (1997a, 1997b) hypothesis that external rules imposed
on citizens can crowd out intrinsic motivation and lead to worse outcomes
than reached through voluntary agreements.6
As the authors conclude:
We have presented evidence that indicates that local environmental policies that
are modestly enforced, but nevertheless are predicted by standard theory to be
welfare-improving, may be ineffective. In fact, such a policy can do more harm
than good, especially in comparison to allowing individuals collectively to confront
local environmental dilemmas without intervention. We have also . . .
presented evidence that the fundamental reason for the poor performance of
external control is that it crowded out group-regarding behavior in favor of
greater self-interest. (Cardenas, Stranlund, and Willis 2000, 1731)
Common-pool resource experiments have also been replicated and extended
upon by Marco Casari and Charles Plott (2003). Casari and Plott
wanted to explore whether an institution that had been used in the Italian
Alps for centuries and was thought to be highly effective would generate
positive results in a laboratory setting. The Alpine system had a relatively
simple structure.
The population of a village developed a contract among themselves, subject to
the approval of the regional government, called “Carte di Regola,” where they
described a system for monitoring and sanctioning those who are discovered
violating or exceeding patterns of use that the villagers agreed upon in the contract.
The “Carte di Regola” specified in advance the conditions under which
a sanction could be inflicted on a person found in violation of the contract and
the amount of the fine. . . . Any villager could report a violation but he usually
incurred a cost in the form of a monitoring effort to discover the violator and
additional costs to bring him to court. A share of such a fine usually went to
the person who discovered the violator in order to give an incentive to monitor.
(Casari and Plott 2003, 218)
  
C HA PT ER TH RE E96
Casari and Plott used the same functional form for a payoff function as
we had earlier used with Indiana University subjects (see equation 1
above), but they increased the monetary incentives by more than three-
fold.7
Using Cal-Tech subjects, they first ran a baseline experiment that
closely paralleled our earlier baseline experiments. Without communication
or sanctioning, they found—as we had earlier found—that the resource
was substantially overused, even more than the Nash equilibrium.
Subjects earned only 28.4 percent of the optimal return, while the Nash
equilibrium would have earned them 39.5 percent. They also found substantial
variations among individual subjects in the amount they overused
the resource, as we had earlier.
Casari and Plott then changed the transformation function and the payoffs
of the game. They used two sanctioning conditions—weak and
strong. In both conditions, after the decision regarding harvesting had
been made and the total investment levels had been announced, a subject
could select an option to inspect the decision made by any of the other
subjects at a set cost. After this decision had been made, the harvesting
decision of the inspected subject was made public information, but not
the identity or number of subjects requesting an inspection. A fine was
imposed for each unit appropriated above the announced level and transferred
to the inspector. The “inspector” made a profit when the fee that
had been paid to carry out the inspection was less than the amount transferred,
which was in turn dependent upon the amount in which the inspected
appropriator had exceeded the announced level. The weak sanctioning
option did not change the predicted Nash equilibrium for the
game, but the strong sanctioning option made the predicted Nash equilibrium
approach the socially optimal level of appropriation.
In the experiments conducted with weak sanctions, slightly over half
of the actions were inspected—a much higher level than predicted by classical
game theory. And subjects obtained closer to optimal levels of returns
than they had without sanctions. In the strong sanction condition,
the efficiency of the joint return was 94 percent, but when inspection fees
are subtracted, the net return was 77 percent of optimal (238). Almost all
actions were inspected. It turned out that the subjects making the lowest
uses were more aggressive inspectors than those making the highest uses.
Casari and Plott found that the subjects behaved in a manner consistent
with having heterogeneous preferences rather than all subjects having
preferences that were monotonically aligned with the available payoffs.
Some individuals appeared to be more spiteful than others. This helps to
explain the success of the “Carte di Regola” system, as it would appear
that it was able to use the “heterogeneity of preferences to socially advantageous
ends” (241). By sharing the fine with the person who reported
the violation, the system channeled the behavior of the more spiteful indi-
  
S TU DY IN G A CT IO N S IT UATI ON S I N T HE LA B 97
viduals into socially useful purposes. Overall, they found that the experiment
mimicking the set of rules used in the Italian Alps greatly improved
the efficiency of resource use as contrasted to the baseline experiment
without the sanctioning options.
Casari and Plott provide a cogent and important theoretical explanation
for their findings based on a model of individual choice called a “heterogeneous,
linear other-regarding model.” Their model predicts the outcomes
of their experimental designs well, while the model of individual
behavior focused only on monetary returns does not explain the behavior
in their common-pool resource experiments (or that of our own earlier
experiments) well at all. Given that the results from both types of experiments—as
well as many other extensively replicated experiments—are
not consistent with what is predicted from classical game theory, it is time
to discuss the deep problem of animating institutional analysis. We will
do so in chapter 4.
Conclusions
My intention in writing this chapter was to provide several concrete illustrations
for the reader of the working components of an action situation
and how they combine to generate a clear-cut structure of extrinsic incentives.
The experimental researcher must create all of the working parts of
an action situation in the protocols for an experiment and attempt to
isolate the experiment from external, confounding variables. The experimental
lab is thus an excellent environment to learn about action situations
and how changes in one component of a situation affect interactions
and outcomes.
We have also learned a great deal about interactions and outcomes in
two particular games: the Trust Game and the Commons Dilemma. Basically,
we have learned that:
1. In a two-person Trust Game, both Investors and Trustees engage in more
cooperative behavior than predicted when using a model of behavior based on
purely monetary returns.
2. The level of trust and reciprocity achieved in a Trust Game depends on
many factors associated with the relative position of the subjects, the information
they have, and the type of sanctions made available to participants.
3. When participants in an N-person Commons Dilemma are held apart and
unable to communicate on a face-to-face basis (or via the type of signaling that
is feasible in two-person situations), they overuse a common-pool resource.
4. Participants initially use an opportunity for face-to-face discussions to
share their understanding of how their actions affect the joint outcomes and
arrive at a common understanding of the best joint strategy available to them.
  
C HA PT ER TH RE E98
5. Participants are willing to promise others whom they assess as being trustworthy
that they will adopt a joint plan of action. Most individuals keep their
promises (in situations where substantial advantage can accrue for breaking
the promise).
6. If agreements are broken, individuals become indignant and use verbal
chastisements when available. They are also willing to use costly sanctions
when they have the opportunity to select this option, and even tend to overuse
them.
7. Participants use heuristics in dealing with complex problems.
8. Heuristics vary in their capabilities to cope with a changing configuration
of actions by other participants.
9. When given an opportunity to communicate, craft their own rules, and
sanction nonconformance to these rules, some groups were willing to do so.
Through their own efforts, these groups achieve close to optimal results.
Those who forego such an opportunity are not able to sustain a high level of
performance.
In other words, a substantial number, but not all, of the individuals in
these carefully controlled experiments are trustworthy and reciprocate
trust if it has been extended. When behavior is discovered that is not
consistent with reciprocity, individuals are willing to use retribution in a
variety of forms.8
Individuals also initially rely on a battery of heuristics
in response to complexity. Without communication and agreements on
joint strategies, these heuristics lead to overuse. On the other hand,
individuals are willing to discuss ways to increase their own and others’
payoffs over a sequence of rounds. Many are willing to make contingent
promises when others are assessed as trustworthy (Bendor and
Mookherjee 1990).
These conclusions are not consistent with predictions derived from classical
game-theoretic models of participants focusing entirely on monetary
returns in these situations. They are, however, consistent with evidence
gathered from empirical research in the field (Van Vugt et al. 2000). Thus,
it is time that we address the question of how institutional analysis should
be animated. In chapter 4, I dissect the other holon—the participant—
which, together with an action situation, creates an action arena. To make
a prediction about the likely outcomes to be achieved in a particular
arena, one must animate the analysis by specifying key assumptions about
the individual actors holding positions in the situation.
  
Four
Animating Institutional Analysis
IN THE FIRST three chapters of this book, I focused on the exogenous
variables that underpin all action situations and the components of action
situations at operational, collective-choice, or constitutional-choice levels
of analysis. So far, I have provided only a minimal sketch (in chapter 3)
of the contemporary theory used to explain and predict how participants
in action situations are expected to choose among actions. It is this theory
(and models of this theory) that analysts use to predict likely actions of
diverse participants and their cumulative outcomes. Participants are the
second holon of an action arena and the animators of institutional worlds.
Without humans who make decisions in a situation, there is nothing but
the biophysical world to explain.
In this chapter, I discuss the approaches taken by scholars to the puzzle
of animating diverse types of situations ranging from highly competitive
markets to various types of social dilemmas. We will start with situations
that are relatively well understood—open, competitive, posted price markets.
In these settings, a theory of full-information, rational behavior focusing
on material outcomes has been shown to be a powerful engine of
prediction and is consequently very valuable for institutional analysis.
Then, I add complicating assumptions that redefine the information processes,
the valuation mechanisms used by individuals, and/or the selection
processes that individuals adopt.
An institutional theorist must self-consciously posit the kind of information
participants possess, the relevant preference structure of the participants,
and the process they use for choosing among actions. Assumptions
about information, preferences, and choice mechanisms are thus the
essential components of this holon. All three need to be specified in order
to generate hypotheses about interactions and outcomes that can be tested
in a particular type of action situation or linked set of action situations.1
As a scholar trained extensively in both political science and economics,
I have used (and plan to continue to use for many purposes) the basic
assumptions about human behavior in models developed to represent
what is called “rational choice theory” as it has currently evolved in contemporary
microeconomics and game theory (for basic textbooks see
Gardner 2003; Dixit and Skeath 1999; Gintis 2000a).
  
C HA PT ER FO UR100
Animating Open, Competitive Processes
First-generation rational choice theory and related models have proved
valuable for predicting human behavior in stable, competitive market settings
and in competitive electoral and legislative settings where the issue
space is constrained (Aldrich 1995; Nardulli 1995; V. Smith 1991, 2000).
In these stable and repetitive settings, individuals are able to learn about
the full, relevant structure of the situation and attach positive or negative
preferences to actions and outcomes. When one is explaining behavior in
familiar and often-used situations with complete information, one is not
faced with the problem of explaining how individuals learn about the
situation and its likely outcomes and payoffs. One can assume that learning
has taken place and proceed with an explanation of behavior by informed
participants using a mental model of the situation that is at least
roughly approximate to the external situation itself. Explaining how individuals
learn turns out to be extremely challenging.2
Predictions from these models are empirically supported at an aggregate
level in open, competitive, posted-price market settings and at an
individual level in carefully designed experimental settings of competitive
market situations (see V. Smith 1982; Kagel and Roth 1995; Lian and
Plott 1998; Noussair, Plott, and Reizman 1995). If open, posted-price,
competitive markets for the exchange of goods (or similar situations) were
the only type of action situations that individuals faced, then rational
choice theory and its currently accepted models would clearly be the only
theory (and related models) of human behavior that one would need for
prediction of outcomes. Frohlich and Oppenheimer (2001, 22) assess “the
traditional spare economic model of decision making” as being “useful
and robust in predicting behavior in contexts of choice which are relatively
stable, hence in which subjects have learned to call up particular
representations on a repeated basis” (22).
Thus, it is important to learn about relatively simple situations that
have been successfully analyzed and shown to have empirical support
before venturing off into the interesting but difficult work of understanding
and explaining behavior in more complex settings.3
To some extent,
we can think of learning how to analyze behavior in these games as somewhat
similar to a young chemist learning how to make simple compounds
or a biologist learning how to dissect a frog. These tools are not all that
are needed to explain complex chemical and biological phenomena, but
they are a useful starting place. Further, the ways of analyzing relatively
simple situations as formal games have already been developed. It makes
little sense to try to utilize a complex form of analysis, or develop entirely
new forms of analysis, when tools are already available for the analysis
and explanation of behavior in many situations.
  
A NI MATI NG IN ST IT UT IO NA L A NA LY SI S 101
Thus, I strongly advise institutional analysts to learn how to use the
working assumptions of models developed to apply contemporary rational
choice theory. These are:
1. Individuals possess as much information about the structure of a situation
as is contained in the situation itself.
2. Individuals assign a complete and consistent, internal valuation to outcomes
that is a monotonic function of an individual’s own net external payoff.4
3. After making a complete analysis of the situation, individuals choose an
action in light of their resources to maximize expected material net benefits to
themselves given what others are expected to do.
We will use the term rational egoist to describe a participant in a situation
whose behavior can be predicted using these three assumptions. Knowing
the assumptions of a rational egoist and how to use them enables an analyst
to begin using theoretical tools to predict how individuals make
choices and generate outcomes in competitive situations. Further, some
individuals are present in most situations whose choices of strategies can
be predicted using a model of a rational egoist who focuses on the immediate
material payoffs to self and not on other values.
Armen Alchian (1950) demonstrated long ago that those who do not
behave as rational egoists in an open, competitive market will not be present
in any significant number once demand and supply have led to an
equilibrium. More recently, Gode and Sunder (1997) demonstrated that
it is not even necessary to assume that all individuals are rational egoists—
even at equilibrium—to establish efficient markets. Rather, they show that
the efficiency of market exchanges derives from the set of rules constituting
a market, rather than the sophisticated calculation of the participants.
In their model, they show that “zero intelligence traders” reach efficient
outcomes when seven essential rules constrain the actions of buyers and
sellers and not otherwise (see chapter 7, note 13). Thus Alchian, as well
as Gode and Sunder, show that it is the structure of markets that leads
participants to make efficient choices rather than the assumptions made
in economic theory about the internal structure of individual valuation
and choice.
The Challenge of Imperfect Information
Competitive markets or other full-information, competitive situations are
not the only situations facing individuals that are of interest to institutional
analysts. Further empirical research has uncovered a rich array of
anomalies that systematically occur in situations that were once thought
to have properties leading to clear predictions.5
Situations that are rela-
  
C HA PT ER FO UR102
tively simple social dilemmas that make clear predictions about expected
behavior, for example, frequently evoke positive or negative internal values
for participants who are not monotonically related to the objective
payoffs involved (as discussed in chapter 3).
When other-regarding preferences and/or intrinsic values are assigned
to outcomes and actions, the situation is one of incomplete, rather than
complete, information because other players cannot know exactly how
an individual is valuing these actions and outcomes. In order to analyze
these situations, one needs to make different assumptions about the values
taken into account by individuals. Further, as Kenneth Koford (2003)
illustrated with his research conducted in Bulgaria, the way a situation is
framed may strongly affect how individuals embedded in a particular culture
and history view it and value actions and outcomes (McDermott
2001). Frohlich, Oppenheimer, and Kurki (2004) have analyzed simple
situations where small changes in the structure of the situation itself—
what they call the context of the situation—evoke norms in some participants
that were not expressed in a baseline situation.
Instead of complete information, some situations generate only partial
information. Further, the distribution of the information may be asymmetric.
One may also be confronted with the need to examine how individuals
view risk, uncertainty, and information asymmetries and how they
react to the actions and perceived attributes of other participants. Situations
also differ in regard to the extent of repetition and whether the same
individuals continue to interact with one another over time or whether
interactions are largely with a continuous stream of strangers. How do
all of these affect choice? When rational choice models—as usually operationalized—fail
to predict outcomes, other theoretical tools are needed
(Vanberg 2002).
Political economists thus find themselves at a very interesting juncture.
Experimental researchers have shown that behavior after several rounds
in experimental markets and other tightly constrained, competitive processes
tends to be consistent with the predictions stemming from a rational
egoist model of human behavior. On the other hand, experimental
researchers have also shown that behavior in many forms of social dilemmas
and other games is not consistent with what would be predicted if
all individuals behaved in a manner consistent with the rational egoist
model (see Gintis 2000a for a review). Further, the level of trust and cooperation
exhibited varies substantially from one experimental design to
another. The combination of the extensive experimental research and the
strong theoretical arguments made first by Alchian (1950) and then by
Gode and Sunder (1997) that the institutional structure of a market,
rather than the model of the individual, leads to the outcomes predicted
by market theory challenges all political economists. We need to know
  
A NI MATI NG IN ST IT UT IO NA L A NA LY SI S 103
when to use the assumption of a rational egoist for all participants or to
assume some individuals may have other-regarding preferences or value
norms such as trust or reciprocity.
Assumptions Used in Animating Participants
Consequently, I strongly advise institutional analysts not to rely on one
and only one theoretical tool to explain human behavior unless they wish
to confine their analyses entirely to situations that can be successfully
modeled as simple, competitive, complete-information situations. The
three basic assumptions of rational choice theory are, however, a useful
starting point for doing institutional analysis. They illustrate the components
of any theory of human behavior that an institutional analyst would
use in analyzing situations other than highly competitive situations. Thus,
whether a participant is an individual or a corporate actor, the analyst
must make assumptions about three components of human behavior in
order to animate an institutional analysis. These assumptions are the components
of the holon called “participants” in the Institutional Analysis
and Development (IAD) framework. Each of these assumptions can also
be unpacked into multiple layers. These are:
1. the way that participants acquire, process, represent, retain, and use
information;
2. the valuation that participants assign to actions and outcomes; and
3. the processes (maximizing, satisficing, or using diverse heuristics) that participants
use for selecting particular actions or strategic chains of actions in
light of their resources.
Once one decides to explore alternative assumptions about human behavior,
the number of choices that the researcher has to make are substantial.
Alternative assumptions are not likely to be as clean and elegantly
simple as those of rational choice models. Frohlich and Oppenheimer
(2001, 22) reflect that the “standard rational choice model, simple, elegant
and decisive, is not liable to be replaced with as simple and manipulable
a model. The anomalies which have been identified are broad and
diffuse and they are likely to require more theoretical superstructure for
their explanation” (see also Levi 1997b). The grave hesitation of some
theorists to adopt “more realistic” assumptions stems, to a large extent,
from the messiness of the alternative superstructures.
One strategy adopted by many (including the author) is not to change
all of the assumptions at the same time when trying to model a particular
type of situation. In stable and repetitive situations where intrinsic values
are important, one may retain assumption one—regarding full informa-
  
C HA PT ER FO UR104
tion about the structure of the situation—while focusing primarily on the
types of values of importance to individuals and the resulting strategies
they select (see Ahn et al. 2003; Cox 2004). Alternatively, one may focus
primarily on the information processing and assume individuals are
boundedly rational (Selten 1998; B. Jones 2001). If the situation is uncertain
and complex, one may focus on the third assumption related to how
individuals make decisions about their choices. In many field settings,
participants use heuristics or rules of thumb that they have learned over
time—or were taught to them by their seniors—that give them good
enough solutions that there is little need to pay the costs of a full search
(see Gigerenzer and Selten 2001). Thus, which assumption (or assumptions)
one changes and the particular assumptions made depend on the
situation to be explained.
A thick book could be devoted to elucidating the different theories and
models that posit diverse assumptions about these three broad components
of human choice. In order to keep the present book focused, I will
briefly discuss here only some of the most relevant alternative assumptions
that are used by institutional theorists and will refer the reader to
contemporary literature related to these assumptions. I will conclude with
a discussion of the importance of the institutions within which individuals
interact to the likelihood that individuals with normative orientations and
other-regarding preferences may be able to achieve higher outcomes over
time in social dilemmas and may even change the distribution of strategies
used by individuals over time.
Assumptions about Information Processing and Mental Models
Many of the situations of interest to institutional analysts are uncertain
and lack the kind of rules leading to the selective pressures and information-generating
capabilities of highly competitive processes. In analyzing
these situations, theorists frequently substitute the assumption of
bounded rationality—that participants are goal oriented and try to be
rational but face cognitive limits—rather than the assumptions of complete
information (see Simon 1957, 1972, 1995; Williamson 1985, 2000;
V. Ostrom 1997; B. Jones 2001). In some complex situations, the available
information may greatly exceed the competence of an individual to
compute a solution based on fuller analysis (Heiner 1983). As discussed
in chapter 2, the game of chess exceeds the capacity of any human to
compute an analytical solution. In analyzing the behavior of chess players,
one can assume that they know the current pattern of chess pieces on the
board, but not that they know all of the action-outcome linkages.
Most cognitive scientists stress that humans expend substantial effort
to make sense out of the variety of signals (and symbols) they receive as
  
A NI MATI NG IN ST IT UT IO NA L A NA LY SI S 105
Mental
Model(s)
Perception of
the Situation
Culture
Expected
Outcomes
Revise
Model
Possible
Actions
Participant in a Situation
External Action Situation
Chosen
Actions
Information about
Actual Outcomes
of Prior Action
Information about
Action Situation
Actual Outcomes
Figure 4.1 The relationship between information, action-outcome linkages, and
internal mental models. Source: Adapted from Denzau and North 2000, 36.
they interact within a variety of situations in life (Busemeyer and Myung
1992). Individuals attempt to create a mental model or a representation
of diverse situations so as to be able to make reasonable decisions in these
multiple settings (Holland et al. 1986). Mental models are affected by at
least two basic sources—feedback from the world and the shared culture
or belief system in which an individual is embedded.
As individuals interact in a diversity of situations, they receive information
about the structure of the situations they are in prior to making a
decision and usually receive some kind of feedback after they take an
action. Participants need to discover an appropriate model of the situation
they are in through repeated interactions in it or similar structures (Plott
1996). Drawing on Denzau and North (2000), one can envision a participant
in a situation receiving information about its structure (how many
participants, who they are, the benefits and costs of action-outcome linkages,
etc.) (see figure 4.1). The participant initially relies on earlier mental
models formed of this situation to calculate expected outcomes of diverse
actions. If satisfied with the outcomes, they may not search for further
information. What they learn about the outcome of their own and others’
actions, however, will potentially stimulate an effort to revise their mental
model if there is an incongruity or a lack of satisfaction.
  
C HA PT ER FO UR106
Cultural belief systems also affect the mental models that individuals
utilize. Most of childhood is spent in a combination of observing others
interacting, being told the prudent or proper way to perceive situations
and to act within them, and taking actions based on both observation and
instruction (Tomasello 2004). Parents actively encourage their children to
learn a culturally appropriate set of strategies, including industriousness,
responsibility, and self-reliance, for doing well in a variety of situations
(Low 1989). The distribution of imparted traits varies by whether a culture
is highly stratified or not, and by the gender of the child (ibid.). When
we indicate that people share a culture, it is a shorthand way of indicating
that the wide diversity of mental models that individuals have invented
has been reduced to a smaller set within those sharing the culture (Benedict
1934). Culture may also be viewed as an intergenerational transfer
of past experience. The mental models that scientists hold come about
initially from what they learn in school as these are modified by their own
research and that of others (Gopnik 2004).
Individuals learn from experience and from shared mental models.
Learning is enhanced in situations that are often repeated. Interactions
with the same set of individuals enable an individual to obtain a better
estimate of the strategies that specific others adopt. Theoretical and experimental
work has shown that individuals tend to experiment with diverse
actions and then adopt those that have returned the highest payoffs in the
past (Selten, Abbink, and Cox 2001; Busemeyer and Myung 1992). If the
situation in which individuals interact is relatively stable and repetition
occurs frequently, and if there is pressure to improve performance, individuals
will tend to discover those strategies that an omniscient individual
would have selected.6
Of course, the larger the number of individuals involved
in a repeated situation, and the more diverse their strategies, the
more difficult it is for anyone to gain an accurate perception of others’
strategic behavior.
The convergence through learning to the same strategy as is predicted
under full rationality is unlikely to happen when the number of participants
in a situation is large and the situation itself is complex, changes
frequently, and/or the individuals do not participate in that situation with
regularity or any induced need to increase performance. Information
search is costly. The information-processing capabilities of human beings
are limited. Individuals must often make choices based on incomplete
knowledge of all possible alternatives and their likely outcomes. With
incomplete information and imperfect information-processing capabilities,
individuals may make errors in perception, in their comprehension
of how a complex structure works, or simply in computations (V. Ostrom
1986, 1997).
  
A NI MATI NG IN ST IT UT IO NA L A NA LY SI S 107
The experiences that different individuals have had and their interpretation
of them may differ substantially. It is thus possible that a single individual
might be able to call up more than one mental model or that the
mental model of participants in the same situation will differ. This presents
a theoretical problem for the analyst trying to understand behavior.
“How can we understand a person’s choices if, when confronting a given
environment, she can have multiple representations, each of which is associated
with different values?” (Frohlich and Oppenheimer 2001, 6).
Denzau and North (2000) stress the importance of communication as
one way to enable individuals to develop a shared mental model. If the
same individuals repeatedly interact in free and open communication, and
if they already share some broad cultural views of the world, mental models
are more likely to converge. The positive impact of communication—
particularly face-to-face communication on joint outcomes in a wide diversity
of social dilemma experiments—has been replicated frequently, as
discussed in chapter 3.
Individuals may not always be able to engage in open, repeated communication,
however. John Schiemann (2002) examines this problem in his
analysis of the mental models used by different age groups coming from
the same culture but who have had vividly different experiences in life.
Older Hungarians, who had directly experienced the bloodshed of World
War II, interpreted the situations following the ending of the Soviet Union
differently than did younger Hungarians who only read about the earlier,
dramatic events. A generational difference also exists on how individuals
view the Chinese leadership’s decision to use force in the Tiananmen incident
in 1989.7
Unless there is open and active communication among
participants over some time, they may simply use different mental models
when interacting in the same external situation. This can lead to gross
misunderstandings and disappointments or even to major violence.
Frohlich and Oppenheimer (2001) address the puzzle of multiple internal
models by asking what factors in an environment are likely to affect
the way a participant perceives a situation and the way the individual
optimizes in that situation given that “paying attention” is costly. Few
humans gain full mastery of all of the potentially relevant details in a
complex, ongoing situation. Frohlich and Oppenheimer single out two
properties of a situation as most important in affecting the way a situation
is perceived. “The first property is the salience of the elements in a choice
situation,” by which they mean “the degree to which an element is linked
to possible changes in the welfare of the decision maker” (8). The second
property is the vividness of the situation or the “amount and quality of
the sensory details of the objects encountered” (8). These attributes are
important in gaining attention given the variety of signals an individual
receives. “In order for something to grab one’s attention it must displace
  
C HA PT ER FO UR108
Mental
Model(s)
Perception of
the Situation
Culture
Expected
Outcomes
Revise
Model
Possible
Actions
Participant in a Situation
External Action Situation
Chosen
Actions
Information about
Actual Outcomes
of Prior Action
Information about
Action Situation
Vividness Salience
Communication
Actual Outcomes
Figure 4.2 The impact of communication, vividness, and salience on the relationship
between information, action-outcome linkages, and internal mental models.
something else to which one is attending. To accomplish this, a new focus
of attention must have a higher claim. Attention shifts from one object of
attention to another as if there were a threshold of value attached to the
former which has to be surpassed for the competitor to displace it” (8).
The repeated finding that face-to-face communication in social dilemmas
is more effective than written communication may be related to its being
a more vivid form of communication.
The role of vividness of symbols or rituals in solidifying the shared
mental models of large groups of individuals is an important theme in the
work of anthropologists (Rappaport 1979, for example). The philosopher
Allan Gibbard (1990) stresses the importance of rituals in achieving
shared norms of what participants should do in particular situations and
helping to increase the probability that they will do so in the future partly
because they share a vivid memory of what should be done.8
Thus, the
vivid ritual or symbol helps to select one mental model over others. The
resulting congruence in behavior and outcomes helps to reinforce that
model among those sharing the same culture. The role of communication,
salience, and vividness in affecting mental models can be represented as
in figure 4.2.
  
A NI MATI NG IN ST IT UT IO NA L A NA LY SI S 109
The architecture of information processing broadly reflected in figure
4.2 illustrates that individuals do not respond immediately to all information
but rather to that which is communicated in a salient and vivid man-
ner.9
On the basis of his own and others’ research, Bryan Jones argues
that individuals and the organizations with which they work are disproportionate
information processors. “Disproportionate information processing
means that inputs into a decision-making process do not link directly
to outputs. As a consequence, there is an imperfect match between
the adaptive strategies people devise and the information they receive.
This mismatch is the inner cognitive and emotional architecture of the
human brain ‘showing through’ in responding to information” (Jones
2001, 9). Jones illustrates this mismatch with a variety of imaginative
quantitative studies of disproportionate information processing in shifting
electoral responses over time, in public budgetary changes, and in
the level of coverage of national policy issues by Washington-based news
publications. The substantive interpretation of his results for an institutional
analyst is that: “change in human institutions tends to be quite
conservative—most cases cluster around a central peak—but is subject to
occasional quite large punctuations (the tails). . . . It would seem that a
hypothetical decision maker would have to be prepared either for virtually
no change or a very large change—he or she could not hope for moderate
adjustments to changing circumstances” (184; see also Jones,
Baumgartner, and True 1996).
Denzau and North (2000) reflect a similar view of the way change happens.
They use the term “punctuated equilibrium” to reflect their sense
that the internal models used by individuals remain relatively similar until
some event triggers a large change in the mental model and resulting
changes in the actions that individuals (or organizations) take. Denzau
and North, as well as Jones, also illustrate how individuals are able to
overcome some of their own information-processing limits by establishing
rules and routines that structure situations in such a way as to enhance
the likelihood that individuals will share a mental model of the situation
and take actions that lead to better rather than worse outcomes (North
2005). Humans are thereby able to compensate for—as well as replicate—
cognitive processing limits by the way they organize themselves and the
procedures they follow (see also Dawes 1988).
Assumptions about Valuation Processes
How individuals value actions and outcomes has become a very active
discourse among scholars interested in institutional questions (Fehr and
Falk 1999; Fehr and Ga¨chter 2000a, 2000b; Bolton and Ockenfels 2000;
Casari and Plott 2003; Frohlich, Oppenheimer, and Kurki 2004; Janssen
  
C HA PT ER FO UR110
and Ahn 2003). Assuming that all individuals are self-interested utility
maximizers without exploring how individuals reach utility judgments
was a satisfactory scholarly strategy for many years when explaining behavior
in highly competitive market settings where one could implicitly
equate utility with profits. As mentioned above, substantial field and experimental
research supports the second assumption of rational choice
when individuals are interacting in a stable, competitive environment
with full information about the short-term options available to them.
When one turns to the large realm of social dilemmas, however, little
empirical support exists for an assumption that all individuals value only
the material outcomes that flow only to them.10
If this were the case, we
would not have seen the extensive evidence presented in chapter 3 that
Investors frequently send funds to Trustees and Trustees often send funds
back to the Investor at a personal cost to themselves to fulfill the trust
that the Investor placed in them. Nor would we have seen subjects keeping
promises that they have made in face-to-face communication rounds to
other subjects in a Commons Dilemma experiment to keep their appropriation
levels low. Many subjects appear to have other-regarding preferences
and to accept norms of behavior backed up by emotions including
pride, guilt, shame, and anger. Whether norms are invoked and lead to
cooperative behavior varies across experiments that differ only in regards
to relatively small structural features.
The results discussed in chapter 3 are hard to explain using the standard
rational choice model that all individuals who face the same objective
game structure evaluate decisions the same way!11
And, we cannot simply
resort to the easy criticism that undergraduate student subjects are being
paid a modest sum (or nothing at all) and thus the game is not a valid
test of the theory. Many subjects were not students in these and other
experiments conducted in multiple countries (Henrich et al. 2004).12
Increasing
the size of the payoffs offered in experiments does not appear to
change the broad patterns of empirical results obtained.13
Nor, on the
basis of experimental evidence, can we simply change from an assumption
of universal selfish behavior to an assumption of universal altruism. As
Fehr and Ga¨chter (1998, 847) stress, “Homo Oeconomicus Also Exists.”
In all social dilemma experiments, a “non-trivial minority of subjects exhibits
selfish behavior” (847).
Thus, the results of the experiments summarized in chapter 3—and
many others documented by Camerer (2003)—substantially challenge the
second assumption of contemporary rational choice models that valuation
of all players is always focused entirely on extrinsic, immediate, net
benefits to the individual. These well-substantiated facts provide an essential
foundation of a more eclectic (and classical) view of human behavior.
Further, as Ben-Ner and Putterman (2000a) stress, humans invest substan-
  
A NI MATI NG IN ST IT UT IO NA L A NA LY SI S 111
tial time in their attempt to “mold the character” of their offspring. If this
were truly impossible, it would be irrational for adult humans to spend
so much time trying to transform the preference structure of their offspring.
On the other hand, adopting the closely related, internal norms
of trustworthiness and trust are not the unchanging, universal attributes
of all individuals. It is important that we develop tools of analysis that
reconnect the normative side of individual orientations with the calculation
of individual benefits.14
Recent research by Rilling and his colleagues (2002) provides an even
further challenge. Freely recruited and paid subjects participated in a series
of repeated Prisoner’s Dilemma games while connected to a magnetic
resonance imaging (fMRI) scanner. In each round, subjects pressed a button
to indicate their choice of cooperation or defection leading to a payoff
per round of $2 each if both cooperated, or of $3 if a subject defected
and if the other cooperated, or $1 if both defected. The researchers compared
levels of cooperation and brain activities in several designs including
one in which the second player was a programmed computer strategy.
The results of each round were projected onto a screen that the player in
the scanner could see and onto the computer screen of the other player
(where applicable). The observed behavior was similar to that of other
repeated dilemma experiments.
While the observed interactions and outcomes were not novel, linking
the portrait of brain activity to behavior in social dilemmas adds a new
dimension to the body of findings. And what they found was surprising
to the research team (as reported to Angier, July 26, 2002) as well as to
many readers. Pressing the cooperate key in a reciprocating relationship
generated the brightest reactions in the pleasure zones of the brain (the
anterovential striatum and the orbitofrontal cortext). The researchers also
tested whether the same level of brain activity was associated with receiving
a payoff of $2 in a nonsocial context. They found different patterns
of neural activation depending on whether the scanned player thought
she was playing a real human or a computer (see also McCabe 2003).
They concluded that they had “identified a pattern of neural activation
that may be involved in sustained cooperative social relationships, perhaps
by labeling cooperative social interactions as rewarding, and/or by
inhibiting the selfish-impulse to accept but not reciprocate an act of altruism”
(Rilling et al. 2002, 403). These findings are consistent with the
earlier theoretical work of Robert Frank (1988), who posited that human
emotions underpin a commitment to reciprocity so that humans feel good
when achieving mutual cooperation and feel bad when defections (their
own or others) occur.
Given the extensive evidence from many empirical studies, at least some
social scientists now tend to follow Amartya Sen’s (1977) advice to stop
  
C HA PT ER FO UR112
assuming that all individuals are “rational fools.” Instead, they assume
that at least some individuals in some situations do have other-regarding
preferences and do follow norms of behavior (see, for example, Frohlich
and Oppenheimer 1996; McCabe and Smith 2003). As discussed in
some detail in chapter 5, norms can be thought of as shared concepts
of what must, must not, or may be appropriate actions or outcomes in
particular types of situations. Individuals add positive or negative values
to objective payoffs invoked by emotions such as pride (when positive)
or shame (when negative). Norms such as those of trustworthiness,
trust, and fairness will lead individuals to take actions in some situations—like
those reported earlier—that are directly contrary to those predicted
using a rational egoist assumption. While “foolish” according to
these models, if those following norms can identify others with similar
norms, they will actually earn more than those seeking their own shortrun
selfish objectives.15
The behavior of many participants can be thought of as partially based
on intrinsic preferences related to how they prefer to behave (and, obviously
would like others to behave) and the kind of outcomes they wish
to see themselves and others obtain. Norm-following individuals take into
account other individuals’ interests as well as their own in the decisions
they make (Frey 1994, 1997a). These “nonselfish” individuals, however,
differ among themselves in terms of the extent to which they depart from
purely selfish motivations and do not adopt an unchanging strategy when
interacting with others in repeated situations. Further, individuals differ
in regard to how they interpret actions by others that reduce their own
payoffs. Some individuals can easily find themselves enmeshed in everexpanding
threat systems (Boulding 1963) whereby one bad deed cannot
rest unpunished. Reciprocity can have a very dark side when bads have
to repay bads over time. Unfortunately, “envy, vengeance and the desire
to dominate are not less intrinsically motivated than altruism, conscientiousness,
and love. All of these motives contribute to immediate satisfaction
rather than to achieving externally set goals” (Osterloh and Frey
2000, 540).
Turning to the bright side, psychological research provides evidence
that positive intrinsic motivation is increased when individuals feel that
their own self-determination or self-esteem is enhanced (Deci and Ryan
1985; Deci, Koestner, and Ryan 1999). This leads to the possibility that
intrinsic motivation can be “crowded out” in situations where individuals
do not perceive themselves to have sufficient self-control over the actions
they take. The recent experiments by Fehr and Rockenbach (2003) and
Cardenas, Stranlund, and Willis (2000), discussed in chapter 3, provide
strong evidence for the crowding out of reciprocity by the imposition of
external sanctions. In a review of crowding-out theory, Frey and Jegen
  
A NI MATI NG IN ST IT UT IO NA L A NA LY SI S 113
(2001, 591) identify the conditions that affect the level of intrinsic motivation
that individuals may feel:
1. External interventions crowd out intrinsic motivation if the individuals
affected perceive them to be controlling. In that case, both self-determination
and self-esteem suffer, and the individuals react by reducing their intrinsic motivation
in the activity controlled.
2. External interventions crowd in intrinsic motivation if the individuals
concerned perceive it as supportive. In that case, self-esteem is fostered, and
the individuals feel that they are given more freedom to act, which enlarges
self-determination.
Just as individuals may have different mental models of the situations
they are in, they may differ in regard to their internal valuation patterns—
the extent they take others into account in the decisions they make and
the intrinsic valuation they may place on taking particular types of actions
(e.g., being trustworthy) or reaching particular types of outcomes (e.g.,
more equitable). Further, the form that normative and other-regarding
behavior takes “may be substantially context dependent. No ‘single’
model which explicates a particular set of values is likely to be adequate
to capture behavior in all contexts” (Frohlich, Oppenheimer, and Kurki
2004, 116).
Assumptions about Selection Processes
Reviewing the assumptions about information and valuation processes
has already opened a Pandora’s box of complexity. Examining alternative
assumptions about the selection processes that individuals use does not
help reduce the complexity. One can make statements like: “Individuals
will try to do as well as they can given the information they obtain.” Or,
“Individuals will use heuristics that have been proved to work in
the past.”
In Rules, Games, and Common-Pool Resources, for example, we examined
the possibility that individuals in an experimental common-pool resource
situation were using a heuristic that we called “measured reaction”
(E. Ostrom, Gardner, and Walker 1994, 199). After engaging in face-toface
discussions where subjects had reached a simple agreement regarding
what they all should do in future rounds, each subject had to make a
personal decision as to whether to keep to the agreement and what to do
if someone deviated from their agreement. Most subjects kept to their
agreement. When deviations did occur, most subjects first reacted with a
moderate increase in their own harvest rates.
A different strategy that scholars have posited that participants will use
in coping with this type of situation is called the “grim trigger.” A person
  
C HA PT ER FO UR114
using a grim-trigger strategy would react to any deviation from a verbal
agreement by withholding cooperation on every play for the rest of the
game. While subjects discussed such an option, they always rejected it. In
trying to understand how they were behaving, we posited that players
were reacting mildly (if at all) to a small deviation from an agreement. If
defections continued over time, subjects using a measured response slowly
changed from keeping the agreement toward actions consistent with the
Nash equilibrium (E. Ostrom, Gardner, and Walker 1994, 200). An analysis
of the round-by-round decisions by subjects to the deviations that did
occur generated strong statistical support for the use of the measured
reaction heuristic (200–215). By following this heuristic, individuals
achieved a much higher payoff than predicted if they had used a maximization
of expected net benefits assumption.
That experience certainly increased the warrantability of the claim that
individuals use heuristics, in my view. And, the extensive research by
Gigerenzer and his research team on “fast and frugal heuristics” provides
substantial evidence across a diversity of situations (Gigerenzer, Todd,
and the ABC Research Group 1999; Gigerenzer and Selten 2001).
There is, however, a “serious and perhaps intractable induction problem
in inferring the structure of a black box from the structure of the behavior
it produces” (J. Anderson 1991, 471). In other words, it is hard to tell
from behavior which of a variety of potential heuristics individuals are
really using.
In an effort to answer this question, scholars are using a variety of
ingenious methods. Rieskamp and Hoffrage (2003), for example, devised
a series of experiments where they asked subjects to make decisions about
a business firm’s profitability based on their best estimate obtained from
a cursory examination of an array of cues. By observing the cues subjects
used—by opening a box on a computer screen—and how much time they
spent, Rieskamp and Hoffrage could assess the heuristics that the subject
used. As part of the experimental conditions, they also changed the
amount of time allowed to subjects in order to observe how time constraints
affect the selection process.
Rieskamp and Hoffrage identified the eight heuristics shown in table
4.1 as among those that subjects might use in this kind of a choice situation.
As one can see from table 4.1, there is a rich array of potential
heuristics that subjects could potentially use. Under low time pressure,
Rieskamp and Hoffrage found that subjects tended to use noncompensatory
heuristics—the PROS heuristic shown on table 4.1. The simple heuristic
LEX was the best behavioral model to describe their behavior under
high time pressure. They then conducted a Monte Carlo simulation to
evaluate the performance of the eight heuristics listed on the table against
using a multiple regression method to make the same decision. They
  
A NI MATI NG IN ST IT UT IO NA L A NA LY SI S 115
TABLE 4.1
Description of various decision strategies
Strategy* Description of strategies
LEX The lexicographic strategy (Fishburn 1974) selects the alternative
with the highest value on the cue of the highest validity. If two or
more alternatives have the same highest cue value, then for these
alternatives the cue with the second highest validity is considered,
and so on. LEX is the general form of the “Take the Best” heuristic
investigated in Gigerenzer and Goldstein (1996).
LEX-Semi Lexicographic semiorder (Luce 1956) works like LEX, with the
additional assumption of a negligible difference (in the present
paper, this difference was set at 1). Pairs of alternatives with a
negligible difference between the cue values are treated as not
discriminative.
EBA Elimination by aspects (Tversky 1972) eliminates all alternatives
that do not exceed a specified value on the first cue examined (in
the present paper, this value was set at 2). If more than one alternative
remains, another cue is selected. This procedure is repeated
until only one alternative is left. Cues are selected in order of their
validity.**
Features The features strategy (Alba and Marmorstein 1987) selects the
alternative with the highest number of good features. A good feature
is a cue value that exceeds a specified cutoff (in the present
paper, this cutoff value was set at 3).
ADD The additive strategy calculates for each alternative the sum of
the cue values (multiplied by a unit weight of 1) and selects the
alternative with the highest score.
LEX-ADD The LEX-ADD strategy is a combination of two strategies. It
first uses LEX-Semi to choose two alternatives as favorites, then
evaluates them by the ADD strategy and selects the one with the
highest sum.
PROS The weighted pros strategy (Huber 1980) selects the alternative
with the highest sum of weighted “pros.” A cue that has a higher
value for one alternative than for the others is considered a pro for
this alternative. The weight of each pro is defined by the validity of
the particular cue.
WADD The weighted additive strategy calculates for each alternative the
sum of the cue values multiplied by the corresponding cue validates
and selects the alternative with the highest score.
Source: Rieskamp and Hoffrage 2003, 50.
* Strategies occasionally do not end up with one single prediction; in this case, it was
assumed that the strategy would randomly choose between the remaining alternatives.
** In contrast to this deterministic selection, in the original formulation of the EBA heuristic
the cues are selected with a probability proportional to their weights (Tversky 1972).
  
C HA PT ER FO UR116
found that the simple heuristics—particularly the LEX heuristic, which
actually requires the least information—did very well in comparison to
the “optimal” method for making this decision.16
In their effort to understand
the use of heuristics, Jager, Janssen, and Viek (2001) used the dimensions
of uncertainty and satisfaction to determine which heuristic would
be used. The more an individual is both satisfied and certain about the
environment, the less likely they will put much effort into making a calculated
choice.
Another puzzle in regard to the use of heuristics is: What is the effect
of diverse heuristics on outcomes achieved? Gigerenzer and his colleagues
(1999) have identified a number of heuristics that they have found to
enable individuals and firms to make quick decisions and do very well
over time. Other researchers are somewhat more skeptical of the efficiency
of heuristics—especially in highly volatile environments. Gu¨th and
Neuefeind (2001), for example, explore the over-time efficiency of a heuristic
that appears to be used in many consumer choices—a form of directional
learning. This involves continuing or increasing an action if past
actions have generated good outcomes and decreasing if bad outcomes
occurred. In relatively stable environments, they found that the heuristic
helped individuals move toward an efficient set of decisions. In a highly
volatile environment, however, learning from the use of this heuristic only
slowly improves achieved outcomes.
Variety and Complexity: An Asset or a Liability?
What I hope the reader gains from this brief overview of the broad set of
assumptions used by theorists to animate analyses of multiple situations
is a good sense that research in the behavioral social sciences is gradually
increasing our capabilities to understand and predict interactions and outcomes
in a diversity of action situations. The large number of replicated
experiments provide substantial evidence for theorists to use in testing
out a diversity of assumptions about basic human behavior (see Bolton
and Ockenfels 2000; Fehr and Schmidt 1999; Frohlich and Oppenheimer
2001; Charness and Rabin 2003; Bowles 1998; Cox, Friedman, and Gjerstad
2004; Cox and Sadiraj 2004; Gintis 2000a; Rabin 1993; E. Ostrom
1998; Camerer 2003). What is also encouraging is that research in biological
evolution (Maynard Smith and Szathma´ry 1997; Kurzban 2003), cultural
evolution (Boyd and Richerson 1985), and their interrelationships
(Henrich 2004; Richerson and Boyd 2002; Sethi 1996) is helping to provide
a better foundation for understanding how humans could have
evolved such a rich set of preferences in addition to the universal goal of
seeking to do well in regard to material payoffs.
  
A NI MATI NG IN ST IT UT IO NA L A NA LY SI S 117
The reaction of some scholars, however, to the growing richness of
behavioral theories of human decision making is to view the variety as
leading to messy complexity. Given that humans may be viewed as pursuing
multiple values in diverse situations depending to some extent on their
prior life’s experiences, does it matter anymore which assumptions one
uses? Does theory underlie all of these models? Is it just a hodgepodge of
assumptions? Or, are there some major lessons to be learned?
My answers to these questions are: Yes, there is a broad underlying
theory of human behavior that is itself developing over time as a result of
the extensive empirical research and theoretical effort. Yes, it matters a
lot that we can make multiple assumptions about how participants in a
situation make decisions. The variety of assumptions is an asset and not
a liability when used to design research so as to test out the implications
of one set of assumptions versus others (see McCabe and Smith 2003; Cox
2004). We are beginning to make some real headway in understanding the
behavior of the extremely complex animal called Homo sapiens. In this
effort, we can learn a lot from engineers and how they draw on the general
laws of physics.
Engineers work with multiple types of motors that are used to propel
vehicles. To predict energy use, length of service, fragility to impact, and
other important outcomes, an engineer needs to know specific facts about
the situation (the kinds of roads or rail systems, their roughness and steepness,
the amount of traffic, etc.) and about the motor involved (its internal
composition, fuel needs, horsepower, etc). A skilled engineer is trained
to work with multiple types of animation devices—motors—that are all
consistent with underlying laws of physics. The underlying laws are too
broad and general to provide specific guidance as to what to expect in
particular settings. Engineers must design motors to perform well in specified
contexts rather than in any and all contexts. Engineers use very detailed
specifications. This does not mean, however, that the design of motors
lacks a foundation in general physical laws.
Like engineers, social scientists need to recognize that to predict outcomes
we must match the animating assumptions about the participants
to the structure of the situation or linked set of relevant situations we are
analyzing. While social scientists do not design the animators of action
situations, we must understand how humans tend to behave in diverse
situations. Our explanatory models of human behavior will differ when
we explain behavior in a repeated social dilemma among individuals who
have built reputations for trustworthiness and conditional cooperation
as contrasted to total strangers who cannot even communicate with one
another. When the situation relates to private goods and is competitive
and stable; when it generates considerable information about its structure
and the actions of participants; and when participants voluntarily enter
  
C HA PT ER FO UR118
primarily to achieve material outcomes, animating the situation with rational
egoists generates empirically supported propositions. Or, if the situation
is extremely simple and lacks any context that provides a clue as to
who else is involved and the appropriate norms to use, individuals tend
to rely primarily on what is best for themselves (Frohlich, Oppenheimer,
and Kurki 2004).
Unlike motors, however, the animators of social life adapt and change
over time. Humans do not have fixed characteristics. What makes understanding
human behavior so difficult is exactly our capacity to try out
multiple norms, heuristics, or strategies, and to learn to use one set in one
situation while using different mixes in other situations. The reason we
can characterize participants as rational egoists in an open, competitive
market is because of the institution, not that all of the participants are
narrowly selfish! The same individuals who energetically pursue profitmaximizing
strategies from 8 A.M. to 5 P.M. every workday may also volunteer
several evenings a month on neighborhood projects, contribute
substantial funds to diverse charities, regularly vote, and be known to
friends and coworkers as kind, considerate individuals who always do
more than their share of any team project. These individuals are also likely
to have biases in the way the world is interpreted and a variety of other
human failings.
All of these aspects of humans are consistent with a theory of boundedly
rational, fallible individuals who pursue multiple goals for themselves
and others, adopt contextually relevant norms of behavior, and
can learn better strategies in a particular situation over time—particularly
if it generates accurate information about key variables. We need to draw
on and expand the basic work of Herbert Simon, Vincent Ostrom, Douglass
North, Reinhard Selten, Bryan Jones, Oliver Williamson, and others
who have posited and continue to develop a general theory of bounded
rationality.
Like all creatures honed by millennia of evolutionary processes, humans
do seek beneficial outcomes for themselves. Information search is
costly, and the information-processing capabilities of human beings are
limited. Individuals, therefore, often must make choices based on incomplete
knowledge of all possible alternatives and their likely outcomes.
With incomplete information and imperfect information-processing capabilities,
all individuals may make mistakes in choosing strategies designed
to realize a set of goals (V. Ostrom 1986). Over time, however, individuals
can acquire a greater understanding of their situation and adopt strategies
or heuristics that result in higher returns. Bounded rationality, however,
has focused mostly on the information condition related to how participants
choose.
  
A NI MATI NG IN ST IT UT IO NA L A NA LY SI S 119
The theoretical challenge facing scholars today is developing an appropriate
family of assumptions to make about the intrinsic values individuals
place on actions and outcomes—particularly outcomes obtained by
others. Individuals adopt norms of behavior taught them by parents and
others in the situations in which they find themselves. Which norms are
supported and become a strong influence affecting decisions is affected
by the history of experiences that individuals have (or have not) shared
and by the specific experiences they face in a particular decision situation.
Humans tend to partition action situations into those where norms
are not taken seriously into account and those where norms matter (Harsanyi
1955).
Social scientists have to match their initial assumptions about the orientations
of participants to the situation they are trying to understand and
explain. What kind of goods and services are involved, what rules, and
what kind of community surrounds a particular situation? We have to
ask whether the situation is stable or changing, conveys substantial information
about its structure and the behavior of participants, tends to invoke
norms such as trust and reciprocity (or those of an eye for an eye),
and allows participants to adapt more effective strategies over time?
The effort to develop better theory, specify testable propositions, and undertake
carefully designed field and experimental work is crucial to our
enterprise.
The two fundamental lessons from the vast empirical and theoretical
research of the last several decades are: first, humans have complex motivations
including narrow self-interest as well as norms of proper behavior
and other-regarding preferences; and second, institutions matter! To move
beyond these important lessons to better understand institutional diversity,
we need to address three major issues: (1) How can we cut through
the complexity to focus on problems that are ripe for further growth? (2)
How can we include norms in our analysis without falling into the trap
that all that is needed is to assume that individuals learn and use norms?
and (3) How can we gain a better grasp of what we mean when we say
that institutions matter?
A Focus on Collective Action to Overcome Social Dilemmas
First, to cut through the complexity we need to concentrate on broad
areas of nonmarket situations. Market institutions work well when the
goods involved are private goods. Individuals trying to provide public
goods or sustain common-pool resources (as defined in chapter 1) find
themselves facing a variety of social dilemmas that are not easy to solve.
We are likely to make more progress if we do not try to develop a single
  
C HA PT ER FO UR120
model of human behavior that can be used to predict behavior in all market
and nonmarket action situations. A more focused effort to explain
collective action in overcoming social dilemmas appears to have a higher
probability of success in the near future.
After all, traditional rational choice theory emerged from a consistent
effort of many scholars to develop a rigorous theory and models of human
choice within one broad type of institutional arrangement—that of a
competitive market. The expansion of this theory to other types of action
situations did not occur until long after it had proved itself to be successful
in predicting outcomes in one type of institutional arrangement. It was
not until the influential work of Kenneth Arrow (1951), Anthony Downs
(1957), Mancur Olson (1965), William Riker (1962), and James Buchanan
and Gordon Tullock (1962) that scholars began to take seriously
the possibility of using rational choice theory to explain voting, legislative
decisions, decisions within a hierarchy, and engagement in collective
action to overcome social dilemmas. And we have learned a lot from
this effort to apply a systematic theory to explain behavior in nonmarket
situations.
Part of what we have learned is that highly competitive situations exist
outside of the market where the narrow theory of rational behavior generates
useful explanations for important empirical phenomena (Shepsle and
Weingast 1984, 1987; Tsebelis 2002). We have also learned that the predictions
of zero contributions derived from this theory are not supported
in many social dilemmas. Explaining successful and unsuccessful efforts
to engage in collective action to overcome social dilemmas should be a
high priority for institutional theorists.
Explaining the diversity of outcomes in social dilemma situations is a
puzzle that is ripe for further development given the quality and quantity
of relevant research. It is also an important question to pursue if one
presumes that humans are capable of developing, transmitting, and learning
norms of trust, trustworthiness, reciprocity, and equity as well as
learning how to govern themselves. Without further progress in developing
our theories and models of human valuation in social dilemma situations,
those convinced that all human behavior can be explained using
rational egoist models will continue to recommend Leviathan-like remedies
for overcoming all social dilemmas. Hopefully, much of what we
learn from focusing on behavior in social dilemmas will be useful in other
puzzling nonmarket situations.
In focusing on social dilemmas, we need to address how to focus on
the role of norms and other-regarding preferences. Simply explaining
puzzling findings post hoc, as “they must somehow share some norms,”
is not a satisfactory strategy in the long run. Focusing on norms and
  
A NI MATI NG IN ST IT UT IO NA L A NA LY SI S 121
other-regarding preferences is not enough, however, to explain fully how
individuals do overcome social dilemmas. Rules are needed to back up
these norms (or counteract dangerously escalating negative reciprocity).
We then need to dig into the analysis of institutions so that we can understand
how individuals adopt norms as well as rules to overcome social
dilemmas.
In the remainder of this chapter, we will focus on the role of norms
in overcoming social dilemmas—how to represent norms, why they are
important, why norms alone are not sufficient to cope with many collective-action
problems, and how norms may evolve. Then, the rest of this
book digs into the concept of institutions and tries to clarify the meaning
of key concepts, the reason we need to develop a theoretically relevant
way of classifying rules, the way we identify classes of rules used in the
field, and the manner in which rules may be used as tools to affect the
orientations of participants over time.
Norms Fostering Collective Action
Some scholars are hesitant to use the concept of norms because a generally
accepted method for representing norms in formal models does not yet
exist. Sometimes norms are simply used as a casual explanation after observing
behavior that is not consistent with that predicted by noncooperative
game theory. Arguing that “Oh, they must have been using a norm
of reciprocity” as a post hoc explanation of puzzling behavior is not sufficient
in the long run for arriving at empirically warrantable theory.
Representing norms in formal theory and then positing testable propositions
from these theories is not immensely difficult. Many contemporary
theorists add one or more symbols to the payoff function they examine
in a formal game to represent the internal valuation that participants may
place on outcomes that others may receive—other-regarding preferences—or
on actions or outcomes to which an individual assigns an intrinsic
value that differs from its extrinsic value (see Fehr and Ga¨chter 1998;
Frey 1997a, 1997b; Bolton and Ockenfels 2000).
Sue Crawford and I tackled this problem in 1995 in our American Political
Science Review article “A Grammar of Institutions,” which has been
revised and updated as chapter 5 of this book. As we define them, norms
are prescriptions held by an individual that an action or outcome in a
situation must, must not, or may be permitted. Norms can be represented
in formal analyses as a delta parameter that represents the intrinsic benefits
or costs of obeying a normative prescription in a particular situation.
The changes may occur as a result of intrinsic motivation such as pride
  
C HA PT ER FO UR122
Figure 4.3 The Snatch Game with norms.
when keeping a norm or guilt when breaking a norm.17
The delta parameter
may also occur as a result of the action being observed by others leading
to esteem for following a norm or shame for breaking it.
The Snatch Game—with Norms
As an example for the reader, it would be useful to illustrate how one
could explicitly introduce the concept of norms into a formal game—the
Snatch Game—presented in chapter 2. Norms change the internal value
that participants place on an action or outcome in a situation (see figure
4.3). In the initial Snatch Game, participants did not possess norms
against snatching goods. Thus, their internal preference functions could
simply be represented as the value placed on the goods they received (or
as a monotonic function thereof). Given this preference, Household 1
would predict that Household 2 would snatch any goods put out for exchange.
Thus, Household 1 would not offer any goods in the first place.
Let us now assume that Household 2 has a norm against snatching
goods. The preference function for Household 2 could now be represented
as:
u2 = π2 − δb
, where
π2 = payoff obtained by Household 2,
− δb
= decrease in the value of π2 for breaking the norm.
Now, if Household 1 has good reason to believe that Household 2 has a
norm against snatching, and that the size of the norm is greater than the
value of the goods that could be snatched, then Household 1 is likely to
put the goods out for exchange. Both Households are better off than they
would be simply keeping their own commodities.
  
A NI MATI NG IN ST IT UT IO NA L A NA LY SI S 123
Whether the strength of the norm is sufficient to motivate Household
2 to refrain from snatching the commodities depends on the size of the
delta parameter and not simply its presence. In this example, Household
2 would need to suffer a loss of more than five payoff units to offset what
could be obtained by snatching. If the two households were embedded in
a community with an effective observation and gossip network, so that
anyone who snatched goods offered for sale would be likely to be observed
and shamed, the shame of snatching might be enough to lead
Household 2 to prefer an exchange rather than to snatch the goods.18
If
the intrinsic value of the norm is not high enough to compensate for the
potential gain of material payoffs in a Snatch Game, then the behavior
will be the same as in the earlier game. Thus, in addition to specifying
that norms exist, the theorist needs to posit the relative size of the costs
or benefits of following a norm. Norms can definitely change behavior
but may not do so. Whether norms have an impact on behavior depends
both on the strength of the norm and the context of the situation.
For norms against snatching goods offered for exchange to be effective
without any rule enforcement present, a community in which the households
are located needs to be relatively stable. All households must teach
children the prescription against snatching other people’s goods. All
households must share the knowledge that all the other households are
doing this. In other words, the presence, size, and sign of delta parameters
must become common knowledge over time. A small clique of households
who obtain a positive internal reward from the daring act of
snatching goods could unravel years of stability in the exchange of
agreements among the households.19
In most market economies, where
rules are relatively well enforced, shared norms play an important complementary
role in enabling participants to engage in a wide diversity of
economic transactions without relying entirely on external rule enforcers
(Cooter and Ulen 1996).
The Problem of Heterogeneity
More puzzling than finding ways of representing norms in formal analyses
is coping with the wide heterogeneity of norms that individuals learn and
the internal strength that they attach to these norms. As illustrated by the
experiments discussed in chapter 3, a substantial proportion of participants
in social dilemmas is observed to take actions that are more cooperative
than predicted for rational egoists. One must stress, however, that
not all of the participants appear to have intrinsic delta parameters
attached to the actions of trusting others, reciprocating trust, cooperating
to solve collective-action problems, or sanctioning others who are not
  
C HA PT ER FO UR124
cooperative. Nor can we assume a uniform “delta” that everyone who is
not a rational egoist would share.
Different participants have their own orientations and adopt diverse
strategies in the same situation. In other words, we need to assume heterogeneity
of preferences. The classic model of noncooperative game theory
now becomes a special case for modeling situations where one assumes
that individuals do not attach any internal value to following norms or
to the payoffs received by others. Further, one can expect some individuals
to behave in most situations in a manner consistent with this model.
Many new models have been posed in an effort to devise another general
model of human behavior (see Rabin 1993; Ito, Saijo, and Une 1995;
Chan et al. 1997; Levin 1998; Fehr and Ga¨chter 2000a; Bolton and Ockenfels
2000; Gintis 2000a; Casari and Plott 2003). None of these general
models are yet sufficiently well supported by experimental and field data
that we can just substitute a new general model for the old classical general
model. What is exciting, however, is that scholars are developing
careful experimental designs precisely to examine how these competing
models fare in critical experiments (see Cox 2004; Cox and Sadiraj 2004;
Cox, Friedman, and Gjerstad 2004).
Many reasons exist for a concern that all of the explanations for cooperative
behavior can be placed on a new model of individual behavior.
Fundamentally, one has to know key information about the situation before
one can specify the kind of values participants are likely to adopt. As
discussed above in relation to the Snatch Game with norms, the norms
that individuals may adopt in general may or may not be strong enough
to overcome a temptation to break a norm in a particular situation. Further,
for norms such as those leading an individual to be a conditional
cooperator to generate positive returns, some arrangement for clustering
conditional cooperators together is needed. Studies that look at niches,
or spatial relationships among participants, often show that conditional
reciprocity can lead to cooperation to overcome social dilemmas (Laland,
Odling-Smee, and Feldman 2000).
A key requisite for successful cooperation is that conditional cooperators
must be able to find one another—due to either institutional or
spatial connections. If too many rational egoists surround conditional
cooperators, cooperation can just collapse. It is also the case that some
norms are potentially destructive rather than constructive (Boulding
1963). McGinnis (forthcoming) demonstrates for us how individuals engaged
in threat systems can find themselves in ever deepening conflict
situations.
What is important about recognizing that not all individuals in all situations
act like rational egoists is that we can begin to take institutions very
seriously. As Colin Camerer (2003, 117) captured the current situation:
  
A NI MATI NG IN ST IT UT IO NA L A NA LY SI S 125
“Institutional arrangements can be understood as responding to a world
in which there are some sociopaths and some saints, but mostly regular
folks who are capable of both kinds of behavior.” Thus, institutions are
among the tools that fallible humans use to change incentives to enable
fallible humans to overcome social dilemmas. Sadly, however, institutions
can also exacerbate bad outcomes as well as ameliorate them. We need
to recognize that not only are humans complex systems; so are the structures
they build. Thus, we need to really dig in and examine what we mean
by institutions and how they can be used to overcome social dilemmas—a
task that we will begin in chapter 5 and continue through the rest of the
volume. Before we turn to this task, however, we need to examine the
question of how human norms could survive in settings where rational
egoists are initially a large part of a population.
Emergence and Survival of Norms in Evolutionary Processes
While evolutionary theory has been used for years to explain why only
rational egoists can survive in a competitive process, recent developments
in evolutionary theory are coming to a different set of conclusions (Maynard
Smith and Harper 2003; Marcus 2004; Richerson, Boyd, and Paciotti
2002). Contemporary evolutionary theories provide useful ways of
modeling the emergence and survival of multiple strategies in a population
of participants. In a strict evolutionary model, individuals inherit
strategies and do not change strategies in their lifetime. Those carrying
the more successful strategies for a particular environment reproduce at
a higher rate (Axelrod 1986). After many interactions, the more successful
strategies come to prominence in the population.20
Human evolution occurred mostly during the long Pleistocene era that
lasted for about 3 million years to about 10,000 years ago. During this
era, humans roamed the earth in small bands of hunter-gatherers who
were dependent on each other for mutual protection, sharing food, and
providing for the young. Survival was dependent not only on aggressively
seeking individual returns but also on solving many day-to-day collectiveaction
problems. Those of our ancestors who solved these problems
most effectively and learned how to recognize who was deceitful and
who was a trustworthy reciprocator had a selective advantage over those
who did not (see Barkow, Cosmides, and Tooby 1992; Ben-Ner and Putterman
2000a; Brown and Moore 2002). Humans have acquired wellhoned
skills at face recognition and strong abilities to detect cheating.
Research provides evidence that humans keep rough internal accounts—
both in regard to goodwill (McCabe and Smith 2003) and threats (Boulding
1963).21
  
C HA PT ER FO UR126
Evolutionary psychologists who study the cognitive structure of the
human brain speculate that humans do not develop general analytical
skills that are then applied to a variety of specific problems. Humans are
not terribly skilled at general logical problem solving (as any scholar who
has taught probability theory to undergraduates can attest). Rather, the
human brain appears to have evolved a domain-specific, human-reasoning
architecture (Clark and Karmiloff-Smith 1991). For example, humans
appear to use a different approach to reasoning about deontic relationships—what
is forbidden, obligated, or permitted—as contrasted to
reasoning about what is true and false. When reasoning about deontic
relationships, humans tend to check for violations or cheaters (Manktelow
and Over 1991). When reasoning about whether empirical relationships
are true, they tend to use a confirmation strategy (Oaksford and
Chater 1994). This deontic effect in human reasoning has repeatedly
been detected even in children as young as three years old, and is not
associated with overall intelligence or the educational level of the subject
(Cummins 1996).
Recent studies have also examined how strategies are transmitted via
diverse cultural processes and thus can evolve rapidly (Richerson and
Boyd 2002; Henrich 2004; Sussman and Chapman 2004). Cultural transmission
is largely enhanced by language. While evidence is mounting that
some animals have gained simple languages (Kaminski, Call, and Fisher
2004), humans have acquired language skills involving thousands of
words that can be combined, given the rules of a grammar, into an infinite
number of sentences conveying both specific information about a situation
as well as general conceptual and normative information (Bloom
2000). The acquisition of human language provides “a second system of
‘heredity’” (Maynard Smith and Harper 2003, 140). This second system
of heredity links the generations and is able to support rapid cultural
change. “With language, genetic change ceases to be the main basis of
change: history begins” (140).
Developments in evolutionary theory and supporting empirical research
provide strong support for the assumption that modern humans
have inherited a propensity to learn rules and norms similar to our inherited
propensity to learn grammatical rules (Pinker 1994). In their chapter
on the acquisition of language, Maynard Smith and Szathma´ry (1997)
stress the formal similarity between an “action grammar” and a “language
grammar.” Children begin to learn strategies of constructing sentences
and undertaking rule-ordered actions in the world at about the
same time in their development. Which prescriptions are learned, however,
varies from one culture to another, across families, over time, and
with exposure to diverse social norms expressed within various types of
  
A NI MATI NG IN ST IT UT IO NA L A NA LY SI S 127
situations. In most societies, however, the norm of reciprocity—that one
should return cooperative actions of others that benefit self, with similar
efforts to help others—appears to be present to some extent related to
specific activities, locations, and relationships among individuals. The
evidence that individuals do learn norms of reciprocity—including a norm
of punishing others who do not reciprocate cooperation as well as punishing
those who do not punish others—is steadily mounting (Bowles and
Gintis 2004).
The Indirect Evolutionary Approach to Adaptation
through Experience
Recent work on an indirect evolutionary approach to the study of human
behavior offers a rigorous theoretical approach for understanding how
preferences—including those associated with social norms—may evolve
or adapt in shorter time horizons than posited in biological evolutionary
theory (Gu¨th and Yaari 1992; Gu¨th 1995). In an indirect evolutionary
model, players receive objective payoffs (measuring economic or reproductive
success), but make decisions based on the transformation of these
material rewards into their own intrinsic values. Those who place a value
on reciprocity, fairness, and being trustworthy can be thought of as adding
a subjective delta parameter to actions (of themselves or others) that
are consistent or not consistent with their norms. This approach allows
individuals to start with a predisposition to act in a certain way, but it
allows preferences to adapt within a generation (rather than over many
generations) given the objective payoffs received depending on the interpretation
of the evolutionary dynamics.22
Social dilemmas are particularly useful for exploring the indirect evolutionary
approach. An indirect evolutionary approach explains how a mixture
of norm-users and rational egoists would emerge in settings where
standard rational choice theory assumes the presence of rational egoists
alone. In this approach, social norms may lead individuals to behave differently
in the same objective situation depending on how strongly they
value conformance with (or deviance from) a norm. Rational egoists can
be thought of as having intrinsic payoffs that are the same as objective
payoffs since they do not value the social norm of reciprocity.
Conditional cooperators (to add one additional type of player) would
be modeled as placing a positive delta parameter on (1) contributing to a
group outcome in a Commons Dilemma or (2) trusting others when in
the position of an Investor or on being trustworthy when in the position of
a Trustee in the Trust Game. By their behavior and resulting interaction,
however, different types of players are likely to gain differential objective
  
C HA PT ER FO UR128
returns. In a Trust game where players are chosen from a population that
initially contains some proportion of rational egoists and conditional cooperators,
the level of information about player types affects the relative
proportion of rational egoists and conditional cooperators over time.
With complete information regarding the presence and size of a delta parameter,
conditional cooperators playing a trustworthy strategy will more
frequently receive the higher extrinsic payoff, while rational egoists will
consistently receive a lower payoff since others will not trust them.
In indirect evolutionary theory, only trustworthy participants survive
in an indefinitely repeated social dilemma characterized by complete information
about the past actions of all subjects (Gu¨th and Kliemt 1998,
386). If one thinks of this theory as a learning process, those who were
less successful would tend to learn the intrinsic values or delta parameters
of those who had achieved higher material rewards (Bo¨rgers and
Sarin 1997).23
Where a player’s type is common knowledge, rational egoists
would not survive. Conditional cooperators interacting with other
known conditional cooperators will obtain higher payoffs and come to
dominance.
Full and accurate information about all players’ intrinsic preference,
however, is a strong assumption and difficult to achieve. When participants
try to limit who else is involved in overcoming dilemmas to others who
live in a community and who can establish a reputation for being trustworthy,
they may be able to create the conditions needed for the evolution of
conditional cooperators to become the dominant type of player.
If there is no information about player preferences for a relatively large
population, preferences are predicted to evolve so that only rational egoists
survive.24
If information about the proportion of a population that
are trustworthy is known, and the first player has no information about
the trustworthiness of a specific second player, Gu¨th and Kliemt (1998)
derive a prediction that first players will trust second players as long as
the expected return of meeting trustworthy players and receiving the
higher payoff exceeds the expected payoff obtained when neither player
trusts the other. In such a setting, however, the share of the population
held by the norm-using types is bound to decline. On the other hand, if
there is a noisy signal about a player’s type that is at least more accurate
than a random signal, trustworthy types can survive as a substantial proportion
of the population. Noisy signals may result from seeing one another,
face-to-face communication, and various internal accounting mechanisms
that humans have evolved to monitor each other’s behavior. As
we discuss later in this volume, the institutions that humans craft to cope
with diverse problems can affect the kind of information that individuals
obtain about each other’s past actions.
  
A NI MATI NG IN ST IT UT IO NA L A NA LY SI S 129
Evidence Testing the Indirect Evolutionary Theory
Indirect evolutionary theory is able to explain how a mixture of contingent
cooperators and rational egoists would emerge in settings where traditional
game theory predicts that only rational egoists should prevail.
Given the recent development of this approach, direct tests of this theory
are not yet extensive. From the viewpoint of an indirect evolutionary process,
participants in a collective-action problem would begin an interaction
with differential, intrinsic preferences over outcomes due to their
initial predispositions toward norms such as reciprocity and trust. Participants
would learn about the likely behavior of others and shift their behavior
in light of the experience and the objective payoffs they have received.
Several recent experiments provide evidence of these kinds of
contingent behaviors and behavioral shifts.25
In a one-shot, sequential, double-blind, Prisoner’s Dilemma experiment,
for example, the players were asked to rank their preferences over
the final outcomes after they had made their own choice, but before they
knew their partner’s decision. Forty percent of a pool of 136 subjects
ranked the cooperative outcome (C,C) higher than the outcome if they
defect while the other cooperates (D,C), and 27 percent were indifferent
between these outcomes, even though their individual monetary payoff
was higher in the latter outcome (Ahn, Ostrom, and Walker 2003). This
finding confirms that not all players enter a collective-action situation as
pure forward-looking rational egoists who make decisions based solely
on individual outcomes.26
Some bring with them a set of norms and values
that can support cooperation.
On the other hand, preferences based on these norms can be altered by
bad experiences. One set of 72 subjects played 12 rounds of a finitely
repeated Prisoner’s Dilemma game where we randomly matched partners
before each round. Rates of cooperation were very low. Many players
experienced multiple instances where partners declined to cooperate
(Ahn, Ostrom, and Walker 2003). In light of these unfortunate experiences,
only 19 percent of the respondents now ranked (C,C) above (D,C)
while 17 percent were indifferent (ibid.). In this uncooperative setting,
the norms supporting cooperation and reciprocity were diminished by
experience, but not eliminated.27
In another version of the Prisoner’s Dilemma game, Cain (1998) first
had players participate in a “Dictator Game”—in which one player divides
a sum of money, and the other player must accept the division, whatever
it is—and then a Prisoner’s Dilemma game. “Stingy players,” defined
as those who retained at least 70 percent of their endowment in the earlier
Dictator Game, tended to predict that all players would defect in the Pris-
  
C HA PT ER FO UR130
oner’s Dilemma game. “Nice players,” defined as those who gave away
at least 30 percent of their endowment, tended to predict that other nice
players would cooperate and stingy players would defect. Before playing
the Prisoner’s Dilemma game, players were told whether their opponent
had been “stingy” or “nice” in the dictator game. Nice players chose
cooperation in the Prisoner’s Dilemma game 69 percent of the time when
they were paired with other nice players and 39 percent of the time when
they were paired with stingy players.
Finally, interesting experimental (as well as field) evidence has accumulated
that external imposed rules tend to “crowd out” endogenous cooperative
behavior (see Frey 1994). For example, consider some paradoxical
findings of Frohlich and Oppenheimer (1996) from a Prisoner’s
Dilemma game. One set of players—the control group—played a regular
Prisoner’s Dilemma game (some with communication and some without).
A second set of players used an externally imposed, incentivecompatible
mechanism designed to enhance cooperative choices. In the
first phase of the experiment, the second set of participants gained higher
monetary returns than those in the control group, as expected. In the
second phase of the experiment, both groups played a regular Prisoner’s
Dilemma game. To the surprise of the experimenters, a higher level of
cooperation occurred in the control groups that played the regular Prisoner’s
Dilemma in both phases, especially for those who communicated on
a face-to-face basis. The greater cooperation that had occurred due to
the exogenously created incentive-compatible mechanism appeared to be
transient. The authors speculated that removing the external mechanism
undermined subsequent cooperation. Having an effective rule imposed
on them, even though it induced them to improve their outcomes, appeared
to leave the players worse off once it was removed, in comparison
with the players in the control group who relied entirely on face-to-face
communication.
The studies by Fehr and Rockenbach (2003) and Cardenas, Stranlund,
and Willis (2000), discussed in chapter 3, have confirmed the notion that
external rules and monitoring can crowd out cooperative behavior. These
studies typically find that a social norm, especially in a setting where there
is communication between the parties, can work as well or nearly as well
at generating cooperative behavior as an externally imposed set of rules
and system of monitoring and sanctioning. Moreover, norms seem to have
a certain staying power in encouraging a growth of the desire for cooperative
behavior over time, while cooperation that is primarily there due to
externally imposed and enforced rules can disappear very quickly. Finally,
the worst of all worlds may be one where external authorities impose
rules but are able to achieve only weak monitoring and sanctioning. In
  
A NI MATI NG IN ST IT UT IO NA L A NA LY SI S 131
a world of strong external monitoring and sanctioning, cooperation is
enforced without any need for internal norms to develop. In a world of no
external rules or monitoring, norms can evolve to support cooperation. In
an in-between case, a low level of external monitoring discourages the
formation of social norms, while also making it attractive for some players
to deceive and defect, given the low risk of being caught.
Conclusion
The institutional analyst at the beginning of the twenty-first century faces
a challenging task in animating analyses of expected behavior in a wide
diversity of situations. To do a detailed analysis of expected behavior, the
analyst must first ask some crucial questions about the action situation.
First, does the situation generate substantial information about the structure
of the situation itself? Second, do individuals voluntarily enter the
situation in order to compete for valued, objective outcomes such as income,
commodities, winning an election, beating others in a sports tournament?
Third, do those who engage in this type of situation place primary
value on obtaining the immediate objective outcomes of the
situation? If the answers to these three questions lead the analyst to conclude
that the situation is an open competitive situation focused primarily
on objective payoffs, then using the assumptions of a rational egoist to
animate the model is the best strategy. When all of these factors are present,
the theorist is likely to make predictions about behavior and outcomes
that are supported by empirical data.
On the other hand, if the situation is a social dilemma rather than an
open competitive process, I would urge the analyst to animate an initial
analysis assuming that participants hold multiple value orientations and
use strategies ranging from those used by rational egoists to those used by
players who value trust, reciprocity, and equity very highly. The relative
proportion of each type of player that would survive after repetition of
the situation would depend heavily on multiple aspects of the structure
of the situation and the initial distribution of types as affected by the
biophysical world, the rules in use, and the community in which it is
embedded.
If the dilemma involved many individuals located in diverse settings
around the world who have little opportunity to communicate and share
no common rules—like an open-access ocean fishery or the global atmosphere—then
the best predictions and explanations of behavior would be
derived from assuming that most participants are rational egoists. Conditional
cooperators can do little without an appropriate institutional struc-
  
C HA PT ER FO UR132
ture to support their norms. Those who would like to cooperate with
others find themselves unable to do anything but follow the dominant
strategy. Recent evidence related to the massive depletion of fisheries in
the open oceans confirms this prediction (Myers and Worm 2003).
This points to the importance of larger institutions that enable participants
in social dilemma situations to have sufficient autonomy that
they can change the rules that affect their ongoing situations. If individuals
face a social dilemma situation repeatedly and they have the autonomy
to change the rules that structure it so as to enhance the probability
that the proportion of conditional cooperators and willing punishers can
grow over time, many individuals have crafted ingenious institutions that
help them reach mutually productive rather than mutually unproductive
outcomes.
Given the importance of institutions in affecting the structure of a situation
and the population of types of individuals who are most likely to
come to prominence in a particular type of situation, we now need to
tackle a major focus of this book—the rules that humans can use as tools
to fashion the action situations they engage in repeatedly. The biophysical
world and the attributes of a community work together with rules to
constitute action situations that enhance or reduce the likelihood of individuals
reaching better outcomes. Rules are the tools, however, that fallible
humans can use to try to change situations to achieve better outcomes.
In chapter 5, we will focus on how rules, norms, and strategies share
some attributes and differ on others. A high level of confusion exists in
the literature as to what these concepts mean. It is hard to develop better
policies to overcome collective-action problems when scholars disagree
on fundamental terms such as these. In chapter 6, we will then focus on
the question of why we should classify rules themselves. In chapter 7, we
will develop a systematic language for identifying and classifying rules.
In chapter 8, we will illustrate many of the rules actually used in field
settings and develop a theoretical calculus for how appropriators from a
common-pool resource decide to change the rules affecting their interactions.
Chapter 9 will then focus on what types of resource governance
regimes tend to be robust in a changing world.
As scholars and as policy analysts, we need to learn the artisanship of
working with rules so as to improve how situations operate over time.
Human beings are neither all-knowing saints nor devilish knaves. The
institutions they grow up in—families, schools, playgrounds, neighborhoods—differentially
reward or punish them over time so that intrinsic
and extrinsic motivations are learned and developed over time. The situations
they find themselves facing as adults in the workplace and their
community also affect which norms they use and the outcomes they reach.
  
A NI MATI NG IN ST IT UT IO NA L A NA LY SI S 133
When individuals learn the artisanship of crafting rules, they can experiment
and learn to create more productive outcomes (as well as participants)
over time. Learning to craft rules that attract and encourage individuals
who share norms of reciprocity and trustworthiness, or who learn
them over time, is a fundamental skill needed in all democratic societies.
  
  
Part II
F O C U S I N G O N R U L E S
  
  
Five
A Grammar of Institutions
S U E C R AW F O R D A N D E L I N O R O S T R O M
CHAPTER 4 focuses on the challenge that social scientists face in animating
analyses of social situations so as to generate understandings and predictions
that are then tested.1
The need to study holons within holons makes
this a difficult process. Until recently, political economists had at least one
unchanging constant in their analyses—the model of the individual used.
One paid primary attention to the focal action situation and then asked
what participants modeled as rational egoists would do in this situation.
The general strategy recommended in this book is similar but more difficult.
First, one needs to examine the structure of the situation. Then one
asks how boundedly rational, fallible but adaptive individuals would interact
in that situation over time.
In this chapter, we return to the task of analyzing the structure of situations
so as to better model the interaction of actors in those situations.
Earlier chapters establish the basic components of diverse action situations.
Here we turn to analysis of specific kinds of institutional statements
that shape incentives in action situations. Specifically, we elucidate a syntax
that illustrates the similarities and differences between shared strategies,
norms, and rules. These concepts are not clearly distinguished in
much contemporary social science literature. In this chapter, we show
why there is so much confusion and provide tools to clarify the distinct
influences of each kind of institutional statement on human interaction
in diverse action situations. As in other chapters, we illustrate how the
grammar works for institutional analysis that uses various research
approaches.
Parsing Institutional Statements
We view the concept of an institutional statement as a broad term encompassing
three types of statements—rules, norms, and shared strategies.
These statements describe opportunities and constraints that create expectations
about other actors’ behavior. In other words, an “institutional
statement” encompasses a broad set of shared linguistic constraints and
  
C HA PT ER FI VE138
opportunities that prescribe, permit, or advise actions or outcomes for
participants in an action situation. We stress the shared nature of these
concepts. Many written statements have the form of a rule (or a norm or
strategy) but are not known to participants and do not affect behavior.
Such statements are considered rules-in-form rather than rules-in-use
(Sproule-Jones 1993). We concentrate our attention here on rules-in-use.
As the Institutional Analysis and Development (IAD) framework stresses,
these shared linguistic constraints interact with influences from the biophysical
world and those attributes of the community not easily understood
as institutional statements (for example, ethnic heterogeneity) to
shape the structure of the action situation.
The grammar tool allows analysts to distinguish more systematically
between institutional statements that are best understood as attributes of
the community (strategies and norms) and those that are best understood
as rules. This distinction recognizes that rules operate in such a distinct
way in action situations that analysts need to know when an institutional
statement is a rule and when it is not.
We focus on rules for two basic reasons. First, institutional analysts
conducting policy analysis are frequently asked to analyze the impact of
some change in rules—either a change that has already occurred or the
possible impacts of a proposed change. Second, institutional analysts
working to craft solutions to negative outcomes in an action situation
recognize that changes in the rules may be easier or more stable than
attempts to change the situation through changes in the biophysical world
or attributes of the community. The grammar of institutions provides a
tool to help those crafting institutions to ensure that the institutional statements
that they craft are indeed rules, so that they can better predict the
influence that the institutional statement will have on the action situation
and better tap into the institutional strengths that rules provide.
We assume with von Wright (1968) that rules can be expressed using
two basic linguistic forms. One is a generative form: “let there be an X.”
Rules that create positions (e.g., voter, judge, mayor) or organized bodies
(e.g., the U.S. Senate) are generative rules. The other is a regulatory form,
which has a more complex syntax elucidated in this chapter. We return
to a brief discussion of generative rules in chapter 7 when we examine the
challenge of classifying and using rules.
In this chapter, we posit that regulatory rules are one grammatical step
away from norms and two steps away from strategies. Our explicit recognition
of the differences between rules, norms, and strategies clarifies analytical
questions about the similarities among these concepts and about
the existence and origin of institutional rules. Using this grammar, one
can ask clear questions about when strategies or norms evolve into rules
and why.
  
A GR AM MA R O F I NS TI TU TI ON S 139
The grammar of institutions thus provides a syntax for analyzing and
expressing institutional statements that can be used to distinguish systematically
among rules, norms, and shared strategies. It draws on and speaks
to discussions of rules, norms, and institutions from a wide range of the
social sciences. But first, a disclaimer: we do not assume that institutional
statements affecting behavior can always be articulated easily and fully
by participants. Knowledge of institutional statements often becomes habituated
and part of the tacit knowledge of a community (Epstein 2001).
Moreover, we do not assume that all individuals recognize the existence
of an institutional grammar and explicitly use it to formulate institutional
statements. The grammar of institutions is a logical tool that complements
other ways of representing institutional phenomena. It is a useful tool
for summarizing and analyzing the content of institutional statements,
distinguishing between types of institutional statements, and studying the
formation and evolution of institutional statements.
The Syntax of a Grammar of Institutions
In this section, we identify the syntax and components of a grammar of
regulatory rules that can be used in the analysis of institutional statements.
To ease the discussion of the syntax, we provide five examples of
typical institutional statements. We refer to these examples throughout
this chapter.
1. All male, U.S. citizens over eighteen years of age must register with the
Selective Service by filling out a form at the U.S. Post Office or else face arrest
for evading registration.
2. All senators may move to amend a bill after a bill has been introduced, or
else the senator attempting to forbid another senator from taking this action
by calling him or her out of order will be called out of order or ignored.
3. All villagers must not let their animals trample the irrigation channels, or
else the villager who owns the livestock will have to pay a fine.
4. If you use the microwave, you must clean up your own mess!
5. The person who places a phone call, calls back when the call gets dis-
connected.
Let us now turn to the syntax that we will use to analyze these and other
examples.
The general syntax of this grammar includes five components: [ATTRIBUTE],
[DEONTIC], [AIM], [CONDITIONS], and [OR ELSE] where,
A ATTRIBUTES is a holder2
for any value of a participant-level variable that
distinguishes to whom the institutional statement applies. Examples include
  
C HA PT ER FI VE140
eighteen years of age, female, college educated, 1-year experience, or a specific
position, such as employee or chairperson.
D DEONTIC is a holder for the three modal verbs analyzed by von Wright.
These are “may” (permitted), “must” (obliged), and “must not” (forbidden).
I AIM is a holder that describes particular actions or outcomes in the action
situation to which the deontic is assigned. An AIM may include a formula specifying
an amount of action or outcome or a description of a process for an ac-
tion.
C CONDITIONS is a holder for those variables that define when and where
an action or outcome is permitted, obligatory, or forbidden.
O OR ELSE is a holder for the institutionally assigned consequence for not
following a rule.
There are several advantages of using this ADICO syntax. First, elements
from the syntax make up all three types of institutional statements.
Second, elements from the syntax also distinguish among these three types
of statements. Rules contain all five components (ADICO). Norms contain
four components (ADIC) and shared strategies contain three components
(AIC). Third, the syntax provides a format for writing institutional
statements in a consistent manner. Regardless of how institutional statements
are expressed in natural language, they can be rewritten or summarized
in the ADICO format.
All regulative rules can be written as: [ATTRIBUTES] [DEONTIC]
[AIM] [CONDITIONS] [OR ELSE]; all norms can be written as: [ATTRIBUTES]
[DEONTIC] [AIM] [CONDITIONS]; and all shared strategies
can be written as: [ATTRIBUTES] [AIM] [CONDITIONS]. By writing
the statements in a consistent manner, we can then better compare the
institutional statements in use in a variety of settings.3
Finally, since the
addition or subtraction of components switches institutional statements
from one type to another, the syntax provides a tool for analyzing the
evolution of institutional statements from one type to another.
The Syntax Components
We now turn to a discussion of the components of the ADICO syntax.
Understanding the components helps analysts develop common methods
of distinguishing between rules, norms, and strategies to cumulate knowledge
for key questions such as: What difference does it make if the prescription
is a rule or a norm? and What difference does it make if an
institutional statement is a shared strategy or a norm? The syntax does
not eliminate all gray areas. One gray area of particular concern to many
scholars since the original publication of the syntax has been the precise
  
A GR AM MA R O F I NS TI TU TI ON S 141
point at which a norm can be said to have evolved into a rule. The discussion
of the OR ELSE component here clarifies this distinction between
norms and rules to address some of these concerns.
ATTRIBUTES
All institutional statements apply to a subset of participants in an action
situation. The subset can range from one participant to all participants.
A set of ATTRIBUTES establishes the subset of the participants affected
by a particular statement. If individuals make up the participants in an
action situation, the ATTRIBUTES will be individual-level values. Individual-level
ATTRIBUTES include values assigned to variables such as
age, residence, sex, citizenship, and position.4
When the participants governed
by a set of institutions are corporate actors, rather than individuals,
the ATTRIBUTES refer to organizational variables such as size of membership,
geographic location, or who owns the residuals.
In the first example, the relevant ATTRIBUTES are: male, citizen of the
U.S., and over eighteen years old. Some version of the fourth example, the
microwave cleanup statement, can often be found taped to the door of a
community microwave. The beginning of that statement, “If you use the
microwave,” could be parsed as the ATTRIBUTE of “microwave users.”
In the last example, the ATTRIBUTE is the caller who placed the call. The
other examples list no specific attribute. When no specific attribute is
listed, the default value for the ATTRIBUTE component is: all members
of the group.5
This means that the ATTRIBUTE component always has
something in it, even when a specific attribute is not contained in the statement.
Thus, the second example applies to all senators in a legislature, and
the third example applies to all villagers in a particular village.
Within a particular action situation, the ATTRIBUTE component of
institutional statements maps the authority or prescription of an institutional
statement to particular positions or to all positions. In order for
this assignment to work, then, there must be other institutional statements
that assign participants into positions (discussed as boundary rules in
chapter 7). When analyzing an institutional statement with specific positions
in the ATTRIBUTES component, it is important to extract the rules
assigning participants into positions in order to understand how a specific
institutional statement with that position actually works in the action
situation.
DEONTIC
The DEONTIC component draws on the modal operations used in
deontic logic to distinguish prescriptive from nonprescriptive statements
  
C HA PT ER FI VE142
(see von Wright 1951; Hilpinen 1971, 1981). The complete set of DEONTIC
operators, D, consists of permitted P, obliged O, and forbidden F.
The logical relationships among the DEONTIC operators include the
following:
D = P ∪ O ∪ F.
F ∩ P = Ø; O ∩ P = O; and F ∩ O = Ø.
If O then P.
Institutional statements use the operative phrases may, must/should, and
must not/should not to assign these operators to actions and outcomes.
“Should” and “must” are both commonly used to oblige a person to act.
Similarly, “must not” and “should not” both forbid. For the sake of simplicity,
we use “must” and “must not” throughout this chapter in nearly
all examples. However, the deontic terms can be used equally well for
“should” statements. Generally, in everyday language, “must” obligates
someone more strongly than “should,” and “must not” forbids someone
more strongly than “should not.” Later in the chapter we describe how
delta parameters allow more precision in the weight of the Oblige or Forbid
and thus can be used to distinguish between “should” and “must”
when needed in analysis.
The statement that all members may vote assigns the DEONTIC permitted,
P, to the action of voting. The assignment of a DEONTIC operator
to an action [ai] may be represented as [D] [ai], where D stands for P,
O, or F. Similarly, [D] [oi] represents the assignment of a deontic to an
outcome.
DEONTIC OPERATORS RELATE TO THE PHYSICALLY POSSIBLE
Regulation refers to statements about what is presumed to be physically
possible. A person cannot logically be required to undertake an action
that is physically impossible for anyone to do. As expressed by von Wright
in regard to actions: “The notion of ability or can do . . . signifies ability
so to say in its ‘naked form,’ subject only to the restrictions imposed by
the laws of nature (including the limits of man’s innate capabilities of
growth and learning). Within this broad concept of ability (can do) one
can distinguish a narrower concept. When, in this narrower sense, we say
of an agent that he can do or that it is possible for him to do a certain
thing, we mean that his doing of this thing will not violate a set of rules
(norms) or conditions such as, for example, the rules of a certain legal
order or moral code” (von Wright 1966, 33). The DEONTIC operators
are related to each of the components of an action situation. In the above
  
A GR AM MA R O F I NS TI TU TI ON S 143
paragraph, the DEONTIC operators are related to actions. In scope rules,
the operators refer to outcome variables. In information rules, the operators
refer to communication channels.
DEONTIC OPERATORS ARE INTERDEFINABLE
The three DEONTIC operators are interdefinable (von Wright 1968, 143).
In other words, if one of them is taken as a primitive, or the initial starting
point, the other two can be defined in terms of this primitive. For example,
let us use permission P as a primitive. If we are referring to a possible
action [ai], then [P] [ai] would be read: One is permitted to do ai, or one
may do ai. The statement that an act is forbidden [F][ai] can be restated
using P as the primitive as [∼P][ai]. In other words, when an action is
forbidden, one is not permitted to do [ai]. On the other hand, if the negation
of an action [∼ai] is forbidden, one is obliged to take the action. The
statement that an act must be done, [O][ai], can be defined as [∼P][∼ai]. If
an action is obligatory, one is not permitted to not do [ai]. Alternatively,
we could use F as the primitive. Then, P can be defined as [∼F][ai] and O
can be defined as [F][∼ai]. With O as the primitive, P can be defined as
[∼O][ai or ∼ai], while F can be defined as [O][∼ai]. This same interdefinability
exists for prescriptions that refer to outcomes instead of actions. Any
prescriptions with a DEONTIC assigned to some OUTCOME, oi, can be
restated using either of the other two DEONTIC operators.
The first four examples listed above can be restated using F as the primitive
yielding the following:
1. U.S. citizens with [ATTRIBUTES] [F] [∼a1] [CONDITIONS] [OR ELSE]
All male, U.S. citizens, over eighteen years of age are forbidden not to register
with the Selective Service by filling out a form at the U.S. Post Office, or else
face arrest for evading the draft.
2. Senators [All] [∼F] [a2] [CONDITIONS] [OR ELSE]
All senators are not forbidden to move an amendment to a bill after a bill has
been introduced, or else the senator attempting to forbid another senator from
taking this action by calling him or her out of order will be called out of order
or ignored.
3. Villagers [All] [F] [o3] [CONDITIONS] [OR ELSE]
All villagers are forbidden to let their animals trample the irrigation channels,
or else the villager who owns the livestock will have to pay a fine.
4. [Microwave users] [F] [∼a4] [CONDITIONS]
Microwave users are forbidden to not clean up their own mess.
Notice that the fifth example stated earlier has no DEONTIC. The fifth
statement does not state a “must” and “must not” or a “may.” This means
  
C HA PT ER FI VE144
that the strategy of the person, who initiated a telephone call being the one
who “calls back” if the call is interrupted, does not contain a DEONTIC
component. It is a strategy rather than a norm or a rule.
MAKING SENSE OF PERMISSION RULES
The meaning of the DEONTICS Obliged (“must”) and Forbidden (“must
not”) fit well into most conceptions of normative statements. The meaning
of Permitted (“may”) is more perplexing for many scholars. For example,
Susan Shimanoff (1980, 44) concludes that “it is incongruous to talk
of rules prescribing behavior which is merely permitted.” Her conclusion
begs the question that we address here—namely, what does it mean for a
rule or norm to permit an action?
Statements that assign permission (P) to an action influence the structure
of action situations in at least three ways. Most often, rules and
norms that assign permission define constraints on permission by establishing
limited CONDITIONS in which permission exists. In the United
States, consumers above the age of twenty-one may purchase alcohol in
most states. In some states, Sunday liquor laws constrain this permission
further and allow consumers above the age of twenty-one to purchase
alcohol only Monday through Saturday. In effect, these rules establish the
settings in which permission exists and thus forbid the action in circumstances
that do not meet the stated CONDITIONS. A rule that grants
permission to cut trees with a permit from a forestry agency implies that
the absence of a permit renders tree cutting forbidden.
Less often, assigning a “may” to an action is the equivalent of “constituting”
that action (Searle 1969). For example, a statement that an individual
may vote in an election creates an action—voting—that did not
exist before. The rule assigning the permitted action constitutes the action.
These permission rules, then, add new action options to the action
situation that did not exist before.
Finally, some permission rules change the action situation by granting
participants with particular attributes a right to take an action. For example,
voting in some political systems is a legal right. As Commons ([1924]
1968) warns, however, the permission to vote does not operate as an effective
legal right unless at least one other rule assigns a duty to someone
else. Others, who have a duty to recognize a person’s right, are the ones
who are forbidden or required to take actions or affect outcomes.6
The
Voting Rights Act of 1965 included rules that assigned the United States
the duty of ensuring that African Americans were permitted to vote. That
legislation converted a simple permission to vote into a right to vote.
John R. Commons ([1924] 1968), drawing upon the work of Wesley
Hohfeld (1964), further develops the meaning of permission in the form
  
A GR AM MA R O F I NS TI TU TI ON S 145
Capabilities
Right Duty
Exposure Liberty
Party A Party B
Correlatives
L
i
m
i
t
s
Limita
tions
Figure 5.1 Authorized relationships: authority to act. Source: V. Ostrom and
E. Ostrom 1999, 46.
of rights by clarifying the relationships between rights and duty. He argues
that a right depends on a correlative obligation or duty on the part of
others to act in accordance with the right being asserted. Rights are subject
to limits. Limits bearing upon any right define the area of decision
making where a claimant stands exposed. Thus, Commons has defined
the limit of a right as an exposure. The correlative of an exposure is beyond
the limit of a duty. A person who is no longer under duty is at liberty
to act. Thus, the correlative of an exposure is a liberty. When these are
represented in a boxed space, as in figure 5.1, the correlatives indicate the
reciprocal interdependencies among two different legal parties or sets of
legal parties. The limits are applied to each particular party. Taken together,
the diagonal or reciprocal relationship represented by right and
liberty establish the capabilities assigned to both parties. The duty and
exposure establish the limitations assigned to their respective decisionmaking
capabilities. The correlates in figure 5.1 result from rule configurations,
not from a single permission rule. One would need to go beyond
the parsing of single rules to define the full set of rights, liberties, duties,
and exposures of actors within an action situation.7
If a permission rule exists in an action situation, with no rule that assigns
a duty to some position to ensure that those permitted to act are
able to do so, then the permission rule simply establishes that one (or
more) participants are allowed to take this action without having a right
to take it. Others are exposed to this liberty. The U.S. Constitution permits
people who meet certain attributes to run for the office of president. Labor
laws permit individuals over a set age to be legal employees. These rules
  
C HA PT ER FI VE146
establish eligibility. They do not confer “rights” because another rule does
not assign someone the duty to allow anyone meeting those attributes
to be president. Nor do U.S. laws assign duties to anyone to ensure full
employment of those of legal working age. Although these permission
institutions do not confer rights, because they do not assign a duty to
anyone to ensure that the actions can be carried out, these permission
institutions do create liberties. They constrain efforts by others to forbid
a participant to take the permitted AIM. Notice that the rule in the second
example contains an OR ELSE that indicates what happens when senators
attempt to treat the permitted action as forbidden (the offending senator
is called out of order or ignored).8
Even without an established right, the
assignment of permission influences the opportunities and constraints in
the action situation.
REPRESENTING DEONTICS IN FORMAL ANALYSIS
What does including a DEONTIC in an institutional statement imply for
the way a formal game is analyzed? As initially discussed in chapter 4, it
implies that the payoff structure for individuals in situations where they
share common understandings related to prescriptions will differ from
similar situations in which players merely accept a shared understanding
of prudent, rational action. Our way to capture this change in the payoffs
is to add parameters, called delta parameters, to the payoffs related to
conforming or not conforming to a shared prescription. In simple terms,
delta parameters are added to an individual’s payoff to represent the perceived
costs and rewards of obeying (o) or breaking (b) a prescription.
The delta parameters can thus be defined as:
∆ = δo
+ δb
, where
∆ = the sum of all delta parameters
δo
= the change in expected payoffs from obeying a prescription
δb
= the change in expected payoffs from breaking a prescription.
One can further divide these rewards and costs into those that arise from
external versus strictly internal sources of valuation. Thus:
δo
= δoe
+ δoi
and δb
= δbe
+ δbi
, where
e = changes in expected payoffs originating from external sources
i = changes in expected payoffs originating from internal sources.
The distinction between external and internal sources of delta parameters
is similar to that of Coleman (1987) between “internalized norms”
and “externally sanctioned norms.” Delta parameters originating from
external sources are a way to capture the benefits and costs of establishing
a particular reputation (see Kreps 1990). The delta parameters originating
  
A GR AM MA R O F I NS TI TU TI ON S 147
from internal sources can be thought of as the guilt or shame felt when
breaking a prescription and the pride or “warm glow” felt when following
a prescription—particularly if it is costly to follow in a particular
situation (Andreoni 1988; Ledyard 1995). “People who have developed
an internal sanctioning system, for example, feel guilt and shame at behaving
in a deviant way” (Bicchieri 1997, 19).
The analyst may not wish to focus on all four parameters in any particular
analysis. Three of the four delta parameters could be assigned a zero
value in a game-theoretic analysis involving a norm or a rule. In order to
analyze the impact of a DEONTIC on expected outcomes of a game,
however, at least one of the delta parameters must have a nonzero value.
In situations where it is reasonable to assume that all players who break
the prescription feel the same cost, the delta parameters can be modeled
as if they were the same for all players and as if their magnitude is public
information. As discussed in chapter 4, the theorist can also model players
as having different orientations (Harsanyi 1967–68) for situations where
actors react differentially to breaking prescriptions. One player can perceive
the costs of breaking a prescription (δbi
or δbe
) to be high while
another perceives costs (δbi
or δbe
) to be low. Coleman’s (1988) zealot, for
example, is a player with high external deltas for obeying norms (a high
positive δoe
).
If an action is forbidden by a norm and an individual engages in that
action, we expect that player to experience some type of cost represented
by at least one component of delta. If norms indicate that an AIM is
permitted, then we expect that others, who treat that action as if it were
forbidden, will experience some cost. In other words, norms or rules that
forbid or require some AIM will be reflected in a cost parameter to the
individuals to whom the prescription applies. Agreements that establish
permission for an action place the cost parameter on others. Others may
experience a cost if they try to obstruct an individual when a shared norm
indicates that the individual is permitted to take that action.9
The existence of a DEONTIC implies the presence of additional information
that individuals use in developing their expectations about others’
behavior and thus their own best response. If players share a norm, the
payoff structure looks different from the payoff structure for a similar
situation in which the players do not share a norm. The payoffs may even
change enough so that the predicted outcome of the game differs entirely
from that predicted by a similar game that includes no delta parameters.
Uncertainty about whether other actors, who have accepted certain
norms, are present in a situation may be sufficient grounds for changing
the behavior of players. Kreps and colleagues (1982) have analyzed repeated
Prisoner’s Dilemma games where information asymmetries exist
among players concerning the probability that other players will play tit-
  
C HA PT ER FI VE148
for-tat.10
In such games, players who are “perfectly rational” (i.e., the
players’ payoff functions have a zero value for delta components associated
with playing tit-for-tat) will adopt behavior consistent with the norm
for most of the game.
AIM
The AIM is the specific description of a working part in an action situation
to which an institutional statement refers. The description can include
information about a process (filling out a form at the U.S. Post
Office) or a formula (pay $10 per hour worked). In order for an institutional
statement to influence behavior, the AIM must be physically possible
and its negation [,ai] must also be physically possible. An individual
cannot logically be required to undertake a physically impossible action,
and prescribing an action can only influence behavior if it is physically
possible to not do that action. The capability of voting implies the capability
of not voting. Voting for candidate A implies the option of not voting
for candidate A. The AIM sometimes specifies states of affairs in the
world or an outcome instead of an action. Outcomes, like actions, must
also be possible and avoidable to be parts of a well-formed institutional
statement. Moreover, any particular outcome [oi] implies the existence of
its negation [,oi].
In the first example given earlier, the AIM is the action (a1) of registering
for the Selective Service by filling out a form at the U.S. Post Office and
the DEONTIC operator required, O, is assigned to the action for all individuals
with the ATTRIBUTES listed in the rule. In the second example,
the AIM is the action of offering a motion to amend a bill and the DEONTIC
operator is P, or permitted for all senators. The AIM in the fourth
example also includes an action—cleaning the microwave. The third example
assigns the DEONTIC F, or forbidden, to the outcome of livestock
damage. The AIM of the rule does not specify actions that an irrigator
must or may take. The AIM specifies only the forbidden outcome. Villagers
may select any actions that are not forbidden by another rule to keep
their livestock from damaging the irrigation channel. Finally, in the fifth
example of a strategy, the AIM, a5, is the action of calling back.
The AIM often supplies the focus for formal and empirical studies.
Once the set of actions or outcomes is selected, the next step is developing
the analysis of institutional statements related to those AIMs. Scholars
decide to study the impact of institutions on behavior for some subset of
actions or outcomes. Studies of agenda setting and voting institutions, for
example, focus on those actions related to setting agendas and voting
within a particular forum. In studies of voting rules, it becomes particularly
important to specify the formulas in the voting rules that indicate
  
A GR AM MA R O F I NS TI TU TI ON S 149
when the collective body has permission to change the status quo (Shepsle
1989). The AIM component for a majority rule will be different from the
AIM component for a consensus voting rule. Analyses of the impact of
different voting rules will often focus on the impact of different formulas
here. In common-pool resource action situations, the AIM formulas are
particularly important. Formulas in the AIM component of institutional
statements indicate how much a participant may extract (e.g., three units
per acre of land) and how much participants must contribute (e.g., two
days per acre of irrigated land). Analysis of the effectiveness of rules, then,
often turns to analysis of the effectiveness and costs of different formulas
(Yandle and Dewees 2003).
CONDITIONS
CONDITIONS indicate the set of variables that define when and where
an institutional statement applies. For example, the CONDITIONS for a
statement might indicate when a statement applies, such as during certain
weather conditions, at a set time, or at a particular step in some process.
Likewise, the CONDITIONS might indicate where a statement applies,
such as a particular jurisdictional area. If an institutional statement does
not specify particular variables, the default value for the CONDITION is
at all times and in all places covered by that rule, norm, or strategy. Thus,
like the ATTRIBUTE, the CONDITION component always has some
value in it even when the institutional statement fails to overtly specify it.
The CONDITIONS component in the second example indicates when
the prescription applies. After a bill has been introduced, the prescription
of the second rule applies. Thus, any senators may move to amend a bill
after it has been introduced. The strategy in the fifth example applies
when a telephone call is disconnected. The first, third, and fourth examples
do not specify specific CONDITIONS; therefore, we assume that the
rules apply for U.S. citizens, members of the village, and microwave users
under all circumstances.
OR ELSE
The final component of our institutional syntax is the consequence that an
institutional statement assigns to detected noncompliance with the other
components of that statement. In some cases, the OR ELSE specifies a
range of possible punishments if a rule is not followed. Individuals in the
community know that if they violate a rule, they face the probability that
a sanction in a specified range will be applied and that others in a similar
situation face the same range. Only rules include an OR ELSE. This com-
  
C HA PT ER FI VE150
ponent, consequently, plays a crucial role in discerning what a rule is and
how rules differ from other institutional statements.
Sanctions for breaking a rule are a common type of OR ELSE, but the
OR ELSE may take other forms. The OR ELSE might also shift the DEONTIC
assigned to some other action. For example, a violator might be forbidden
to vote or engage in some other action that would otherwise be permitted.
The OR ELSE might also shift the DEONTIC assigned to some activity
for an individual from permitted to obligatory (P to O). The violator might
be required to allocate resources to a public jurisdiction (i.e., pay a fine),
or another actor might be required to check on the violator. Those actions
might be permitted (P) under all other CONDITIONS, but obligatory (O)
when the CONDITION of a violation of the rule is met.
Although the OR ELSE often refers to physical punishments, the OR
ELSE may also involve institutional actions, such as taking away a position
or refusing to accept an amendment as legal. For example, one of the rules
governing the amendment process may state that legislators with [ATTRIBUTES]
[must] [take a particular action] [when voting for an amendment]
[OR ELSE—the amendment fails].
Three qualifications must be met for an OR ELSE to exist. First, the
consequence stated in the OR ELSE must be the result of collective action.
A collective decision must have been made in a relevant collective-choice
arena to determine the consequence.11
Second, the threat in the OR ELSE
component of a rule must be backed by another rule or norm that changes
the DEONTIC assigned to some AIM, for at least one actor, under the
CONDITION that individuals fail to follow the rule. Often the actions
threatened in the OR ELSE are forbidden under most CONDITIONS
(e.g., imposing a fine, incarcerating a citizen, or taking someone’s livestock
and putting them in a village pen). The prescription backing the OR
ELSE makes these actions permitted or required in the CONDITION that
someone breaks a rule. The shift in the deontic is not always from F to P
or O. The OR ELSE might involve forbidding some action that is usually
permitted; a shift from P to F. For example, the OR ELSE might forbid a
government agency from providing a tax incentive to a corporation on
the condition that the corporation violated the specific provisions authorizing
the incentives in the tax code (see Lederman 2003).
Third, in order for an OR ELSE to exist, a prescription must affect the
constraints and opportunities facing an actor or actors with the responsibility
of monitoring the conformance of others. Although the actors who
monitor frequently sanction nonconforming actors, they may only report
nonconformance to someone else responsible for sanctioning. We do not
consider government sponsorship or government backing to be a necessary
condition for a statement to include an OR ELSE. Many self-orga-
  
A GR AM MA R O F I NS TI TU TI ON S 151
nized, communal, or private organizations develop rules that include (1) a
sanction, (2) backed by another rule or norm that changes the DEONTIC
assigned to some AIM for at least one actor if individuals fail to follow
the rule, and (3) a norm or rule (a sanctioning prescription) that affects
the constraints and opportunities facing an actor or actors to take the
responsibility to monitor the conformance of others to the prescription
(a monitoring prescription).
Turning again to the examples of institutional statements listed above,
the first three examples appear to contain an OR ELSE. Of course, we
would need to check to be sure that there are rules or norms regarding
monitoring and sanctioning backing the stated OR ELSE before we would
be sure that the stated sanction fully qualifies as an OR ELSE. For example,
the potential punishment for villagers who let their livestock trample
the irrigation channels qualifies as an OR ELSE only when rules or norms
accepted in that village prescribe others to monitor and to employ the
sanctions defined in the OR ELSE. Without the establishment of positions
with the authority for monitoring and sanctioning, phrases that contain
the words “or else” fail to constitute an OR ELSE that distinguishes an
institutional statement as a rule as defined herein.
Institutional statements with content in the OR ELSE slot, then, are
institutional statements that add information to the action situation about
what will happen if a participant violates the prescription. The Senate
example of a rule indicates that participants who violate the prescription
will be called out of order. A participant in the Senate action situation,
then, knows that if he or she chooses to violate the rule that he or she will
have the consequence of being called out of order or at the very least
having his or her attempts ignored. In effect, the institutional consequence
of breaking this rule is the removal of legal standing. The senator has no
legal standing to restrict another member from moving an amendment in
this simple rule.
Compare the information about the consequences in the Senate action
situation to that in the microwave example. The sign on the microwave
does not provide any specific information about what will happen to a
participant who chooses to ignore the norm and leave a puddle of soup
at the bottom of the microwave. Office workers seeing the sign on the
microwave have only information about their own internal costs or benefits
of following the norm and their beliefs about how others in the office
are likely to respond when they open the door and see the crusty soup
spill there.
The OR ELSE component of a rule is frequently linked to a CONDITIONS
component that specifies the number of times that a rule has been
violated (Dana 2001). The range of sanctions is likely to be lowest when
  
C HA PT ER FI VE152
someone has committed their first offense against a specific rule. A different
rule, specifying a more stringent range of sanctions, will then be applicable
if that individual has committed a second, or a third, or more offenses.
Many rules against drunk driving use this form of graduated
sanctions and increase the sanction for this offense substantially for second
or third offenders.
The content of the OR ELSE affects the very nature of a rule. Rowe
(1989) discusses the difference between a speed limit law with minor sanctions
and a speed limit law with the death penalty as the sanction. The
prescription is the same. The severity of the sanction in the OR ELSE is
the only difference. Yet a speeding law with a death penalty is quite a
different rule than a speeding law with a minor fine. Given the same level
of enforcement, the type of OR ELSE involved may make a substantial
difference on the behavior of participants.12
This example is not meant as
an argument for severe sanctions. Rather, it points out that the OR ELSE
is a fundamental part of rules and that the implications of the sanctions
that are assigned to prescriptions are important for the analysis of institutional
arrangements. This is particularly true for studies of rule compliance
and stability.
Use of the grammar thus far suggests that the OR ELSE component of
the grammar has been the most challenging for scholars to apply, but
the discussions of determining whether this component exists have led to
productive wrestling about the important theoretical differences between
rules and norms in action situations.13
The presence of an OR ELSE, then,
is absolutely crucial to the effective use of the grammar for analysis of
rules. Identifying when this component exists distinguishes when an institutional
statement includes “rule information.” These statements, then,
influence the action situation so that one would describe the situation as
“rule-governed” or “shaped by a rule.” Thus, the OR ELSE clarifies the
difference between institutional statements that influence the action situation
as norms or strategies alone and those that can influence the action
situation as rules.
Applying the Grammar
No one will want to spend time learning the intricacies of the grammar
of institutions developed herein without a sense that it is useful for at
least some purposes. In the remainder of this chapter, we illustrate initial
applications of the grammar to three broad endeavors: (1) game-theoretical
analyses, (2) the synthesis of diverse theoretical approaches to the
study of rules and norms, and (3) learning about norms and rules in empir-
  
A GR AM MA R O F I NS TI TU TI ON S 153
ical research. The remaining chapters of the book will dig into these applications
still further.
Using the Grammar in Game-Theoretical Analyses
Our first application demonstrates the grammar for a very simple game.
By using a game familiar to most contemporary scholars, we illustrate the
analytical usefulness of the working parts of the grammar. The concept of
a strategy, as developed in the institutional grammar, is the same as the
concept of strategy used in contemporary game theory. Thus, a game-theoretic
analysis of a situation in which there are no norms or rules would
not differ in any way from current usage. The Snatch Game, first presented
in chapter 2, is such an analysis. To incorporate the syntax into formal
analyses of behavior, the payoffs for actions governed by norms alone need
to include delta parameters that capture the DEONTIC (as illustrated in
chapter 4), and for rules they must also specify the institutionally assigned
consequence defined in the OR ELSE (as we now illustrate).
If the enforcing players are brought into the analysis, the enforcing
player would have a delta parameter assigned to the action of “not sanctioning,”
since there is a norm or rule that prescribes sanctioning. If the
OR ELSE is backed by a rule, then we expect the payoffs for sanctioning
or not sanctioning to include delta parameters and a variable representing
the cost of the sanction defined in the OR ELSE of the sanctioning rule.
A game that includes the enforcement players must also specify the monitoring
institution influences on the monitor. Once a monitoring norm or
rule comes into play, the delta parameters and possibly the OR ELSE
parameters become part of the monitors’ payoff formulas.
If it is costly to monitor the actions of others and/or to impose sanctions
on them, those assigned these tasks may not be motivated to undertake
these assignments unless (1) the monitor or sanctioner face some probability
of themselves being sanctioned for not monitoring and/or sanc-
tioning,14
(2) social pressure to monitor or sanction is large and is salient
to the monitor and sanctioner (large δoe
and δbe
), (3) the monitor or sanctioner
hold some strong moral commitment to their responsibilities (large
δoi
and δbi
), or (4) the payment schemes for the monitor or sanctioners
create prudent rewards high enough to offset the costs. When norms back
an OR ELSE, monitoring and enforcement rests solely on the value of the
normative delta parameters and on the payment schemes for the monitor
and sanctioner (i.e., Are the monitors and sanctioners paid? If so, is payment
a set fee regardless of the number of defectors they catch, based on
the number of defectors caught and punished, or some mixture of a set
fee and a commission?).
  
C HA PT ER FI VE154
COLLECTIVE-ACTION PROBLEMS
To illustrate applications of the syntax to ongoing research, our discussion
of applications uses shared strategies, norms, and rules related to an abstract
social dilemma problem rather than the five examples discussed
earlier in this chapter or the Commons Dilemma discussed in chapter 3.
The scholarly discourse about social dilemmas and their solution through
collective action has involved terms such as common understanding,
shared beliefs, scripts, norms, rules, procedures, institutions, informal
rules, informal institutions, conventions, internal solutions, external solutions,
as well as a wide diversity of highly technical terms related to particular
solution theories.
Collective-action problems can be represented by many different game
structures (see discussion in Taylor 1987; E. Ostrom, Gardner, and
Walker 1994; Gintis 2000b). Because almost all social scientists know the
Prisoner’s Dilemma (PD) game well, however, we can more easily jump
into existing debates and rely on extensive earlier work. We start with a
simple two-person Prisoner’s Dilemma game and use the ADICO format
to illustrate differences in the research issues, the game structures, and
the predicted outcomes that arise from (1) changed expectations of other
players’ behavior only (AIC statements); (2) changed normative views of
the appropriate actions to be taken or the adoption of norms (ADIC statements);
and (3) changes in the rules (ADICO statements). Table 5.1 summarizes
the institutional and payoff characteristics of four games based
on a two-person Prisoner’s Dilemma situation. The first game is the base
two-person Prisoner’s Dilemma game. The shared strategies game adds a
set of shared strategies that equate to the grim trigger strategy. The norms
game adds a cooperating norm to the base situation. The rules game adds
a cooperating rule, a monitoring norm, and a sanctioning norm to the
base Prisoner’s Dilemma game. These four examples represent only one
way to add the ADICO statements to a two-person Prisoner’s Dilemma
game. We do not develop a new solution theory. Rather, our effort is to
illustrate how an application of the ADICO syntax distinguishes between
three structural adjustments to social dilemma situations.
In figure 5.2, we present both the extensive and the normal forms of
the base two-person PD game. We make only the assumption that the
payoffs are related in the following way: 1 > c > d > 0. Both players are
better off choosing D, no matter what the other player chooses, so D is the
dominant strategy for both players. The Commons Dilemma discussed in
chapter 3 is frequently represented as an N-person Prisoner’s Dilemma
game under particular cost and benefit assumptions.
The game-theoretic solution to this game, if played only once, is for
both players to choose D and receive d instead of the more desirable c
  
A GR AM MA R O F I NS TI TU TI ON S 155
Table 5.1
Game summaries
Institutional statements Payoffs
Base game
NONE (physical world) Player 1 or 2
C = c if other C
= 0 if other D
D = 1 if other C
= d if other D
Shared strategies game
AIC statements Player 1 or 2
[All players] [ ] [C] [first round] [ ] C = c + t (c) if other C
[All players] [ ] [C] [if all C in previous = 0 + t (d) if other D
round] [ ] D = 1 + t (d) if other C
[All players] [ ] [D] [all rounds after a D] [ ] = d + t (d) if other D
Norms game
ADIC statement Player 1 or 2
[P1* and P2] [must] [C] [always] [ ] C = base game payoffs + δoi
+ δoe
if P3 ⇒ M**
= base game payoffs + δoi
if P3 ⇒ ∼M
D = base game payoffs − δbi
− δbe
if P3 ⇒ M
= base game payoffs − δbi
if P3 ⇒ ∼M
Player 3
M = E if (P1 and P2) ⇒ C
= R − E if (P1 or P2) ⇒ D
∼M = 0
Rules game
ADICO statement Players 1 and 2
[P1 and P2] [must] [C] [always] [f] C = norm game payoffs
ADIC statements D = norm game payoffs + f if (P3 ⇒ M) and
[P3] [must] [monitor] [always] [ ] (P4 ⇒ S)
[P4] [must] [impose f on defector] [when = norm game payoffs if (P3 ⇒ ∼M) or
P3 reports a D] [ ] (P4 ⇒ ∼S)
Player 3
M = norm game payoffs + δo
m
∼M = norm game payoffs − δb
m
Player 4
Only plays if P3 ⇒ M
S = δo
s − Es
∼S = − δb
s
Source: Crawford and Ostrom 2000, 134.
* P1 refers to player 1, and so on.
** (P3 ⇒ M) indicates that player 3 chooses M.
  
C HA PT ER FI VE156
BASE GAME
C
D
0,1
d,d
c,c
1,0
C D
22
1
C D C D
C D
c
c
d
d
0
1
1
0
Figure 5.2 Base game. Source: Crawford and Ostrom 2000, 143.
that they could have received if they had both chosen C. In the Commons
Dilemma game, this would mean that subjects would overinvest rather
than investing at the optimal level. Even if repeated for a finite number
of times, the solution is for both players always to choose D.
SHARED STRATEGIES
Predictions that individuals will select C rather than D in a PD game based
on shared strategies rely upon changes in players’ expectations about each
other’s future behavior. In order to incorporate those expectations into
formal analysis, we use an indefinitely repeated version of the base game.
By making the game repeated, we can include future expected payoffs as
part of a player’s calculation at any one round. It is now logically possible
for individuals to adopt shared strategies involving plans of action to cooperate
in the first round, and to defect for all periods thereafter, if the
other player defects in a prior round.
Figure 5.3 illustrates the extensive form of an indefinitely repeated
game with a set of shared strategies that create the famous grim trigger:
“all players cooperate in each round of the game or else all other players
will defect for the rest of the game.” Both players cooperating in every
  
A GR AM MA R O F I NS TI TU TI ON S 157
22
1
C D C D
C D
c+t(c)
c+t(c)
0+t(d)
1+t(d)
1+t(d)
0+t(d)
d+t(d)
d+t(d)
t = number of expected future rounds
t( ) = expectation of payoffs from future rounds of the game
Institutional Statements
Shared Strategies
[All Players] [C] [First Round]
[All Players] [C] [All Rounds in which All Players Play C in the Previous Round]
[All Players] [D] [All Rounds after a D]
Figure 5.3 Repeated game with shared strategies. Source: Crawford and Ostrom
2000, 144.
round is the predicted result if and only if c + t(c) > 1 + t(d) where t is the
expected number of future rounds, assuming that players do not discount
future payoffs.15
Whether all C is the predicted outcome to this game
depends on the relative size of 0, d, and t, all features of the physical
world.
Shared strategies, even the crude trigger strategy, are institutional statements.
They require shared understanding. If all players do not consider it
prudent to defect for all rounds after someone initially defects, the trigger
strategy is not shared and will not work. Little empirical evidence exists
that individuals share a belief in the prudence of a grim trigger (E. Ostrom,
Gardner, and Walker 1994). Herein lies the frailty of the grim trigger as
a resolution of PD situations.
The words “or else” in the grim trigger shared strategy might cause one
to wonder whether it is a rule. The shared strategy with a grim trigger is
not a norm or a rule, using the ADICO syntax because there is no DEON-
  
C HA PT ER FI VE158
TIC and no OR ELSE here. If the advice to cooperate to avoid a trigger
response is not discussed in terms of obligation nor backed by monitoring
and sanctioning institutions, the massive defection that is threatened by
the trigger can only be a prudent response of a player’s defection.
The shared strategy with a grim trigger would be a rule, using the
ADICO syntax, if we were to assume that there were a prescription to
play the trigger strategy that is in turn backed by another rule or norm
that changes the DEONTIC assigned to some AIM for at least one actor,
and there is an additional institutional statement that affects the constraints
and opportunities facing an actor to monitor the conformance of
others. The institutional statements that could back a trigger rule might
be a sanctioning rule, such as: “all other players must defect for the rest
of the game when one player defects in any of the rounds OR ELSE the
other players face the probability of a further sanction” and a monitoring
norm that “all players must monitor all other players.” Notice that the
sanctioning rule changes the DEONTIC assigned to C from obligatory to
forbidden in the CONDITION of a defection in the prior round.
NORMS
Predictions that individuals will select C rather than D in a PD base game
based on norms rely upon changes in players’ payoffs because of the addition
of at least one delta parameter to the players’ payoffs. Figure 5.4
illustrates a game in which the base PD structure has been modified by
the addition of delta parameters in the payoffs for players 1 and 2. In
order to make the discussion more applicable to situations with more than
two players, we add a third player, a Monitor who chooses to monitor
(M) or not to monitor (< M) and assume that external reinforcements for
obeying or breaking a norm occur only when the Monitor reports the
defection. In the simple two-person game, this assumption is usually not
necessary; players 1 and 2 know whether the other player cooperated by
simply looking at their own payoffs. As soon as the number of players in
a PD is larger than two, however, identifying who defects is no longer
trivial. To make the game more applicable to common’s situations with
multiple actors, we “blind” the actors here and assume that defection
becomes common knowledge only when the Monitor does his or her job.
Consequently, the negative external effects of breaking the norm only kick
in when the Monitor chooses to monitor.
In this game, the Monitor is motivated solely by prudential rewards
associated with discovering defection and not by any normative motives
(delta parameters). In other words, we do not assume a monitoring norm.
The absence of a monitoring norm is appropriate here because the game
  
A GR AM MA R O F I NS TI TU TI ON S 159
Institutional Statements
22
1
C D
M ~M
C D
R= reward for detecting defector
E = expense of monitoring
Norm
[Players 1 & 2] [Must] [C] [Always]
3 3
c+*oi+*oe
c+*oi+*oe
!E
c+*oi
c+*oi
0
0+*oi+*oe
1!*bi!*be
R!E
0+*oi
1!*bi
0
C D
M ~M M ~M M ~M
1!*bi!*be
0+*oi+*oe
R!E
1!*bi
0+*oi
0
d!*bi!*be
d!*bi!*be
R!E
d!*bi
d!*bi
0
Payoffs to Monitor (Player 3) Deltas
*oe= external changes in payoffs from obeying prescription
*oi = internal changes in payoffs from obeying prescription
*be= external changes in payoffs from breaking prescription
*bi
= internal changes in payoffs from breaking prescription
3 3
Figure 5.4 Game with a norm and monitoring. Source: Crawford and Ostrom
2000, 145.
represents the addition of a cooperating norm only, not a cooperating
rule. Freelance reporters are an example of this type of monitor. They
receive payment for detecting and reporting nonconformance with accepted
norms. Rewards include fees for stories accepted and increased
probabilities of receiving prizes for good reporting. It is, of course, possible
to illustrate the addition of norms without a player who is assigned a
specialized role as a Monitor by simply assuming that the existing institutional
statements allow players to monitor each other (Weissing and Ostrom
1991a, 1993). To do this, however, one needs to model a sequential
structure, which introduces more complexity than we desire in this initial
application.
In the game represented by figure 5.4, predictions about players’ strategies
depend on the relationships among the original payoffs in the base
game, the added delta parameters, and the benefits the Monitor receives
for reporting nonconformance. This game has many equilibria. Assuming
that all of the delta parameters are symmetric (players 1 and 2 have the
same values for each delta parameter) and that the sum of the external
parameters is greater than the sum of the internal parameters (i.e., the
social pressure to follow the prescription is greater than the internal pres-
  
C HA PT ER FI VE160
(I)
All D
All M
(IV)
All C
All ~M
C
~M
R
E
R=E
0 L'L
(II)
All D
All ~M
)
1!c
L: ) = 1!c
L': *oi+*bi = 1!c
(III)
Mixed (C/D)
Mixed (M/~M)
Figure 5.5 Equilibrium diagram: game with a norm and monitoring. Source:
Crawford and Ostrom 2000, 136.
sure), four equilibrium regions exist as shown in figure 5.5. The vertical
axis is the ratio of the Monitor’s reward for detecting defection to the
expense of monitoring (R/E). The vertical axis is divided into regions
above and below the point at which the reward to the Monitor equals the
expense (R = E).
The reward is higher than the cost above this point and lower than the
cost below this point. The horizontal axis on figure 5.5 is the ratio of the
sum of all delta parameters to the advantage of defecting (1 − c). As one
moves to the right, the size of the delta parameters relative to the advantage
of defection increases. At point L, the sum of the delta parameters
equals the advantage of defection when the other player cooperates (δoi
+
δoe
+ δbi
+ δbe
= 1 − c). Point L′ on the horizontal axis represents that point
where the internal deltas just equal the advantage of defection (δoi
+ δbi
=
1 − c), and thus the area to the right of L′ represents the region where the
internal deltas offset the advantage of defecting. When (c + δoi
) > (1 − δbi
),
it is clear that both players will choose C (equilibrium region IV in figure
5.5). This is the case in which C is the dominant strategy because the
internal costs and rewards for players 1 and 2 are sufficiently high.
  
A GR AM MA R O F I NS TI TU TI ON S 161
Such a norm could be shared in a stable, small community for some
time but would be highly vulnerable to immigration from other communities
where individuals did not share the same norm. A few instances of
individuals taking actions contrary to the norm would generate a cascade
of responses. Unless there were a forum in which the importance of
keeping the norm could be discussed with newcomers so they could learn
the importance of this norm before it disappeared, communities that regulate
resources using only norms may find themselves exposed to substantial
changes in regularized behavior when exposed to considerable
inmigration.
In region III, both players 1 and 2 select a mixed strategy between C
and D.16
As one moves from left to right in region III, the sum of the delta
parameters increases and thus the probability that players 1 and 2 assign
to selecting C increases. In this region, the Monitor also selects a mixed
strategy. The relative amount of the Monitor’s reward decreases as one
moves from north to south. Since the Monitor receives a reward only if
defection is detected, the probability of obtaining a reward decreases as
one moves from west to east because the probability of defection decreases.
Thus, as one moves from “northwest to southeast” in this region,
the probability assigned to M decreases. The combined effect is that the
Monitor has the least incentive to monitor when relative rewards are low
and the probability of defection is low (in the southeast corner) and the
greatest incentive to monitor when the relative rewards are high and the
probability of defection is high (in the northwest corner of this region).
In equilibrium regions I and II of figure 5.5, D remains the dominant
strategy for players 1 and 2, as it was in the base game, but for different
reasons. In region II, the expenses of monitoring are higher than the expected
reward of the Monitor. Thus, the Monitor will choose < M in
region II. Since < M is the dominant strategy, players 1 and 2 need
not consider the external cost parameters (δoe
and δbe
). Given that the
internal deltas are relatively low in relation to the advantage of defecting
[(δoi
+ δbi
) < (1 − c)], D is the dominant strategy for players 1 and 2.
Region I, on the other hand, represents a socially perverse outcome
whereby players 1 and 2 always defect because the advantage of defecting
(1 − c) is greater than the sum of all delta parameters. At the same time,
the Monitor has a dominant strategy of M because the rewards received
from detecting defection exceed the monitoring costs and are guaranteed
to occur, assuming perfect detection, because players 1 and 2 face the
dominant strategy of D.
This analysis demonstrates that simply introducing norms or monitoring
is not sufficient to change predicted results in a PD base game, and
that a change in predicted results is not always socially beneficial. The
only equilibrium regions where players 1 and 2 select a pure strategy of
  
C HA PT ER FI VE162
cooperating is when internal norms generate high internal cost parameters
relative to the advantage of defecting. The presence of a Monitor who
is motivated to select a mixed strategy boosts the level of cooperation in
one region. There is no region where the actions of the Monitor totally
prevent defection. Moreover, the lower the probability of defection, the
higher the monitoring rewards (R) need to be in order to offset the reduced
probability of receiving the reward. These results hold because the
reward to the Monitor comes only if there is defection to be reported.
Changes in the assumptions about the rewards to the Monitor substantially
affect the outcomes. If the Monitor is rewarded specifically for monitoring,
regardless of whether defection is discovered, there are two additional
equilibrium regions. In these regions, the reward for simply
monitoring is greater than the costs of monitoring. In one of these, the
result of the choice of a pure strategy to monitor is to make C a dominant
strategy for the two players. In the other, a pure strategy to monitor is
combined with players 1 and 2 both adopting mixed strategies.
If one wanted to analyze the incentive structure found in many field
settings where monitors are hired as external, disinterested guards, one
could change the game so that the Monitor receives a salary regardless of
whether he or she detects defection or shirks. In such a setting, the Monitor
has little incentive to monitor and thus the rate of cooperation depends
heavily on the size of the internal delta parameters for players 1 and 2.
Empirical studies and formal models suggest several other motivational
schemes for monitors. Some motivate monitors by embedding them in a
series of nested institutions that reward monitors who actively and reliably
monitor with positive returns from the increased productivity that the
rules generate (see Milgrom, North, and Weingast 1990). Monitors may
also be direct participants in ongoing relationships where efforts are made
to keep monitoring costs low, to reward one another for monitoring, and
to ensure that monitors participate in the greater returns that all achieve
when temptations to defect are reduced. In such situations, monitors may
achieve sufficient benefits from monitoring that they induce a high level of
conformance (but never 100 percent) in an isolated system without recourse
to central authorities (Weissing and Ostrom 1991a, 1993).17
RULES
Predictions that individuals will select C rather than D in a PD base game
based on rules rely upon (1) changes in players’ payoffs because of the
addition of at least one delta parameter to the players’ payoffs, (2) the
addition of an institutionally assigned consequence for breaking a rule,
(3) the possibility of detection, (4) at least one player who has authority
to monitor, (5) at least one player who has authority to impose the OR
  
A GR AM MA R O F I NS TI TU TI ON S 163
22
1
C D
M ~M
Institutional Statements
Rule: [Players 1 & 2] [Must] [C] [All Rounds] [OR ELSE f]
3 3
c+*oi+*oe
c+*oi+*oe
!Em+*om
0
c+*oi
c+*oi
!*b
m
0
0+*oi
1!*bi
!*b
m
0
C D
M ~M M ~M M ~M
1!*bi
0+*oi
!*b
m
0
d!*bi!*be!f
d!*bi
!*be
!f
Rm!Em+*o
m
d!*bi
d!*bi
!*b
m
0
Rm
Em
*o
m
*b
m
Payoffs to Monitor (Player 3)
3 3
Payoffs to Sanctioner (Player 4)
Es
*o
s
*b
s
Deltas for Player 1 or 2
*oe
*oi
*be
*bi
OR ELSE for Player 1 or 2
f
4
S ~S
0+*oi+*oe
1!*bi!*be!f
Rm!Em+*o
m
*o
s!Es
0+*oi
+*oe
1!*bi
!*be
Rm!Em+*o
m
S ~S
4
1!*bi
!*be
!f
0+*oi
+*oe
Rm!Em+*o
m
*o
s!Es
1!*bi
!*be
0+*oi+*oe
Rm!Em+*o
m
!*b
s
S ~S
!*b
s *o
s!Es
4
d!*bi
!*be
d!*bi!*be
Rm!Em+*o
m
!*b
s
C D
= reward for detecting defection
= transaction cost of monitoring
= change in payoff from obeying
monitoring norm
= change in payoff from breaking
monitoring norm
= expense of sanctioning
= change in payoff from obeying
sanctioning norm
= change in payoff from breaking
sanctioning norm
= external changes in payoffs from
obeying prescription
= internal changes in payoffs from
obeying prescription
= external changes in payoffs from
breaking prescription
= internal changes in payoffs
from breaking prescription
= fine for Defection
Monitoring Norm: [Player 3] [Must] [Monitor] [Always]
Sanctioning Norm: [Player 4] [Must] [Impose f on defecting player] [When Monitor Reports D]
Figure 5.6 Game with a rule. Source: Crawford and Ostrom 2000, 146–47.
ELSE, and (6) the base game payoffs. In figure 5.6, we present a game
that shifts the norm game in figure 5.4 to a game with a rule backed by
two norms. The rule that structures this game is: [Players 1 & 2] [must]
[Cooperate] [Always] [OR ELSE f]. The rule adds a fine (f) to the payoffs
for players 1 and 2 for D if their defection is monitored and sanctioned.
This rule is backed by both a monitoring norm ([Player 3] [must] [Monitor]
[Always]) and a sanctioning norm ([Player 4] [must] [Impose f on a
Player] [When Player 3 reports that a Player has defected]). The addition
of a monitoring norm adds delta parameters to player 3’s payoffs. The
sanctioning norm adds another player, player 4. Player 4 (the Sanctioner)
  
C HA PT ER FI VE164
faces a cost of sanctioning and receives delta parameters as payoffs.18
In
other words, the Sanctioner in this game is a volunteer who is “rewarded”
solely by normative interests.
A wide variety of mixed-strategy equilibria are possible that depend on
(1) the relative expected value of the fine and the relative size of the delta
parameters originating from external sources for players 1 and 2, (2) the
relative strength of the reward and deltas associated with monitoring to
the costs of monitoring, and (3) the value of the deltas associated with
conforming with the sanctioning norm minus the costs of sanctioning as
compared to the value of the delta parameters for not sanctioning. Even
in a rule-governed game, if Monitors are not motivated to monitor and
Sanctioners are not motivated to sanction, cooperation rests substantially
on internalized norms of the players.
We set aside the tasks of analyzing the many possible equilibrium regions
and focus here on the simpler task of establishing conditions for
equilibria in which players 1 and 2 always cooperate. The rule adds a
Fine and a new player, yet the Monitor still plays a crucial role. The parts
of the game that come from the OR ELSE (the fine and the sanctioning
norm) do not even enter the game on branches in which the Monitor
chooses −M.19
As in the norm game, we assume that players 1 and 2 do
not see each other’s choices, and that the external delta components occur
only when player 3 monitors. In order to be sure that the Monitor will
choose M when players 1 and 2 are expected to choose C, the sum of
the delta parameters for the Monitor must be higher than E, because the
Monitor will not expect to receive rewards (R) for detecting defection. As
in the base game with norms, in the absence of monitoring, players 1 and
2 will select only C as a pure strategy when their internal delta parameters
are high enough to offset the advantage of defecting when others cooperate.
Thus, the selection of C as a pure strategy depends either entirely on
internal sources of normative constraint for players 1 and 2 (δoi
and δbi
),
or it depends on the configuration of the size of the normative constraint
for the Monitor (om and bm) relative to the monitoring costs, the sum of
internal delta parameters for players 1 and 2, the expected value of the
external delta parameters, and the expected value of f. This latter configuration
matters only when the normative constraint for the Monitor is
higher than the monitoring cost. Cooperation can be a pure strategy only
when either of the following conditions is met:
1. δoi
+ δbi
> 1 − c
2. [(δo
m + δb
m ) / E ] > 1 and
[(δoi
+ δbi
) + (p(M)* (δoe
+ δbe
) + (p(S)* f ) ) > 1 − c ].
The first condition is the same as in the game in figure 5.4. The second
condition was not possible in the earlier game as the Monitor was not
  
A GR AM MA R O F I NS TI TU TI ON S 165
motivated by a norm. Defection by both players continues to be an equilibrium
of this game as well as of all the games considered so far.
In settings where players in the base game develop strong internalized
norms (high internal delta parameters), the presence of even a lowto-moderate
f may be sufficient when combined with Monitors and Sanctioners
who are motivated to do their job, to encourage a high rate of
cooperation. Consequently, the effect of f depends both on its relative
size, the size of the deltas for players 1 and 2, and the behavior of the
Monitor and the Sanctioner. If players 1 and 2 expect the Monitor or
the Sanctioner to break their respective norms, then the expected probability
of S (the Sanctioner choosing to sanction) tends toward zero and f
drops out of the decision calculus for players 1 and 2. In order for the
Monitor and the Sanctioner to be motivated to do their jobs, their value
of following the monitoring and sanctioning norms has to be greater
than the relative cost of doing their jobs. In the case of the Sanctioner, the
value of the delta parameters needs to be greater than the cost for imposing
that sanction: (δo
s + δb
s) > Es.20
Clearly, recognizing rules in formal
analysis of dilemma situations does not automatically “solve” the dilemma
and end analysis. Instead, adding rules suggests a whole new set
of research questions.
FURTHER RESEARCH QUESTIONS
The game in figure 5.6 is intended to stimulate further formal and empirical
research. It suggests questions such as: How do changes in the level
of internalization of rules (δbi
and δoi
) affect the levels of monitoring and
sanctioning required to bolster cooperation at given levels of social pressure
(δoe
and δbe
)? What size do external delta parameters need to be in
order to ensure cooperation at various rates of monitoring and sanctioning
with a given value of f that is less than the advantage of defection?
How do the incentives to monitor and sanction differ if we assume that
players 3 and 4 are the same person? And, what are the empirical equivalents
of delta parameters and the external fine in similar situations?
The disjunction between theoretical predictions of complete free-riding
in PD situations and the rates of cooperation in case studies and laboratory
experiments have provoked much intellectual ferment and development
(Ude´hn 1993; Camerer 2003; Boyd et al. 2003). The syntax facilitates
empirical analysis of the components in institutions that foster
cooperation in dilemma situations by providing a language with which
to discuss differences between institutions in various settings and changes
in institutions over time. When an ADICO component is added to an
institutional statement, the statement has changed from one type of institutional
statement to another. Similarly, when a component drops out,
the type of the institutional statement changes. Key questions then in-
  
C HA PT ER FI VE166
clude: What are the processes that lead to such additions or deletions
(Burns and Dietz 1991; Opp 1982)? Do the additions or deletions increase
or decrease the level of cooperation (Orbell, van de Kragt, and Dawes
1991)? What aspects of the content of the components explain differences
in the level of cooperation (e.g., the level of the punishment in the OR
ELSE [Rowe 1989], the inclusiveness of the ATTRIBUTE)? A focus on
the presence or absence of components, the content of components, and
the source of the components promises to be more replicable by scholars
than some of our current dialogue that uses the distinctions of “formal
and informal” or “internal” and “external.”
Using the Grammar for Synthesis
Once we have developed a syntax for expressing institutional statements
in a systematic fashion, we can better address a number of important
theoretical questions. The above analysis of social dilemmas is just one
example. In any science, however, understanding what others have already
discovered is an important part of research. Synthesis and discovery
go hand in hand.
DISENTANGLING FORMAL LAWS AND INFORMAL INSTITUTIONS
Synthesizing findings from the different subfields that relate to each type
of institutional statement is one important task for those interested in
institutions. The ADICO syntax and its game applications can be used to
help disentangle discussions of formal laws, informal institutions, and
ordered behavior. Written laws or procedures often conform to the
ADICO syntax requirement for a rule, yet before analyzing the law as a
rule in use, one would first examine whether those listed in the ATTRIBUTES
share an understanding of the law, and whether shared understanding
also exists for the related monitoring and sanctioning institutional
statements. As Cooter (1994) stresses, the complexity of modern
economies is so great that centralized law creation cannot effectively cope
with the need to achieve normative regulation among communities of
individuals who repeatedly face collective-action problems (see also Ellickson
1991). Thus, individuals frequently adopt norms or rules without
enacting them as written law. These prescriptions are sometimes called
“informal” institutions, but they may well be binding if challenged in a
court (Cooter 1994).
We argue that if the prescriptions are shared, then they are either norms
or rules. Both formal and informal prescriptions can be classified as
shared norms or rules depending on the presence of the OR ELSE component.
Whether or not the formality of an institution influences the level
  
A GR AM MA R O F I NS TI TU TI ON S 167
of compliance raises an interesting set of questions that disappear if we
use the terms “formal” and “informal” without attention to what those
distinctions mean. Do we find differences between situations governed
by a similar ADICO statement that arise from the formalization of the
institutional statement? Do differences in delta parameters explain the
difference in behavior? Do we regularly expect delta parameters to increase
or decrease when an institutional statement becomes formalized in
some way?
LEGITIMACY AND COMPLIANCE
The language of the syntax enables researchers to address important
questions related to the legitimacy of rules (Tyler 1990). If rulers impose
rules primarily by force and fiat, individuals subject to these rules are
unlikely to develop internal delta parameters associated with breaking
the rules. Nor are delta parameters stimulated only by observation by
others in such settings likely to enhance the rate of rule conformance. If
those who are supposed to follow a rule view it as illegitimate, they may
even reward one another for actions that break the rules (a positive be
)
instead of adopting the type of metanorm envisioned by Axelrod (1986)
(a negative be
).
The complementarity of deltas and the OR ELSEs emerges as important
in analyses of compliance. When delta parameters are close to zero, the
costs of maintaining compliance with the OR ELSE drastically increase
(Levi 1988; Margolis 1991; Ayres and Braithwaite 1992). Without a relatively
high level of voluntary, contingent compliance to rules, Margaret
Levi (1988) explains that rulers can rarely afford the continuing costs
involved in hiring enough monitors and sanctioners, motivating them to
be efficient, and achieving compliance by the actual imposition of sanctions
in a sufficient number of instances that citizens try to conform to
the rules rather than risk the chance of detection and punishment. If violators
can expect to reap the benefits of violating prescriptions without facing
the probability of some established punishment (if there is no OR
ELSE), however, then the experience of feeling the “sucker” may erode
the value of the delta parameter (Levi 1988, 1990; E. Ostrom 1990;
E. Ostrom, Gardner, and Walker 1994; Mansbridge 1994).
BASIC NORMATIVE ASSUMPTIONS
A growing body of work on cooperation and compliance considers the
mix of normative and material motivations that individuals consider
when faced with choices.21
These works treat the normative aspects of
decisions up-front as a significant part of the analysis. Margolis (1991,
  
C HA PT ER FI VE168
130) argues for the necessity of such an approach: “[i]f we analyze everything
in terms of strict self-interest and then include some social motivation
only if we get stuck or if there is something left over, it is not likely
to lead to nearly as powerful a social theory as if the two things are built
in at the base of the analysis.” Delta parameters provide a conceptual
language with which to incorporate explicitly normative considerations
into analysis from the beginning and to discuss differences in studies that
incorporate normative incentives along with extrinsic payoffs.
In table 5.2, we list studies that have addressed three types of questions
about normative motivations. The top section of the table lists different
assumptions regarding the meaning and sign of delta parameters. Knack’s
(1992) analysis of voter turnout, for example, illustrates the insight possible
from a careful study of the normative influences captured by the delta
parameters. He offers empirical evidence of the substantive content of
internal deltas and external deltas associated with a turnout norm and of
the influence of these delta parameters on the probability that an individual
will vote. Interestingly, his interpretation of the declines in voter turnout
echoes the importance of monitoring and sanctioning discussed earlier.
He finds that social pressures (external deltas) are a key influence on
voting turnout. Social pressure operates, however, only when voters expect
to be in situations where someone may ask them if they voted, and
may express disapproval if they did not. As the percentage of individuals
in organizations, in relationships with neighborhood residents, and in extended
family situations drops, this monitoring decreases, and the power
of the social sanction (the external delta) diminishes (see also Amaro de
Matos and Barros 2004).
The middle section of table 5.2 cites work that addresses the implications
of assuming different individual orientations in a situation as discussed
in chapter 4. For example, a rational egoist would assign a zero
value to praise or blame for obeying or breaking prescriptions. One interesting
variant of this analysis ties the size of the delta to the number of
others who conform to the prescription; conformance is conditional (Elster
1989a).
Scholars cited in the last section of table 5.2 discuss variables that influence
the creation and maintenance of delta parameters. Offe and Wiesenthal
(1980) offer an interesting substantive application. They consider
the influence of the costs labor unions face in building and maintaining
shared commitment to participation norms as this in turn affects their
ability to compete with other interest groups. Several other authors ask
whether normative incentives increase or decrease with use. Olson (1991)
views the delta parameters as scarce resources that can be dissipated with
too much use, while Hirschman (1985) and Mansbridge (1994) come to
the opposite conclusion. They argue that the normative constraints in-
  
A GR AM MA R O F I NS TI TU TI ON S 169
TABLE 5.2
Delta parameters and normative concepts used in recent literature
Concepts used by other
Delta parameters authors Citations
Size, sign, and interpretation of delta parameters
+δoi
warm glow Andreoni 1989; Ledyard 1995
+δoe
encouragement Coleman 1988
status improvement / Coleman 1988
reputation enhancement
honor Ullmann-Margalit 1977
−δbi
duty Knack 1992
−δbe
cost of being punished Axelrod 1986*
(P and P′)
social sanctions Knack 1992
third-party sanctions Bendor and Mookherjee 1990**
+δoi
and −δbi
internalized norms Coleman 1987
public-spiritedness Mansbridge 1994
moral duty Etzioni 1988
duty Commons [1924] 1968
+δoe
and −δbe
externally sanctioned norms Coleman 1987
reputation Kreps 1990
responsibility Commons [1924] 1968
moral judgment Sugden 1986
Types of players
+δoe
large zealot Coleman 1988
∆ = 0 selfish rational individual Elster 1989a
+δoi
and/or −δbi
large everyday Kantian Elster 1989a
+δo
large when number elite participationists Elster 1989a
of cooperators low
+δo
large when number mass participationists Elster 1989a
of cooperators high
∆ larger when number of people motivated by Elster 1989a
cooperators > threshold fairness
Creation and maintenance of delta parameters
∆ affected by labor union Offe and Wiesenthal 1980
activities
∆ are scarce resources that Olson 1991
erode with use
∆ are resources that Hirschman 1985; Mansbridge 1994
increase with use
∆ affected by external fines Frey and Jegen 2001
∆ lower when rules come Frey 1994
from outside authority
Source: Crawford and Ostrom 2000, 131.
* In some cases, these sanctions may meet the criteria of an OR ELSE.
** As with the P and P′ of Axelrod’s, the third-party sanctions may at times meet the criteria of an OR
ELSE.
  
C HA PT ER FI VE170
crease in size as they are used repeatedly by individuals in a group. Frey
and Jegen (2001) contend and provide evidence that external interventions,
such as fines, adversely affect the size of delta parameters, particularly
the internal deltas. Frey (1994) speculates that the deltas associated
with rules will be higher when individuals participate in making their own
rules than when rules are made by higher authorities. The other works in
this section debate whether deltas increase or decrease with use.
FREEDOM AND CONSTRAINT
By partitioning sets of actions or outcomes into required, forbidden, and
permitted subsets, rules both constrain and expand the levels of freedom
for actors. Frank Knight reflected on the relationship of institutional rules
to freedom through the constraints that they impose. “But freedom is like
other traits of human nature, in that it is created by a social situation or,
in more technical terms, a complex of institutions. This also sets limits to
freedom” (Knight 1965, 304). Freedom is restricted by rules that limit the
actions an individual can take in an action situation. In a market, for
example, an individual is only legally authorized to conduct specific commercial
transactions and not to take goods by force. Legislators can vote
only on bills presented to them through a formalized set of procedures.
The restrictions on the freedom of one person, however, open up opportunities
for creative actions by others. So individual freedom of action is
also expanded by enforceable rules since rules enhance the predictability
of the actions taken by others in a decision situation. A buyer may obtain
credit from a seller to buy goods for which the buyer does not yet have
sufficient funds because the seller knows that a contract for future payment
can be enforced in a court. The capacity to bind oneself to future
performance through a set of enforceable rules thus opens opportunities
not available without a system of enforceable rules. It is, of course, this
relationship of the freedom to accomplish something by being able to
constrain oneself that underlies the myth of Ulysses tying himself to the
mast in order to resist the call of the sirens and thus safely guide his boat
through a rocky passage (Elster 1979; see also V. Ostrom 1996, 1997).
INSTITUTIONAL CONFIGURATIONS
So far, the descriptions of the components of rule and norm statements
focus primarily on single statements as if the content of the institutional
statements is independent. The focus on single statements is for expository
reasons only. When we examine the interactions of individuals in a
situation, we expect to find that a configuration of rules, norms, and
shared strategies influences the choices of individuals at any one point in
  
A GR AM MA R O F I NS TI TU TI ON S 171
time. In fact, we often find nested institutional configurations, professional
norms, or agency rules nested within government regulations, for
example.
In some cases, the CONDITION component of an institutional statement
explicitly states the linkages between statements in a configuration.
For example, a rule permitting some action may state as a CONDITION
of the rule that the individual follow some procedure outlined in another
rule. In other cases, the linkage between statements is implicit. For example,
the CONDITION component of a voting rule for legislation may not
overtly make reference to the quorum rule, but the specific quorum rule
in place strongly influences the effects of the voting rule.22
A rule stating
that a majority must approve before a bill becomes a law affects behavior
differently depending upon (1) the quorum rule that states how many
members must be present and voting for a vote to be legal and (2) the rule
that states what happens if no positive action is taken (e.g., the OR ELSE
rule for a collective-choice aggregation rule that states a return to the
status quo or an OR ELSE that states some alternative outcome) (see
tables 2.2 and 2.3).
A syntax for individual prescriptions serves useful functions in configurational
analysis. It provides an initial method for sorting the configuration
of institutional statements into rules, norms, and shared strategies.
The ability to identify particular parts of the prescriptions and to state all
prescriptions in the same DEONTIC makes organizing and comparing
the institutional statement in a configuration easier. This facilitates analysis
of inconsistencies in configurations, such as instances when institutions
simultaneously permit and forbid the same action. The syntax can also
be used to help in the organization of types of rules—as we discuss in
chapter 7.
Using the Grammar in Empirical Field Research
In empirical field studies, the researcher’s task is to discover the linguistic
statements that form the institutional basis for shared expectations and
potentially for the observed regularity in behavior. Essentially, this entails
discovering which of the components exist in these statements and the
contents of those components. This frequently requires research employing
qualitative methods including in-depth interviews or the reconstruction
of historical and case materials.
In explaining established patterns of interaction, the researcher has to
decide whether it is reasonable to use an institutional statement that assigns
a DEONTIC to an action or outcome. In the field, the researcher
listens for normative discourse. Is there an articulated sense of moral or
  
C HA PT ER FI VE172
social obligation expressed? If the individuals in a study share only AIC
statements, their discussion of why they would follow such advice focuses
only on prudence or wise judgment. “The best thing to do when faced
with a choice between A and B under condition Y is to choose A because
one is usually better off with this choice.” When individuals shift to a
language of obligation, they use terms such as “must” or “must not” to
describe what they and others should do. When social and moral obligations
are discussed, an empirical researcher initially assumes that it is appropriate
to include a DEONTIC in institutional statements used to explain
behavior. “The obligatory action when faced with a choice between
A and B under condition Y is to choose A, because this is the proper
action.” Occasionally, the analyst can directly pick out rules, norms, and
strategies from written policy statements such as a law code, written regulations,
specific statements in a court decision, or written agreements established
by a village or association. It may also be possible to extract
clear institutional statements from institutional statements retained
through oral traditions. (All of these methods were used to identify the
rules discussed in chapter 8.)
For example, if one were analyzing the use of a commons in workplace
settings, it would be easy enough in many offices to find a sign taped to
the front of the microwave that specifically states “If you use the microwave
you must clean up your own mess!” (as in example 4 above, which
was observed in the commons of the Complex Systems Group at the University
of Michigan in November 2002). In such cases, the institutional
statement can be directly parsed using the grammar of institutions. The
main operational question here is whether or not that statement found
amid the written statements or oral traditions is really a shared institutional
statement that participants know and use. The “know and use”
conditions here are particularly important for analysis of the influence of
the institutions on the outcomes in the action situation, which is the emphasis
of this book (“rules in use” as described in chapter 2).
When the institutional statements relating to a specific action situation
are clear and shared, and when those statements are at an appropriate
level of precision for analysis of the question at hand, then extracting
them is a relatively straightforward process. The second condition here
(appropriate level of precision) returns attention back to the question of
scale. Just as different maps are needed depending on the geographic question
at hand (e.g., whether one is explaining, for example, where an island
is versus explaining to someone how to drive to a specific location on that
island), so too different levels of detail or precision may be appropriate
to different kinds of policy questions. Sometimes the level of precision of
the statements in the empirical setting is quite appropriate for analysis (as
is discussed in chapter 8).
  
A GR AM MA R O F I NS TI TU TI ON S 173
Fieldwork will not always lead to easy extraction of simple and clear
institutions-in-use occurring exactly at the appropriate level of precision.
Consequently, empirical researchers of institutions usually must develop
appropriate summaries of the complex morass of legal, written, or oral
statements that are relevant to the action situation being analyzed. For
some analyses of campaign finance rules, for example, it might be sufficient
to summarize volumes of legal code, federal regulations, and court
decisions concerning campaign finance in the United States into the set of
institutional statements: “Candidates may spend unlimited amounts of
money on the condition that disclosure rules are followed. If the spending
does not occur in a manner that meets this condition, then the candidate
is fined.” Another set of institutional statements that could be extracted
here would be: “Candidates may spend unlimited amounts of money on
the condition that they refuse public financing. If a candidate does not
refuse public financing, then the FEC must fine the candidate for exceeding
legal spending limits.”
The analytical work of identifying the statements that most need to be
summarized and identifying the appropriate level of precision are important
research design questions in institutional analysis analogous to
specifying the appropriate variables and the appropriate precision of variables
in statistical analysis. As we have used the syntax, we find that it
sometimes takes trial and error to extract a useful summary of institutional
statements in an empirical action situation that bring analytical
focus to the appropriate questions at the appropriate level of precision.
Some Next Steps
We started this enterprise in an effort to define the concept of rules clearly.
We found that in order to do this, we needed to clarify how rules were
related to norms and strategies. As a result, we have learned a great deal
about norms and strategies. Delving deeper into each of the components
brought to light connections among these concepts, and the literature that
focuses on them, which had not previously been linked. Moreover, fitting
this all into a grammar helped us to catch inconsistencies and to further
test and refine our understanding of each type of institutional feature and
each component.
A larger question that this research effort must eventually address in
order to return to the core concerns of political science is how an institutional
grammar relates to a theory of knowledge and a theory of action
(see V. Ostrom 1997). Our notion of delta parameters brings normative
considerations into the analysis of action and consequences. This is not
the same as incorporating concerns about the welfare of other actors
  
C HA PT ER FI VE174
into an individual’s calculus. The delta parameters arise from commitments
to the norms and rules of a community, not from the incorporation
of other’s payoffs into one’s own payoff.23
Our notion of OR ELSE brings
institutional consequences into individual-level decisions. All of the institutional
statements affect expectations about others’ behavior, which create
stability in social life. The ADICO syntax illustrates the potentially
cumulative manner in which institutional statements can affect individual
expectations. Thus, improving our ability to analyze and to discuss the
institutional statements prepares us for a more thorough analysis of institutions
and human action. Attending to a grammar of institutions equips
us to return to the core issues of institutions and political order with new
effectiveness.
  
Six
Why Classify Generic Rules?
IN CHAPTER 5, Sue Crawford and I use the ADICO grammar to distinguish
among three essential components of all institutional analyses: strategies,
norms, and rules. Hopefully, the reader has grasped the importance of
understanding how individuals adopt strategies in light of the norms they
hold and within the rules of the situation within which they are interacting.
In chapter 7, we will move forward to discussing a way of consistently
grouping rules so that the analysis of rule systems can be made
much more cumulative. This chapter is a prologue to chapter 7 in that it
addresses why we would introduce still further conceptual tools beyond
those of the general Institutional Analysis and Development (IAD) framework
and the grammar of institutions discussed in chapter 5.
The approach to classifying rules to be discussed in this and the next
two chapters has taken many years to develop. Some of the notes drawn
on in writing these chapters were written in the early 1980s. A lecture
delivered in Bielefeld in 1982 was my first effort to elucidate how rules
were the “hidden” structure underlying games. When I gave a revised
version as my Presidential Address to the Public Choice Society meetings
in March of 1984 (published two years later as E. Ostrom 1986), and
posited the seven types of generic rules that are discussed in this and the
next two chapters, several colleagues strongly criticized me for introducing
so much complexity. They asked, “Why are you driven to do something
so unnecessary?”
Among the answers I have given to these questions is the following list
of “needs”:
• to work further toward solving babbling equilibrium problems for scholars
doing institutional analysis;
• to understand how action situations are constructed so that reasonable reforms
can be considered;
• to move beyond slogan words to describe institutions;
• to cope with the immense diversity of rules by clustering them into seven
generic rules;
• to look at rules as information-transformation mechanisms; and
• to study the underlying universality of rules.
  
C HA PT ER SI X176
Solving Babbling Equilibrium Problems
In his book Convention, David Lewis (1969) clearly laid out the problems
of communicating with one another. Assuming that scholars—as well as
anyone else—are better off when successful communication occurs, the
problem of communication is that of arriving at a shared convention for
the names (or symbols) that we will attach to diverse states of the world.
In essence, communicators are engaged in a signaling game. If the senders
and receivers of a signal do not understand it the same way, they are
unfortunately in a “babbling equilibrium.”
The sender of a signal wishes to convey information about states of the
world to a receiver of that signal. Since terms are not endowed with some
preexisting meaning, the sender could use any of a wide diversity of terms
for the particular states of the world being described. If there were, for
example, three states of the world that the sender wanted to describe, the
sender could use colors to describe the relevant states of the world, such
as red for state 1, green for state 2, and orange for state 3. In essence, this
is what a traffic signal does when it is appropriately engineered. By sending
a green signal to the drivers on one street, it conveys to them the
information that the drivers on the cross street have a red signal and will
not be crossing the intersection. By sending an orange signal to one street,
it is conveying that the signal on the cross street is about to change and
drivers there will begin to cross the intersection. By sending a red signal
to the drivers on the first street, it is conveying that drivers on the cross
street have the right of way and it is dangerous (as well as illegal) to cross
the street.
One equilibrium of a signaling game is that everyone uses and understands
the signals in the same way. This is obviously the optimal equilibrium.
When it comes to traffic signals, the advantage of achieving this
equilibrium is obvious! There is nothing inherent in red, green, and orange
that means stopping, going, and slowing down. The signals used for stopping
and going could easily be different colors (or even the reverse, so
that green means stopping and red means going); the crucial problem of
optimal communication is that senders and receivers use the same signal
to mean the same thing. If they do not, their signals are a form of babbling
and the equilibrium outcome is highly undesirable.
In developing the ADICO syntax as a way of sorting out the diverse
elements involved in institutional statements, we could just as well have
reversed the terms we use for statements containing (or not containing)
“or elses.” In chapter 5, we proposed that a statement containing four
elements of the syntax (ADIC) should be called a norm and five elements
of the syntax (ADICO) should be called a rule. Instead, we could have
  
W HY CL AS SI FY GE NE RI C R UL ES ? 177
called the first a rule and the second a norm. A quick look at table 6.1
shows that several scholars do exactly this. We even observe some scholars
who use the term “norm” at one time for an ADIC statement and at
other times for an ADICO statement.
As senders of a signal, they probably know which attributes are present
when they use one term or another. As receivers of their signals, however,
how are we to know which attributes are present or not? Unless we have
access to the same materials and spend costly effort determining which
attributes are present or absent, we cannot know the meaning of a term
when Author Jones uses all five attributes as reference for a norm and
four attributes to be a rule as contrasted to Author Smith (or even Author
Jones in another text). If senders and receivers have the same attributes
of the states of the world in mind, when they use a particular word to
describe this state, they reach the optimal equilibrium in a signaling
situation.
The similarities between rules, norms, and shared strategies in the
ADICO syntax help us to understand why the literature so frequently uses
these terms interchangeably. They share several of the same features.
Table 6.1 sorts the concepts that other authors use into the types of institutional
statements created by our syntax. All of the terms used by other
authors, shown in the top section of the table, appear to describe institutional
statements that are shared strategies according to the syntax; they
contain AIC components. The need for a consensus in the use of terms is
vividly illustrated by examining the number of different terms in each
section of the table as well as the fact that several terms appear in all three
sections. That these terms have been used in so many different ways is
not a criticism of past work. Rather, it illustrates the difficulty of untangling
key social science concepts that are foundational for institutional
analysis.
Lewis proposes that meaning occurs when senders and receivers of signals
reach such an equilibrium no matter which specific term is being
used. There are, of course, other equilibria possible where there is
agreement on the attributes of one term but not the others. This appears
to characterize the current situation where scholars are more likely to
agree on the use of the term strategy. This is contrasted to the use of rules
and norms for which there is little agreement. Lewis calls these “babbling”
equilibria. Receivers have no reasonable method for imparting
meaning to some of the signals they receive, while there is agreed-upon
meaning for others (see also Crawford and Sobel 1982; Farrell 1993).
The problem of babble about rules and norms is not confined to scholars
who study something other than rules and norms and only occasionally
misuse a term whose meaning has become well established by those
who work extensively on these concepts. In a recently published book
  
C HA PT ER SI X178
TABLE 6.1
Shared strategies, norms, and rules as used in ADICO syntax and in recent literature
ADICO Terms used by
components Crawford-Ostrom Terms used by other authors
AIC Shared Strategies Axelrod 1981—Rules, strategies
Axelrod 1986—Norms
Bourdieu 1977—Doxic elements of action
Calvert 1992—Equilibrium strategies
Hodgson and Knudsen 2004—Conventions
Levi 1990—Norms
March and Olsen 1989—Rules
Meyer and Rowan 1991—Taken-for-granted actions
Myerson 1991—Rules
Rowe 1989—Rules of action
Schank and Abelson 1977—Scripts
Schelling 1978—Focal points
Schotter 1981—Institutions
Ullmann-Margalit 1977—Conventions
ADIC Norms Azar 2004—Norms
Bicchieri 1997—Norms
Braybrooke 1987—Conventions
Braybrooke 1996—Rules
Coleman 1987—Norms
DiMaggio and Powell 1991—Institutions
Levi 1990—Norms
Lewis 1969—Conventions
March and Olsen 1989—Rules
McAdams 2001—Normative attitudes
Meyer and Rowan 1991—Taken-for-granted actions
North 1981—Ethical codes
Rowe 1989—Obligations
Schotter 1981—Institutions
Sugden 1986—Conventions
Ullmann-Margalit 1977—Social norms
Weber 1947—Conventions
ADICO Rules Albert 1986—Laws
Axelrod 1986—Norms backed by metanorms
Coleman 1987—Norms
Commons [1924] 1968—Working rules
Hurwicz 1994—Rules of the game-form
J. Knight 1992—Rules
Levi 1990—Legalistic institutions
North 1990—Rules
Shepsle 1979a, 1989—Rules
Ullmann-Margalit 1977—PD norms, decrees
Weber 1947—Laws
Source: Crawford and Ostrom 2000, 130.
  
W HY CL AS SI FY GE NE RI C R UL ES ? 179
called Social Norms, the editors—who had convened several conferences
and worked extensively with a group of knowledgeable scholars conducting
studies on social norms—introduced the resulting volume by commenting:
“These chapters suggest that consensus about social norms is
limited both across disciplines and within them. To begin, the concept of
social norm clearly means different things to different scholars” (Hechter
and Opp 2001, xii). They go on to bemoan the fact that as “there is no
common definition of social norms, there can be little agreement about
how to measure them” (xiii). Obviously, various authors in the volume
did define social norms—many of them in a way that is consistent with
the syntactic definition provided in chapter 5. The book contained, however,
a multiplicity of definitions of the central concept under study. In an
excellent recent review of how the concept of norms is used in the field
of law and economics, McAdams and Rasmusen (forthcoming) also bemoan
the lack of consensus on the definition of a norm.
The existence of a babbling equilibrium in regard to these terms is a
key problem for the social sciences.1
The concepts of rules and norms are
at the heart of many core theoretical questions having to do with how
individuals coordinate activities with one another. Lacking agreed-upon
definitions of these terms leaves us all in a suboptimal babbling equilibria
rather than in a scientific signaling game where general progress is likely.
If scholars were to accept the syntax as we have laid it out in chapter
5—or an improved version of it—it would help to solve the babbling
equilibrium problem.
At an earlier juncture, Douglass North (1990) dug into another babbling
equilibrium problem. Many scholars tended to use the concept of
an organization and of an institution interchangeably. North has insisted
on a key difference between organizations and institutions. As North described
his approach:
A crucial distinction in this study is made between institutions and organizations.
. . . Organizations include political bodies (political parties, the Senate, a
city council, a regulatory agency), economic bodies (firms, trade unions, family
farms, cooperatives), social bodies (churches, clubs, athletic associations), and
educational bodies (schools, universities, vocational training centers). They are
groups of individuals bound by some common purpose to achieve objectives.
. . . [T]he emphasis in this study is on the institutions that are the underlying
rules of the game and the focus on organizations (and their entrepreneurs)
is primarily on their role as agents of institutional change; therefore the emphasis
is on the interaction between institutions and organizations. (4–5)
We are using terms in a manner consistent with North’s distinction. Rules
are part of the underlying structure that constitute a single action situation
or a series of them. Organizations may be participants in a situation struc-
  
C HA PT ER SI X180
tured by rules and can, in turn, be analyzed by looking at the linked action
situations used by the group “bound by some common purpose to achieve
outcomes.” Most organizations would be composed of multiple simultaneous
and sequential action situations—all constituted by rules as well as
by the physical world.
Sorting out organizations and institutions, and strategies, norms, and
rules is a beginning. It is, however, not enough. The problem becomes
even more substantial when one moves beyond the effort to develop a
general definition of rules and norms to ways of classifying rules. As soon
as one digs below the surface of an action situation to explore the prescriptions
partly responsible for their structure, one is struck with the vast
array of potential rules. The ADICO framework enables the analyst to
classify institutional statements by using their syntax to determine the
difference between strategies, norms, and rules. Now we need to ask,
What is a useful classification based on the semantics of rules?
The Policy Analyst’s Need to Understand How to Reform Situations
Policy analysts must be able to use the semantics of rules. Elected officials,
the staff members of national, state, local governments and citizen interest
groups, and many social scientists are asked what rules should be changed
to solve a particular kind of problem. The questions may be in regard to
problems as global as how to change the incentives facing users of carbon
so as to reduce the likelihood and magnitude of global warming or as
local as how to change the parking regulations in a small town so as
to make it a more attractive location for residents in the region to shop
downtown.
The policy analyst stands in relationship to “repairing” single and
linked action situations as a doctor stands in relationship to a sick patient,
a mechanic in relation to a car that does not run, or a computer technician
in relationship to a computer system producing strange symbols on a
printer.2
In attempting to solve all these problems of poor performance,
the “Doc” has to dig under the surface, begin to think which subsystem
or linkage among systems is most likely to be causing the problem, and
begin to do tests to confirm these initial speculations. For physicians and
mechanics, the systems they need to understand are primarily biophysical
systems and subsystems. The mechanic learns how to take the system
totally apart and reconstruct it subsystem by subsystem. The general physician
does not have the freedom to take his or her patient’s system apart
but certainly learned in medical school how to dissect other living beings,
witnessed many surgical operations and autopsies, and studied a cumulative
body of knowledge about how various parts of the human body
should work. The computer technician has both a mechanical system to
  
W HY CL AS SI FY GE NE RI C R UL ES ? 181
try to understand as well as multiple layers of software languages. He or
she must know what each of the major types of commands does in the
most frequently used languages and which types of commands are likely
to be compatible or incompatible.
The policy analyst’s task is closer to that of the computer technician
than to the auto mechanic or physician. A very large proportion of what
the policy analyst must understand is language-based. When one is studying
rules, after all, the content is entirely language-based. Yet, it is also
essential to know key aspects of the relevant biophysical and social world
to propose changes that are likely to improve outcomes. When we are
asked to propose new rules, however, the request is for something that
will be expressed in language, as Vincent Ostrom (1987, 1993, 1997) has
so often articulated.
Moving beyond Slogan Words to Describe Institutions
The embarrassment that we face is that policy analysis has yet to develop
a coherent understanding of how our subject matter should best be expressed,
how rules fit or don’t fit together to shape observable behavior
and outcomes. All too often, slogan words, such as “privatization,” “centralization,”
and “decentralization,” are used as substitutes for careful
analysis. What are the specific rules that we are talking about when we
talk about a privatization or decentralization policy? What changes in the
incentives of participants will occur if we propose a particular set of new
rules versus other potential sets? The lack of a disciplined language to
be able to analyze, dissect, and propose better reforms was dramatically
illustrated for the world after the collapse of the former Soviet Union.
Western scholars were asked to help Russia and the other “newly industrializing
countries” to create a vigorous and productive private market
economy. Many of our proposals were actually accepted and efforts made
to implement them. But, instead of an open, competitive market, the rules
proposed by policy analysts generated commercial monopolies, massive
corruption, and little economic growth. And, this is not the only time that
the recommendations of policy analysts were accepted and yet did not
produce the predicted results. There is obviously a lot of work to be done.
Coping with the Immense Diversity by Identifying Generic Rules
At the same time as I am committed to digging into and revealing an
underlying universality in the components of human action, I am repeatedly
challenged and amazed by the multiplicity of rules that colleagues
and I have recorded in fieldwork (see chapter 8) and that other scholars
  
C HA PT ER SI X182
have used in their analyses of situations.3
If one focuses on only the detailed
aspects of rules, it is hard to see how they are similar and how they
differ. It is then hard to analyze carefully how changes in a particular type
of rule (as contrasted to changes in several different types of rules) affect
the resulting situations. This recognition and appreciation of variety leads
one to have a substantial wariness related to the capacity of humans to
design optimal systems without a substantial trial-and-error process so as
to learn what works in a particular biophysical environment. I will return
to this issue in chapter 8.
The puzzle has been how to develop a language to express rules that
enables one to capture in a consistent and cumulative manner the most
general characteristics of the basic rules while providing a structure that
enables scholars to enumerate and test the impact of changes in a rule
type. I remember my surprise when I was working on a draft of my Presidential
Address to the Public Choice Society (E. Ostrom 1986) and found
that Charles Plott was unaware that his work on “default conditions”
(Grether, Isaac, and Plott 1979) was an excellent empirical study related
to earlier work by Niskanen (1971) and Romer and Rosenthal (1978) on
“reversion levels.” I had asked Plott to review my draft paper where I
refer to that earlier paper, in which Plott and his colleagues had used the
term “default condition” to refer to an ingenious method they had devised
to examine what would happen if participants in an experiment were not
able to come to an agreement using a unanimity rule.
I thought their “default condition” was the same as the concept referred
to as a “reversion level” by Niskanen (1971) and by Romer and Rosenthal
(1978). In chapter 7, we discuss the need for all aggregation rules to include
a condition stating what decision will hold if multiple participants
do not reach an agreement. We call that rule a “lack of agreement” rule.
What was neat, I thought, was that Plott and his colleagues had actually
tested the impact of changing this rule in the laboratory. They found that
the particular “lack of agreement” rule they used “literally determines the
outcomes in processes such as these” (Grether, Isaac, and Plott 1979, V-
7). Their experimental results thus strongly confirmed Romer and Rosenthal’s
argument in an earlier “theoretical debate” between them and
Niskanen about the likely results when agents did not agree to a proposal
under different “reversion levels.”4
Even with Plott’s substantial acuity
and awareness of the importance of studying rules, he was surprised and
pleased when I pointed out the similarity. What I dream of is a method
for classifying rules that enables scholars to know when they are talking
about the same “variety” of a rule—in this case, the default condition of
an aggregation rule.
To illustrate what I mean by a generic rule, let us take an everyday
example of the wide variety of a specific rule used in practice—sale of
  
W HY CL AS SI FY GE NE RI C R UL ES ? 183
alcoholic beverages. Most state legislatures have passed legislation related
to the sale of liquor—particularly related to who can legally hold the
position of a Buyer of alcoholic beverages. Using the first generic entry
rule enumerated in chapter 7, one would specify a generic entry rule for
buying liquor as:
Persons with attributes a1 , a2 , . . . an , Permitted to hold the position of a Buyer
of alcoholic beverages subject to conditions, c1 , c2 , . . . cn .
What attributes and conditions are used varies widely from one place
to another. To buy liquor at a liquor store, the attributes usually relate to
age and the kind of identity card that may be required. The conditions
relate primarily to the days of the week and the hours that liquor stores
are authorized to be open. To even enter a bar, as contrasted to a liquor
store, in some jurisdictions, the attributes of a Buyer must also be held by
anyone who enters the facility. Younger siblings can accompany older
brothers or sisters when buying liquor at a grocery store, but not at a
local pub. If one were to enumerate all of the variations in attributes and
conditions just for holding the position of being a Buyer of major types of
commodities (bread, prescriptions, pork, liquor, cars, etc.), the list would
become incredibly long very fast. For some purposes, such as trying to
improve policies related to the issuing of food stamps or reducing adolescent
alcoholism, one could identify an important subset of all of these
rules to analyze where they have been used, what other factors affect the
desired outcomes, and what kind of performance has been achieved.
The specificity of the rules that one studies depends on the question one
wants to ask. The generic rules are not specific in terms of identifying the
particular attributes or conditions that affect a type of rule. When a research
team at the Workshop in Political Theory and Policy Analysis was
studying the rules used to regulate appropriation from common-pool resources
around the world and how they affected behavior and outcomes,
we were interested in developing a full enumeration of the specific attributes
and conditions described in the case studies that we read and coded.
As discussed in chapter 8, we did find a shockingly large number of attributes
and conditions used in entry rules. For example, for the first entry
rule discussed above, we identified seven ascribed attributes, two acquired
attributes, four residency attributes, and thirteen conditions that are used
individually or in combination to make an individual eligible to access a
common-pool resource (see table 8.1).
The purpose of classifying generic rules in chapter 7, however, is not to
be exhaustive in regard to detailed attributes and conditions. No generalpurpose
language for classifying rules can do this! A key task is to provide
the equivalent of the components of a general recipe for creating situations.
A good cook knows, for example, that to make a cake requires
  
C HA PT ER SI X184
some proportion of flour, sugar, fat, spices, a leavening agent, and time
and temperature at which a baking pan should remain in an oven. That
is the “general recipe” for making cakes. Pies have quite a different “general”
recipe. Grandmother’s secret recipe for Chocolate Surprise Cake,
on the other hand, lists specific quantities of particular ingredients, the
type of pan that should be used, and the temperature and time of baking.
What is “secret” is the particular type of spices or sugar, some combination
of flours, something about the timing and process of mixing, or any
of a wide variety of other details that makes her cake something special
that her grandchildren fight over. The purpose of the classification system
outlined in chapter 7 is not to get to this level of detail, but rather to
provide a general classification system that can be used by multiple scholars
who are analyzing a specific question and examining rules in order to
address this question.
By identifying the generic structure of entry rules, all of the many ascribed
and acquired characteristics that could be used in a specific entry
rule are considered together as a diverse set of attributes. Similarly, all of
the various relevant events—such as the results of a prior situation—are
simply considered as “conditions.” Thus, the classification proposed in
chapter 7 identifies the most general structure of a rule configuration.
The Role of Rules as Information Transformation Mechanisms
Once we have developed a way of expressing rules in a systematic fashion,
we can begin to address a number of quite important and exciting questions.
One of these has to do with the generative capacity of rules—their
productive and reproductive capacities. If we consider institutional rules
to have a broad similarity to grammatical rules, questions concerning the
information-processing capabilities of institutional rules are similar to
those of a grammar.
Any mechanism that transmits information about how to produce
something (a protein, a sentence, or an action situation) is itself subject
to noise, to random error, and to distortion. Institutional rules are probably
more vulnerable to these problems than either grammatical rules
(since humans are motivated to try to make themselves understood and
thus to follow grammatical rules) or genetic codes (since these “instructions”
do not rely on humans themselves to carry them out, and biologists
have ascertained that the mutation rate is actually very low).
Rules that are repeatedly found in many different types of situations
are apt to be more reliable “building blocks” than rules that are only
infrequently included among the set of rules constructing social arenas.
Once we have a systematic way to classify generic rules, we can study
  
W HY CL AS SI FY GE NE RI C R UL ES ? 185
diverse institutional arrangements in a manner to identify those rules that
are components of a large number of situations but that have a surface
appearance of being different. That may begin to give us some clues concerning
the use of redundancy and of the iteration and reiteration of a
rule to make much more complex structures.
An Underlying Universality?
Another major reason for this effort is my deep conviction, as I state in
chapter 1, that underlying the immense variety of surface differences, all
repetitive situations faced by human beings are composed of nested layers
composed from the same set of elements.
Game theory has already provided us a formal language for expressing
the structure of relatively simple and unambiguous action situations.
Some of the essential working parts of an action situation—such as positions
(as distinct from players) and outcomes (as distinct from payoffs)—
are frequently not acknowledged when a game theory textbook lays out
the working parts of a game. Expanding these parts—as was accomplished
in chapter 2—enables us to be more confident that we have identified
the components of a wide diversity of situations found in all sectors
of organized life. This improves analysis in several ways. First, it requires
the analyst to make hidden or implicit assumptions overt and explicit.
Second, it provides clear elements that need to be generated by rules (or
by physical laws). By not requiring that all situations must be expressed
as a formal game, the concept of an action situation expands the range
of situations that can be compared using the same structural variables.
The attempt to find a way of expressing the most basic rules that generate
action situations will help us illustrate that the generic rules used to constitute
a market come from the same set of generic rules used to constitute
a legislature, a hierarchy, a self-organized resource governance system, or
any of a wide diversity of situations. Having now discussed in some depth
the “why,” I think it is important to define generic rules. Let us move on
to the task of classifying rules by their AIM.
  
Seven
Classifying Rules
E L I N O R O S T R O M A N D S U E C R AW F O R D
THE PURPOSE of this chapter is to develop a useful system for classifying
or naming rules. In our effort to group rules into useful classifications, we
recognize that no single classification can ever be useful for all purposes.
One strategy that is commonly used is to order rules according to the
jurisdiction that created them. All national rules are classified together,
state or provincial rules are a second group, and local rules are a third
group. This is a useful first cut when one is studying the similarities or
differences among multiple domains of a legal system, but does not address
how to change action situations within a jurisdiction. Another strategy
examines how a rule came into being. De jure rules that have been
formally authorized by the legitimate form of government in a particular
jurisdiction are considered as different from de facto rules that are actually
used by participants. Alternatively, some scholars distinguish between
rules that have evolved versus those that have been designed. As institutional
analysts, however, we need to devise a method that draws on the
general Institutional Analysis and Development (IAD) framework to help
link rules to the action situations they constitute.
Thus, the core goal that we seek in developing a useful classification
system for rules is to devise a nested set of rule-concepts that facilitate
building a cumulative body of theoretically and empirically tested research
about human behavior and outcomes in diversely structured situations.
Current tendencies to use various dichotomies (government versus
market, public versus private, formal versus informal) instead of systematic
development of empirically supported theories of human behavior in
diversely structured situations is, we strongly believe, inadequate to the
task.1
We do not pose our classification system as a universally applicable
system, but rather as a useful system for those interested in linking rules
and the action situations (games) created by rules, the biophysical world,
and communities (see Burns and Gomolin´ska 1998 for a related effort).
In this chapter, we posit two conceptual approaches that we have found
most useful in classifying rules. The first approach—and the one on which
we will focus in this chapter—uses the direct AIM of a rule (as defined in
chapter 5) as a method for classifying rules. This method focuses on one
  
C LA SS IF YI NG RU LE S 187
level of action—operational, collective choice, or constitutional choice—
and classifies rules by the part of the relevant action situation that is most
directly affected. We refer to this as the horizontal approach. The second
layer focuses on the level of authority involved in an analysis, which we
refer to as the vertical dimension. Here, we are initially inspired by the
work of John R. Commons ([1924] 1968), who distinguished between
authorized and authoritative relationships. As discussed in chapter 2, the
IAD framework further divides authoritative relationships into those with
the authority to affect operational action situations, called collectivechoice
rules, and those with the authority to affect collective-choice situations,
called constitutional-choice rules (see figure 2.3).
While analyzing rule sets at multiple levels (the vertical approach) helps
us to uncover important authority relationships, to get to a diagnosis for
a particular policy situation, we need to sort the rules at any one level
into basic categories that clarify the links between specific rules at that
level and the structure of the resulting situation (the horizontal approach).
Although this chapter focuses on the ways in which rules directly influence
the structure of an action situation, we need always to bear in mind that
the rules operate alongside the biophysical world and the attributes of the
community and that rules can have indirect effects. So, while the discussion
in this chapter hones in on the direct link between rules and actionsituation
components for the sake of sorting rules by their influence on
different parts of the action situation, the emphasis on rules should not
be taken to imply that rules alone determine these structures.
The Horizontal Approach: Classifying by the AIM of a Rule
As we develop in chapter 5, all regulatory rules have the general
syntax of:
ATTRIBUTES of participants who are OBLIGED, FORBIDDEN, OR PERMITTED
to ACT (or AFFECT an outcome) under specified CONDITIONS,
OR ELSE.
Using this syntax as the basis for classifying rules still leaves multiple ways
to sort rules. One could, for example, use the DEONTIC element (must,
must not, or may) as the foundation for a classification system. All the
rules forbidding something would be classified together. Given that the
deontic operators are interdefinable, however, there does not appear to be
any particular usefulness to be derived from such a classification system. If
a rule were originally classified as a forbidden rule, it could be rewritten
using the permitted or obligated operator.
  
C HA PT ER SE VE N188
Another possibility would be to classify by the “OR ELSE” part of a
rule. For some purposes, this would be useful. One could learn rapidly
what kind of rule infractions in a particular jurisdiction carry a life imprisonment
sentence, a ten-year sentence, or a one-year sentence. For many
purposes, however, classifying by the OR ELSE component of a rule does
not get at the substance of the rules.
It would also be possible to organize a classification system using the
ATTRIBUTES or CONDITIONS sections of a rule. This would again be
useful for some purposes. Young people could potentially look up and
find out what they are allowed to do or not do. One could find out what
rules apply if you own property versus renting it, or the rules that apply
during the period between a party-nominating convention and the general
election. The CONDITION does play an important role in assessing
which rules apply in a given action situation. Consequently, attention to
the CONDITION can help analysts sort out those rules relevant to a given
action situation from those that are not. However, it still leaves us with a
myriad of relevant rules that need further sorting. It does not get us far
enough to solve the policy analyst’s need to know how to repair broken
action situations!
If one wishes to use the syntax as a foundation, this leaves one with the
AIM element of a rule to be used. And this is our plan. Although (as we
note in chapter 5) the syntax fits regulatory rules better than generative
rules, generative rules still do have an AIM, so a sorting mechanism that
uses the AIM works for generative rules too. And, it works for all three
levels of the IAD framework.
Chapter 2 identified the components of action situations that are
used to construct a wide variety of analytical models of markets, families,
hierarchies, legislatures, corporations, neighborhood associations,
common-property regimes, as well as all formal games. The elements
are participants, positions, actions, outcomes, information, control, and
costs/benefits. They are related together in the following manner:
Participants and actions are assigned to positions.
Outcomes are linked to actions.
Information is available about action-outcome linkages.
Control is exercised over action-outcome linkages.
Costs and benefits are assigned to action-outcome linkages.
Participants, who can either be individuals or any of a wide diversity of
organized entities, are assigned to positions. In these positions, they
choose among actions in light of their information, the control they have
over action-outcome linkages, and the benefits and costs assigned to actions
and outcomes.
  
C LA SS IF YI NG RU LE S 189
Information
Rules
PARTICIPANTS
assigned to
POSITIONS
assigned to
ACTIONS
INFORMATION
about
CONTROL
over
Linked to
NET COSTS
AND BENEFITS
assigned to
POTENTIAL
OUTCOMES
Aggregation
Rules
Scope
Rules
Payoff
Rules
Position
Rules
Choice
Rules
Boundary
Rules
Figure 7.1 Rules as exogenous variables directly affecting the elements of an action
situation.
The relationships among the various parts of the action situation are
represented within the square on figure 7.1. When an analyst takes each
of these working parts as givens, no further inquiry is made as to the cause
or source of a particular element. The analyst predicts the likely outcomes
based on the model of individual behavior assumed and potentially evaluates
the expected pattern of actions and outcomes using such criteria as
efficiency, equity, and error proneness.
The analyst examining a one-shot social dilemma among participants
who cannot communicate will predict low levels of cooperation (or none
at all). To dig under that situation, however, to think about changing it,
one needs to know a lot about the underlying structure leading to the
social dilemma. Are the participants a set of prisoners who are being held
apart by a prosecutor trying to get them to confess? So long as the prosecutor
does not use illegal methods and allows the prisoners access to an
attorney, what is inefficient for the prisoners is likely to be socially efficient.
Better to leave the social dilemma as it stands.
If the participants are harvesters from a common-pool resource, on the
other hand, and are led by isolation and individual incentives to overharvest,
the policy analyst needs a coherent scientific language to begin
to think about rules that would change this situation. The recommendation
simply to create “private property” (Demsetz 1967; R. Smith 1981;
L. Anderson 1995) does not tell an analyst anything specific about which
rules could be changed to make the situation one of private property (see
  
C HA PT ER SE VE N190
Rose 2002 and Tietenberg 2002 for good analyses that dig into this question).
Do new positions need to be created—owner versus outsider, for
example? How are individuals chosen to become owners? What actions
may an owner take, and what actions are forbidden? These are the kinds
of questions that we ask at the end of this chapter regarding how to transform
the Snatch Game into a more productive game by assigning property
rights. It turns out that we have to change seven rules in order to modify
this very simple game of nature from one without any rules to one with
some property rights defined. Focusing on the direct link between the
AIM of a rule and the affected component of the action situation classifies
rules in a way that facilitates analysis of structural change.
The action situation mechanism for classifying rules groups rules according
to the element in the action situation that they most directly impact.
Many rules also indirectly affect other components. Figure 7.1
shows the names that we have given to the types of rules based on this
coding scheme. The arrows identify the element of the action situation
directly affected by that type of rule. We also need to remember that two
other major categories of exogenous variables (the biophysical world and
attributes of community) also shape these action situation components.
Using the AIM of a rule for classification leads to the specification of
seven broad types of rules: position, boundary, choice, aggregation, information,
payoff, and scope.2
Position rules create positions (e.g., member
of a legislature or a committee, voter, etc.). Boundary rules affect how
individuals are assigned to or leave positions and how one situation is
linked to other situations. Choice rules affect the assignment of particular
action sets to positions. Aggregation rules affect the level of control that
individual participants exercise at a linkage within or across situations.
Information rules affect the level of information available in a situation
about actions and the link between actions and outcome linkages. Payoff
rules affect the benefits and costs assigned to outcomes given the actions
chosen. Scope rules affect which outcomes must, must not, or may be
affected within a domain.
The direct relationships among rules and the components of an action
situation are shown in figure 7.1 as the set of arrows connecting rules to
specific parts of an action situation. We can also think of a general verb
type that links AIMs directly to different parts of the action situation.
These basic verbs assist in sorting out the various types of rules. A first
assessment in sorting the rules that affect an action situation can be to
determine the general verb that most closely resembles the specific verb
in the AIM of the parsed rule. The basic verb can also be used to translate
diverse specific verbs in rules to general verb types for the sake of uncovering
similarities in rules that, on the surface, look different. Two bound-
  
C LA SS IF YI NG RU LE S 191
TABLE 7.1
The AIM component of each type of rule
Regulated component of
Type of rule Basic AIM verb the action situation
Position Be Positions
Boundary Enter or leave Participants
Choice Do Actions
Aggregation Jointly affect Control
Information Send or receive Information
Payoff Pay or receive Costs/Benefits
Scope Occur Outcomes
ary rules from two different action situations, for example, might use
very different language but may ultimately be identical mechanisms for
regulating how a participant exits the position “voting member.” Table
7.1 lists the basic AIM verb associated with each type of rule.
The discussion of classifying rules by their AIM follows the order of
the list on table 7.1, which starts with position rules and ends with scope
rules.3
These categories simultaneously sort rules for comparative analysis
and identify the immediate link between the rule and a component of the
action situation. However, rules operate together as a configuration. The
clean arrows from one type of rules to one part of the action situation in
figure 7.1 should not be taken as an assumption that these rules operate
independently of one another, or that boundary rules influence only who
enters positions, or that payoff rules influence only net costs and benefits.
While the AIM may hit one element of the action situation, other components
of the rules—particularly the CONDITION and the OR ELSE component—may
impact another component.4
And, one rule may eventually
affect other components of the action situation in addition to the component
directly affected.
An example of a boundary rule that influences many other action situation
components of an “urban taxi game” is a local medallion ordinance.
Such a boundary rule requires that a taxicab display a purchased medallion
from the city before it can legally use the city streets to attract customers.
It has been used in New York City for decades. When a city limits the
number of medallions it will authorize, the entry costs for putting a taxi on
the streets rise substantially, but if the medallions are strictly limited, the
potential returns per unit of time also increase substantially. And the time
that a traveler has to wait before getting a taxi ride may increase substantially.
Thus, one rule can have very substantial impact on the entire structure
of resulting action situations through its direct impact on one of the
  
C HA PT ER SE VE N192
working components of an action situation. This example also illustrates
another advantage of a systematic method for classifying rules, namely,
to help in identifying rules that are similar in function in widely diverse
settings. The medallion entry rule is a specific example of a generic boundary
rule used in a variety of action situations to limit entry. As we will
discuss in chapter 8, it is one of the entry rules used in some locations to
regulate fisheries—requiring fishing boats to purchase fishing permits.
In fact, all rules indirectly impact net costs and benefits since all rules
include DEONTICS that translate into deltas and OR ELSES that generally
have payoff consequences. This complexity does not invalidate the
analytical goal of sorting rules by types in order to get a better handle on
the influence of rules on the action situation components. It does, however,
mean that the initial sorting does not identify all of the rules that
indirectly influence each component. Finding the rules that indirectly influence
each component involves a next step of looking at the other rule
components to find ATTRIBUTES, CONDITIONS, or OR ELSES that
link to that particular component.
Thus, classifying rules by their direct effect on an AIM offers an important
tool for sorting rules into basic types based on their immediate
impact on the action situation, but it does not cleanly classify all rules
into mutually exclusive categories based on the component of the action
situation that is eventually affected. Since rules can often influence more
than one part of an action situation, this simply cannot be done at this
level of sorting. This classification scheme sorts rules into useful categories
for policy analysis by focusing on the AIM first, while recognizing that
further sorting may be necessary for more complete specification of the
rules linked to particular components of the action situation.
With caveats aside, we now turn to an overview of each of these seven
types of rules. We concentrate here on identifying the most general aspects
of each type of rule since so many specific examples of each type of rule
are found in practice. We thus focus on what appear to be the most basic
or generic forms of each type of rule affecting diverse action situations.
The language of rules is, like all of the analytical concepts discussed in
this book, hierarchical in structure. The generic rule forms are special
cases of the seven basic types of rules, and finer rules are elaborations and
special cases of these generic rule forms that we lay out here. Chapter 8
provides examples that illustrate the immense variety of specific rules by
drawing on empirical studies of rules related to common-pool resources
to elaborate the immense variety of specific rules found to exist in a study
of the governance and management of irrigation, inshore fisheries, and
forest institutions around the world.
  
C LA SS IF YI NG RU LE S 193
Position Rules
An initial building block of an action situation is the set of positions or
anonymous slots that are filled by participants and to which specific action
sets are assigned at junctures in a decision process. Position rules
create these positions. Position rules are often not by themselves intrinsically
interesting, as they merely create “holders” for participants to enter
and for the specification of actions that participants in positions can take
at specific nodes in a process.5
As discussed in chapter 2, positions are the
connecting link between participants and authorized actions. Thus, they
create the scaffolding for a wide diversity of situations that can vary significantly
in terms of the number of positions and the relative authority
assigned to each position.
A minimal position rule names a single position or slot—such as “member”—as
the most inclusive position to be held by all participants in a
situation. In the initial Commons Dilemma experiments described in
chapter 3, the experimenters generated only one position, even though
there were eight participants holding that position. A situation with one
and only one position held by all participants is an internally egalitarian
situation even though entry into the position may be difficult. Most situations
contain more than a single position, and sets of rules assign different
kinds of authority to those in different positions. The Trust games also
explored in chapter 3 all had at least two positions: Investor and Trustee.
Differentiation of authority to act or to effect outcomes depends upon the
establishment of multiple positions within a situation.
Number of Participants in a Position
A position rule may also state whether there is a defined number, no limit,
a lower limit, or an upper limit on the number of participants who hold
a position. If all positions in a situation have a defined number of participants,
the maximum number of participants is thus also defined. A defined
number of participants in positions is used in most recreational
games or competitive sports. It is not legal to play with more or less than
the defined number of participants in each position on the field. The position
of a member of a jury must be filled by a defined number of participants
(defined by law, it is frequently twelve jurors, but it varies from
jurisdiction to jurisdiction and according to the type of case involved).
Sometimes a rule states only a lower bound or an upper bound on the
number of participants in a position. Either bound may be left undefined.
When the lower bound is defined and the upper bound is left unspecified,
  
C HA PT ER SE VE N194
a minimum number of participants must be present, but not a maximum.
Most quorum rules define a minimum number of participants who must
occupy a particular position before particular actions must be taken.
When a lower bound is not defined, action can occur without any participants
in a particular position. When the upper bound is defined and the
lower bound is left unspecified, a “lid” is placed on the total number of
participants that can hold a particular position. An operational example
of this type of rule is found in the authorizations given to many public
agencies that they can hire up to a specified number of participants in a
particular position (state trooper, for example). Such rules may or may
not specify whether there is a lower bound. When both bounds are specified,
action may not take place until at least the minimum participants
are in their position and no more than the maximum are present. Most
position rules do not specify conditions, which means that the position
rule holds under any of the likely conditions to be found in the situations.
It is certainly possible that there may be conditions added to a specific
position rule. For example, a position rule in a school district may raise
the number of teacher aides who may be hired by a school in the condition
that the number of students per teacher rises above twenty.
Where position rules do not exactly specify the number of participants,
it is entirely the operation of the boundary rules over time in conjunction
with the type of goods and attributes of a community that affect the number
of participants, their characteristics, and their ease of entry or exit.
Boundary Rules
Boundary rules—frequently called entry and exit rules—define (1) who is
eligible to enter a position, (2) the process that determines which eligible
participants may enter (or must enter) positions, and (3) how an individual
may leave (or must leave) a position. Some entry rules, then, specify
the criteria to be used to determine whether an actor is eligible to fill a
particular position. Ascribed and acquired attributes are frequently used
in this type of entry rule. Individuals may have to meet certain physical
standards, such as height and weight. They may have to meet a certain
wealth standard or pay an entry fee. Individuals may be required to possess
a certain range of experiences, to be above a minimum age, to have
graduated from certain schools, to be the descendants of a particular
group, to possess certain abilities, or to live in certain geographic areas.
Public employment under civil service systems and patronage systems differs
substantially in the entry rules that are applicable.
When those crafting a rule hope to increase the skills and knowledge
held by those in a position, they may list one or more acquired characteris-
  
C LA SS IF YI NG RU LE S 195
tics, such as holding a college degree or passing a test. Exclusive country
clubs may require that a family own a fortune in liquid assets and have
sent their children to Ivy League schools to be eligible to be members.
Immense and important issues have and will continue to erupt over
whether one or another attribute should be included in a rule. Whether
race should be used in an entry rule—biased either for, or against, individuals
of a particular race—has been a major controversy in the United
States for multiple decades (see Greve 2001).
First-order boundary rules define the eligibility of individuals to hold
the position of member. These rules include a set of transformations
that partition a defined set of individuals, usually bounded in space and
time, into subsets of individuals who are eligible and ineligible to hold
the position of member. A simple rule to partition the set is that a participant
must be eligible if the set of ATTRIBUTES of the participant equals
the required set of ATTRIBUTES and the participant meets specified
CONDITIONS. The ATTRIBUTES that may be specified can include ascribed
or acquired characteristics (see chapter 8 for examples). An alternative
formulation of the same rule is that it is forbidden to keep a person
who has certain defining ATTRIBUTES and has met required CONDITIONS
from being eligible to be a member. Since the deontic operators
are interdefinable, the same rule can be stated with any of the deontic
operators. Alternatively, one can state that an individual must not be eligible
unless they have certain ATTRIBUTES and have met defined CONDITIONS.
An example of a rule stated in this manner is the rule in the U.S.
Constitution (article 1, section 2, paragraph 2) regarding the eligibility of
citizens to be a Member of Congress: “No Person shall be a Representative
who shall not have attained to the age of twenty five Years, and been
seven Years a Citizen of the United States, and who shall not, when
elected, be an Inhabitant of that State in which he shall be chosen.”
Second-order entry rules define how the set of eligibles are further partitioned
into subsets of position-holders and nonholders. Nonholders here
are narrowly defined to include those individuals who would be eligible
to hold the position of a member in a particular situation but do not
hold that position. Boundary rules are open when they allow eligibles full
control over the decision whether or not they wish to hold a position.
Most election laws within the United States, for example, are open and
allow eligible voters (those who are above a defined age and have registered,
etc.) to decide whether or not to come to the polls to vote. All
eligible voters who appear at designated places and times are authorized
to participate in an election. Variables such as the length of residency
required for eligibility, registration procedures, absentee voting rules, and
number of hours allowed for voting on election day combine to make the
  
C HA PT ER SE VE N196
act of voting more or less expensive in terms of the amount of time, effort,
and knowledge required to be legally registered prior to election day.
Boundary rules are invitational when they authorize holders of a position
to select future holders from the set of those who are eligible. Thus,
eligibles are divided into two sets—those who have been invited and those
who have not been invited. Entry requires an invitation. Many private
clubs use a rule of this type. Most businesses and public bureaus also use
rules of this type with invitations taking the form of job offers.
Closely related to an invitational entry rule is a competitive rule where
participants are selected as an outcome of another action situation in
which potential members compete against each other to gain entry. In a
competitive election, entry requires receiving the most votes (or some
other definition, depending upon the aggregation rule used in the election
situation). The selection of legislators is normally the outcome of a separate
election in which members are selected by voters from among candidates
who are running in this election.
Boundary rules are compulsory when eligibles have no control over
whether they fill a position or not. When focusing on broad jurisdictions,
the most inclusive class of individuals within a jurisdiction is its subjects—
those subject to its rules. The status of being a subject is not conferred by
choice when the jurisdiction is a general-purpose governmental unit such
as a city, county, state, or province, or the largest, general-purpose jurisdiction,
a nation. Subjects are nonvoluntary members of a public instrumentality.
A voluntary member of an organization in the private domain
of life exercises choice concerning whether or not to join (subject to eligibility
criteria) but is subject to the rules of the arrangement while remaining
a member.
A suspect arrested by the police has no choice as to whether to participate
or not in court proceedings concerning charges made by the state.
The suspect has limited choices: what to plead, whether to hire a lawyer,
whether or not to testify on his or her own behalf. A suspect in a criminal
case cannot take any action independently to exit from the process.6
Compulsory
rules oblige anyone who meets certain CONDITIONS to hold a
position. Being drafted into the army or subpoenaed to serve on a jury
involves the selection of a member by a formal process outside the control
of any individual draftee. A person selected as a defendant in a criminal
trial has entered this position through a compulsory process. Under universal
compulsory entry rules, participation is required for all those who
are eligible to participate. All eligible taxpayers, for example, must complete
tax forms and pay any taxes they owe. A compulsory entry rule is
particular when only a subset of eligibles must become members at any
one time period. Both the draft and jury duty are particular entry rules.
  
C LA SS IF YI NG RU LE S 197
Boundary rules may also assign fees for entry and/or exit. Entry rules
that are open, invitational, or competitive may assign a fee or inducement
to any eligible individual who wishes to enter a position. An example is
an application fee or reward. Poll taxes are fees that eligible voters had
to pay in order to vote at an earlier juncture in the United States. An
example of the second is a membership fee or reward. Many competitions
also require a fee before entry.
Boundary rules may define high licensing fees or large bonding requirements
for entry or exit. These rules also interact with the physical world
and attributes of the community to shape the dynamics of entry and exit
into positions. A fee of $1 per hour to use a tennis court may be considered
a low entry cost in a middle- or upper-class community, while in a poor
community such a fee may preclude the entry of many potential users.
The level or strictness of entry and exit costs is relative to the availability
of an attribute or a resource in a community.
Governmental jurisdictions can also increase entry and exit costs via
other kinds of boundary rules such as certification procedures with many
requirements imposed on all potential entrants, or a limitation on the
number of enterprises licensed (which makes the license itself a very valuable
good like a liquor license or taxi cab medallion). So, one way of
increasing strictness involves more stringent requirements. However,
entry costs are also affected by the production technology of particular
goods. When a high fixed investment is required to produce any quantity
of the good at all, entry costs are relatively high, and the conditions of a
competitive market are rarely met.
Rules Related to Multiple Positions
When situations involve more than a single position, a mix of position
and boundary rules together define relationships between those positions.
Some rules create multiple positions (position rule), require each participant
to hold one of them (boundary rule), and forbid the holding of more
than one position (boundary rule). Such a rule set is used in many recreational
sports. Positions such as pitcher and hitter are defined. Each position
is filled by at least one member. All members hold one, and only one,
position. Such a rule set covers and partitions the set of participants in
the situation. Another rule set is used in many committee settings to assign
one and only one member a unique position, such as the chair of the
committee, and to assign all other members to the other position. The
U.S. Constitution provides an example in article 1, section 6, paragraph
2: “No Senator or Representative shall, during the Time for which he was
elected, be appointed to any civil Office under the Authority of the United
States, which shall have been created, or the Emoluments whereof shall
  
C HA PT ER SE VE N198
have been encreased during such time; and no Person holding any Office
under the United States, shall be a Member of either House during his
Continuance in Office.”
Alternatively, the rule set may assign a member to at least one position
and permit members to hold multiple positions. Such a rule is used in
many organizations when members may hold several different positions
simultaneously. A member of a firm may hold a position as a particular
worker (secretary, foreman, president) as well as a position related to
tenure in office (a probationary versus permanent staff member). This
rule set covers, but does not partition, the set of participants.
Succession Rules
Boundary rules may also define eligibility for entry to positions in terms
of rules that define who is eligible to move from one position to another
and what criteria must be met, often called succession rules. Civil service,
seniority, and patronage institutional rules differ primarily in the procedures
used and criteria applied in regard to succession of individuals into
higher-level positions. In a civil service or “merit” system, those who are
already employees must serve specified periods of time at lower-level positions
and pass examinations in order to be placed on an eligible list for
promotion to higher-level positions. When a seniority rule is used, individuals
who have been in a particular position for the longest period of time
are selected to move into higher-level positions when vacancies occur. Decisions
about upward mobility in a patronage system are made by individuals
who hold the position of “patrons.”
Exit Rules
While entry rules define who is eligible to enter a position and who has
control over entry, exit rules define the conditions under which a participant
must, must not, or may leave a position. In two-player repeated
social dilemma games, the capability to exit—leave the situation and the
position of player—has consistently been shown to make a big difference
in the rate of cooperation reached over time (Orbell, Schwarz-Shea, and
Simmons 1984; Schuessler 1989; Vanberg and Congleton 1992). Exit
allows a participant to extract themselves from having to make a choice
between defecting on someone else or being a sucker when the other participants
defect. Defendants in a criminal trial or a prisoner may not leave
such a position at their own initiative. The results of a trial may be to
release the defendant (allowing the participant in this position to get out
of this position). After a defined period of time has passed, a prisoner may
be released from this position by a parole board or may have simply
  
C LA SS IF YI NG RU LE S 199
served a set term and be automatically released. A citizen of a nation
may not have full control over leaving this position. Many countries have
placed severe constraints on the capabilities of citizens to exit voluntarily.
Rules sometimes set fixed terms of office with stringent rules concerning
the eligibility of a past position holder to be eligible to hold the same
position again. Governors of some states and mayors in some cities may
hold these positions for one term only or may not be eligible to succeed
themselves (even though they may later be eligible after someone else has
served in the office). Under such circumstances, the person in the position
has no control over retaining the position after the fixed term has expired.
For most elected positions, voters have full control over a participant’s
continuation in office. Positions like those of judges, however, may be for
life, subject only to removal for illegal or immoral behavior. Participants
in such positions are assigned very high levels of control over when and
under what conditions they leave office.
The rules related to many positions, however, give both the occupant
and others partial control over whether the occupant continues. Except
under slavery or imprisonment, occupants of most positions are allowed
to exit or resign from positions at their own initiative.7
(Particular rules
may set a limit on the amount of time that must elapse from announcing
a decision to leave and actually leaving or may set a charge associated
with leaving a position prior to fulfilling some aspects of a contract.) The
capability of a participant to leave a position is a fundamental limit on
the power that other participants can exert over a participant.
In regard to civil proceedings, the plaintiff has an initial voice regarding
the instigation of legal proceedings and also a choice concerning their
termination. If the plaintiff wishes to terminate the case by dropping the
charges, the plaintiff will probably have to pay court charges, attorney
fees, and other costs, but can otherwise freely exit. Upon termination of
many civil cases, one or all of the parties are usually assigned court and
attorney fees as part of the costs of the use of the institutional arrangements.
The costs of exit here may be very high. Litigants may use the
potential costs of a trial as a threat to reach a negotiated settlement in
the shadow of the court. When there is sufficient ambiguity as to the
outcome of the trial, both sides may be willing to negotiate outside the
courtroom in order to reduce the possibility of losing in court and paying
the exit costs.
While the holder of a position in most situations may be able to exit
voluntarily, others may also have greater or lesser control over whether
the person continues in or leaves the position. Prior to the establishment
of civil service legislation in many states, appointed local public employees
could be easily removed from office by elected officials. Changes in
the party of locally elected officials frequently meant that public employ-
  
C HA PT ER SE VE N200
ees hired by the other party were fired and new workers loyal to the incoming
party were hired. Civil service legislation changed the relative
“rights” of public employees to their positions. No longer could they be
fired at will or for lack of political loyalty and activity. After an initial
probationary period had expired, a public employee could not be fired
except for “cause.”
Collective bargaining agreements also affect the relative control that
various participants have over exit from positions. Under such contracts,
the terms and conditions of employment and firing for an entire set of
positions are negotiated at the same time. Grievance procedures may be
instituted to provide a forum and procedure for a participant who wishes
to appeal an involuntary termination. A participant holding the position
of a boss may be forced to reemploy an employee (or provide compensation)
if a termination is not considered by the grievance panel to have been
within the power of the boss. Contracts often specify rights to positions
according to seniority, which limits the power of a boss to select which
employees will be terminated during times of financial restrictions. Under
seniority rules, the last person hired into a position is the first to be laid
off, regardless of work performance.
Choice Rules
Choice8
rules specify what a participant occupying a position must, must
not, or may do at a particular point in a decision process in light of conditions
that have, or have not, been met at that point in the process. The
actions that participants must, must not, or may do are dependent both
on the position they hold, prior actions taken by others and/or themselves,
and attributes of relevant state variables.
Rules with action AIM’s partition possible actions in an action situation
into required, permitted, and forbidden acts dependent upon the path
of past actions taken by participants and others and readings on relevant
state variables. Since the classification system focuses on the most direct
link between the AIM of a rule and the components of an action situation,
not all rules with action AIMs fall into the choice rule category. When
the action of an AIM relates directly to entry or exit from a position,
giving or receiving information, joint control over a decision, or giving
or receiving payoffs, then the rule is a boundary rule, information rule,
aggregation rule, or payoff rule, respectively. Choice rules partition all
other actions that do not fit into those specific parts of the action situation.
In complex situations structured by complex systems of rules, however, a
system of choice rules may not completely partition all other possible
actions into required, permitted, or forbidden actions. The partitioning
  
C LA SS IF YI NG RU LE S 201
of actions can also be complicated by complex sets of rules that may
be inconsistent in their ordering of actions with different rules assigning
different DEONTICS to the same action. One rule may forbid an action,
while another rule requires that same action.
Reference to a jurisdiction usually occurs in either the CONDITION
of the rule that states when a particular action is forbidden, required, or
permitted or in the ATTRIBUTES of a rule that define to whom a particular
rule applies. CONDITIONS may specify where and at what time actions
taken by individuals are considered to be within the boundaries of
a particular jurisdiction with the attendant benefits or sanctions that
might result. If a carpenter injures a hand while hammering a nail on the
job, the carpenter may be eligible for Worker’s Compensation because of
the institutional arrangements encompassing the activities of the workplace.
Other individuals have a duty to pay the worker for his injury. The
same injury inflicted at home on an evening hobby project will not entitle
the carpenter to any compensation. Similarly, the penalty for some acts,
such as theft, varies radically from jurisdiction to jurisdiction. Convictions
for stealing property owned by the federal government in the United
States may carry a much stiffer sentence than conviction for stealing private
property.
In many bureaucratic action situations, no one participant is authorized
to take particular positive actions unless specific state variables are above
some minimum or below some maximum. A power-plant employee, for
example, may not be authorized to open a turbine unless water levels
are above a minimum. A social worker cannot authorize food stamps or
welfare payments unless an applicant’s income is below some defined level
given the size of the family and other conditions. Further, specific procedures
must be completed prior to any determination of the eligibility of
a family for welfare payments of any kind.
By widening or narrowing the range of actions assigned to participants,
choice rules affect the basic rights, duties, liberties, and exposures of members
and the relative distribution of these to all. Choice rules may allocate
to positions high levels of control over many different state variables; in
other words, authorize powerful positions. Choice rules empower, but the
power so created can be distributed in a relatively equal manner or a
grossly unequal manner. Choice rules thus affect the total power created
in action situations and the distribution of this power.
One particular type of choice rules, agenda control rules, proves to be
quite important in legislative games. Such agenda rules limit or expand
the authority of participants in particular positions to propose particular
actions (see Shepsle 1979b; Plott and Levine 1978). A closed agenda control
rule limits the number of alternative actions that can be decided upon.
An open rule, on the other hand, allows any feasible action to be consid-
  
C HA PT ER SE VE N202
ered. A “germaneness rule” restricts alternatives to those that affect the
same set of state variables (see Shepsle 1979b for further discussion of
these rules).
Aggregation Rules
Aggregation rules determine whether a decision of a single participant or
of multiple participants is needed prior to an action at a node in a decision
process. In many social games, particularly board games, each participant
is authorized to make a move when it is his or her turn. The player’s
action set at that juncture includes the specific physical moves to be made.
While no single player fully controls the final outcome, individual players
do control the decisions to be made at individual nodes. However, in legislative
and other group action situations, multiple participants jointly control
which actions will be taken at nodes in the decision tree. The decision
whether to amend or not amend a bill is subject to the joint control of
the members of a legislature. Individual participants affect that decision
by casting votes that are then aggregated by an aggregation rule. No single
participant has full control over the move to amend or not amend the bill.
Aggregation rules are necessary whenever choice rules assign multiple
positions partial control over the same set of action variables. The problem
that aggregation rules must clarify for a group is “who is to decide”
which action or set of activities is to be undertaken. Thus, in any action
situation in which multiple members could each potentially have partial
or total control over the selection of an action at a decision node, aggregation
rules are used to determine who will participate in the choice, how
much weight each participant will have relative to others, and the specific
formula to be used in adding up the contribution of each person’s decision
to a final decision about the action. There are many different types of
aggregation rules. Levin and Nalebuff (1995) identified sixteen distinct
methods for aggregating individual decisions into final electoral decisions.
Three major generic forms of these rules are nonsymmetric aggregation
rules, symmetric aggregation rules, and rules that define outcomes in cases
of nonagreement.
Nonsymmetric Aggregation Rules
All nonsymmetric aggregation rules treat the participants in a situation
differently in regard to some decision to be made at some point in a decision
process (Straffin 1977). Some named individual or named subgroup
is designated as the participant(s) who make the decision for the group.
Karotkin and Paroush (1994) analyze nonsymmetric aggregation rules
for a group of four participants and analyze six different nonsymmetric
  
C LA SS IF YI NG RU LE S 203
rules that vary in regard to the weights given to players. When only a
single person is assigned full authority to select the action, the person
can be called an “expert” (or a dictator) for that decision. The dictator
picks which action will be taken by the group. Such a rule gives a single
named individual the capacity to select any of the feasible actions as well
as to avoid any of the feasible actions—full active and blocking capacity.
The individual holding that position can act or make an authoritative
decision without gaining the prior agreement of others. An individual
holding such a position may, of course, consult with others prior to action.
Unless regular expectations have been established about the rules used to
aggregate the expressed preferences of others, such consultation is not
required by the aggregation rule. A single player with veto power may
not be able to direct choice as fully as a player with expert power, but the
power of such a participant is very substantial (Tsebelis 2002; Herzberg
and Ostrom 2000).
A subgroup may be named from the full group and assigned the capacity
to make a decision about actions for the entire group. The subgroup
will need an aggregation rule of its own in order to make its decision.
Such a decision rule may be called an oligarchy rule. The full set of participants
may participate in the decision, but each individual participant may
be assigned a “weighted vote.” This type of nonsymmetric aggregation
rule is used in some types of special districts where each member of a
council votes, but each is assigned a set of votes depending upon some
formula. A subgroup may be named from the full group and a decision
must be agreed to by this subgroup, as well as by the full group, using
one or more aggregation rules. Such a rule would be associated with many
“committee” arrangements wherein committees must approve legislation
before the full body gets a chance to approve the legislation. Members of
such a committee have greater voice in determining group actions.
On the other hand, an aggregation rule may grant certain positions less
voice in group decisions. In most legislative action situations, a presiding
officer is not given the authority to vote on most decisions. Thus, the act
of voting is not among the presiding officer’s set of permitted actions at
many points in a decision process, and so his or her vote does not count
in the aggregation of a regular voting process. However, if a tie occurs in
a vote by regular members, the presiding officer is then authorized to vote
in order to break the tie.
Symmetric Aggregation Rules
Symmetric aggregation rules assign joint control over an action to multiple
participants so that all are treated alike. One symmetric aggregation
rule is that of unanimity—everyone must agree prior to action. A unanimity
rule may be built into a process in such a manner that participants do
  
C HA PT ER SE VE N204
not self-consciously “vote,” but each is required to agree before an action
can be taken. A bank clerk, for example, is not authorized to open safety
deposit boxes unless the owner of the box or an authorized agent signs a
registration form and produces a second key to fit the box. The dual and
equal authority and unanimous aggregation rule are built into the locking
mechanism that requires two keys assigned to different individuals to
open the box. Similar conjunctive authority to act and unanimous aggregation
rules occur in the military when the results of action could be
extremely serious for national security. In addition to the necessity of
receiving positive approval from positions higher in the military hierarchy,
taking some actions—such as launching an intercontinental missile—
cannot physically be undertaken unless multiple persons are present and
all agree.
Once the votes are cast, rules specify what proportion of the total must
be in agreement before an authoritative decision can be made and what
happens if the minimal agreement is not reached. For votes that are
weighted equally, it is possible to conceptualize a simple voting rule as
ranging from allowing any one member of those given joint authority to
make the decision for the collectivity (the anyone rule—used in calling
out public emergency vehicles) to requiring all those given joint authority
to agree prior to a decision (the unanimity rule) (Buchanan and Tullock
1962). Between the two extremes of the anyone rule and the unanimity
rule lies a variety of other specific rules, the most familiar being the requirement
that 50 percent plus 1 person agree (majority rule) or some
large percentage, such as two-thirds or three-quarters (an extraordinary
majority rule). The array of decision rules between these two extremes
can be thought of as the proportions of the persons in the group required
to agree prior to a decision with the most common being majority rule.
Scholars have puzzled why participants in some jurisdictions or legislative
bodies have used a simple majority voting rule to create a new situation—for
example, to approve or not approve a municipal bond—while
others use a supramajority vote rule such as two-thirds of those voting.
Messner and Polborn (2004) attack this problem using the example of
bond issues and show that the voters who may form a simple majority at
time one may not be in the majority five years, ten years, or twenty years
later. Voters who expect a change in population composition and preferences
related to policy issues over time may well use their simple majority
vote at one time to ensure that the size of the winning coalition in the
future requires a supramajority, precisely because this is a more conservative
voting formula. Looking ahead, they fear that their preferences will
be overwhelmed in future votes on costly bond issues if a simple majority
rule is retained.
  
C LA SS IF YI NG RU LE S 205
Lack of Agreement Rules
The formula for determining a joint decision (for either symmetric or
nonsymmetric rules) must also include a “no agreement” condition stating
what decision will happen if no agreement is reached under a rule.
Whenever a decision depends on the approval of more than one participant,
the possibility of no agreement is always present. The no agreement
condition is the reference point for the proposed decision. It states what
will happen if a certain proportion of the participants does not agree to
a proposed action.9
Several types of no agreement rules are possible. One type continues
the status quo distribution of outcome variables. A second presumes that
no one receives any outcome variables if participants cannot agree (all
relevant outcome variables are reduced to zero). A third no agreement
rule is to assign state variables randomly. A fourth type of rule is to apply
some external rule (or turn to some external decision maker) to allocate
outcome variables.
As mentioned in chapter 6, Grether, Isaac, and Plott (1979) used three
“no agreement rules”—which they called default conditions—in a series
of laboratory experiments simulating the allocation of landing slots at
airports. The first rule was the continuation of the status quo allocation
of slots. The second rule was a random assignment of slots. The third rule
was the application of an external rule that would take slots from those
who had the most and give them to those who had none or only a few (a
Robin Hood rule!). These three no agreement rules were combined with
a unanimity rule. What is interesting about the findings from a set of
experiments using each of these no agreement rules, combined with unanimity,
is that the results of the experiment were strongly determined by
which no agreement rule was in use. The committee outcomes were substantially
influenced by the specific lack of agreement rules used in an
experiment (V-7).
That the specific lack of agreement rules that Grether, Isaac, and Plott
used had such an important impact on the outcomes achieved is another
illustration of how one small part of a rule configuration can strongly
affect how the other rules impact the action situations. Scholars need to
study the effects of a full rule configuration rather than assume they can
study the impact of one rule at a time—while assuming that the other rules
are “randomly distributed.” A considerable literature has been generated
about the likely effects of using unanimity rules in different situations.
Very few theorists have explicitly stated the no agreement rule they are
assuming. Yet, one might speculate that once an analyst assumes a unanimity
rule, that the most important assumption driving analysis is the
aggregation no agreement rule presumed in operation. Implicit assump-
  
C HA PT ER SE VE N206
tions about rules, rather than the explicit rules studied, may have been
the most important drivers of the results in earlier analyses of institutional
arrangements.
Information Rules
An important part of any action situation is the information available
to participants about the overall structure of that situation, the current
state of individual state variables, the previous and current moves of other
participants in positions, and their own past moves. Information rules
affect the level of information available to participants. Information
rules authorize channels of information flow among participants, assign
the obligation, permission, or prohibition to communicate to participants
in positions at particular decision nodes, and the language and form in
which communication will take place. Information rules are particularly
important in generating information about past actions of participants
so that other participants can know who is, or is not, trustworthy (see
Janssen 2004).
Channels of Information Flow
Rules concerning the establishment of information channels relate to the
set of all possible channels connecting all participants in a situation. The
connections can be represented as a perfectly connected polygon of whatever
dimension equals the number of participants. If there are five participants,
there are nine possible connections between these participants. Information
rules partition this set of possible connections into subsets of
required (a channel must exist), forbidden (a channel must not exist), and
permitted (a channel may exist). In a paper by Mueller, Chanowitz, and
Langer (1983), for example, they conduct several experiments (at the action
situation level) of communication patterns under different structures
connecting five subjects.
Given that the number of potential communication channels among
any large group is very large and the set of possible rules requiring, permitting,
or forbidding channels is also very large, it will be necessary to identify
specific types of channel rules if one wishes to devise a finer classification
scheme for this aspect of information rules.
Frequency and Accuracy of Communication
In addition to specifying which channels of communication may or may
not exist between positions in a situation, information rules also regulate
the frequency of exchange of information and the accuracy of informa-
  
C LA SS IF YI NG RU LE S 207
tion. In many action situations, regular reports must be filed containing
certain types of information on either a regular basis or at any time that
a participant wishes to obtain certain actions or rewards from others. A
person who is on probation is supposed to report to a probation officer
on a regular basis and provide a report about their conformance or nonconformance
to a set of rules about the actions that they can or must not
do. Most bureaucratic life is filled with requirements to complete regular
reports about recurrent events in these organized settings. The accuracy
rules affect what type of indicators may or must be used as evidence about
the state of the world. Rules establishing audit procedures are intended
to enforce the accuracy of financial information available to top management
and shareholders of a firm.
Subject of Communication
Information rules often limit the topics that can be discussed among participants.
In a courtroom, a witness is forbidden to refer to “hearsay” evidence.
In industrial meetings, participants are not supposed to discuss
price-setting decisions. In many laboratory experimental sessions, rules
frequently limit the subjects that participants can discuss with instructions
like the following: “Some participants in experiments like this have found
it useful to have the opportunity to discuss the decision problem you face.
You will be given 10 minutes to hold such a discussion. You may discuss
anything you wish during your 10-minute discussion period, with the following
restrictions: (1) you are not allowed to discuss side payments (2)
you are not allowed to make physical threats (3) you are not allowed to
see the private information on anyone’s monitor” (E. Ostrom, Gardner,
and Walker 1994, 150–51).
Official Language
Information rules also often specify the official language for communication
in a situation. These types of rules are quite familiar to us in international
settings where there is always an official language in which the
business of an international organization or conference will be conducted.
All nations also have their official languages. But organizations also establish
official languages, including coding systems assigned to products, customers,
order numbers, invoices, and the like.
Payoff Rules
Payoff rules assign external rewards or sanctions to particular actions that
have been taken or to particular readings on outcome state variables. An
  
C HA PT ER SE VE N208
example of a set of payoff rules is the pay schedule that is used by a
government agency or by a private firm to assign salaries to participants
in particular positions. This payoff schedule will vary in terms of the variables
taken into account and the complexity of the schedule. Hourly wage
payoff rules frequently are very simple and specify a computation of a
wage for a certain number of hours considered to be the official working
hours of an employee during a set period. Someone being paid according
to piecework will, on the other hand, be paid by formula, attaching a
weight to a quantity of intermediate or final goods attributed to the work
of the participant. Performance contracts for corporations frequently are
very much more complex. A contract may state that a corporation will
receive x amount if some physical transformation in the world (like a
particular apartment building) is completed to someone else’s satisfaction
by a set date. If the time period is greater than x, then the payment is
reduced according to a formula including the amount of time of delay. In
the field, many payoff rules that involve costs assigned to a forbidden
action (sanctions) involve a low cost for the first infraction, a higher cost
for the second, and further graduated sanctions leading to large costs
assigned for repeated infractions.
Payoff rules have an AIM that involves paying or receiving something
of potential value. As figure 7.1 shows, payoff rules directly impact the net
costs and benefits of action or outcomes for actors in an action situation.
However, payoff rules often are not the only rules that shape costs and
benefits. One could discuss payoff consequences of boundary rules tied
to assignment of actions to positions, payoff consequences of information
rules, as well as payoff consequences tied to choice and scope rules.
Scope Rules
The above sections have focused on the rules that affect the deontic status
of the actions that could physically be taken by a participant in a situation.
Scope rules affect a known outcome variable that must, must not, or may
be affected as a result of actions taken within the situation. Scope rules
define this set, affect the width of the outcome space (number of state
variables affected), and specify the range on each outcome variable included
in that space. The AIM component in scope rules describes an
outcome rather than an action. Since our classification scheme focuses on
the most direct link between the AIM and a component of the action
situation, rules with outcome AIMs directly tied to positions, boundaries,
information, payoffs, or aggregation would be classified as position,
boundary, information, payoff, and aggregation rules, respectively, leaving
all other rules with outcome AIMs in the scope rules category. Thus,
  
C LA SS IF YI NG RU LE S 209
the scope rule category and the choice rule category both work as “all
other” categories. If a rule is not a position, boundary, information, payoff,
or aggregation rule, then it is either a choice rule (if the AIM is an
action) or a scope rule (if the AIM is an outcome). Heiner (1990) focuses
entirely on rules that place constraints on the production possibility frontier
and not on actions in his “Rule-Governed Behavior in Evolution and
Human Society.” Thus, implicitly his analysis of rules constraining the
range of outcomes relates to what we are calling here scope rules.
Because scope rules affect action sets through their effect on outcome
variables, they do not directly enumerate action sets. Rules related to the
operations of cable television stations have frequently listed upper and
lower bounds on particular state variables that a station is forbidden to
exceed. In some situations, the physical world or attributes of the community
may make monitoring of actions more difficult or sensitive than monitoring
outcomes, which would then likely make scope rules more appropriate
than choice rules. The attributes of community may stress
autonomy and liberty, which may make it easier to gain legitimacy for
scope rules than choice rules.
For example, strong norms and rules governing academic freedom render
many university rules governing specific teaching and research activities
suspect, while rules that base promotions on outcomes for professors
may be seen as much more legitimate. Rules generally tie tenure and promotion
most heavily to outcome measures (quantity and quality of publications,
quality of teaching overall), with less evaluative attention given
to choices of actions. Similarly, gaining legitimacy for government regulation
from business communities with strong antiregulation views may be
easier with scope rules that allow firms to choose how to comply.
A recent workshop was held in Washington, D.C., to examine when
scope rules are most likely to be effective regulatory tools. The organizers
identified two basic types of rules that can be used in regulating industry:
“Regulators can direct those they govern to improve their performance in
at least two basic ways. They can prescribe exactly what actions regulated
entities must take to improve their performance. Or they can incorporate
the regulation’s goal into the language of the rule, specifying the desired
level of performance and allowing the targets of regulation to achieve that
level” (Coglianese, Nash, and Olmstead 2003, 706).
The first alternative focuses on regulation via choice rules. The second
alternative—and the one that was examined in some depth at the workshop
described above—is via scope rules. Even though considerable emphasis
in Washington has been on using scope rules rather than choice
rules,10
Coglianese, Nash, and Olmstead point out that there has been
relatively little empirical study of the impact of setting performance targets.
This may be related to lower numbers of scope rules in the real world
  
C HA PT ER SE VE N210
because of the difficulty that regulators face in measuring performance
rather than in determining whether forbidden actions were indeed taken.
Many of the agency officials participating in the discussion were attracted
to the idea of performance-based regulation, but pointed out that it was
difficult to move away from regulation focusing on permitted or forbidden
actions.
In problematic situations, the most immediate questions raised frequently
focus on the behaviors (the actions) that need to change in order
to improve the situation, and so choice rules would tend to be the most
obvious direct response. However, growing emphasis on outcome measurement
in industries, in foundation grants, and in government programs
may well lead to more attention to reforms that use scope rules. Experience
in the field and in formal theory has focused less on scope rules than
on choice rules.11
At this point, we find many fewer instances of explicit
scope rules and thus less need to further divide scope rules into specific
types. If outcome regulation continues to expand, and thereby increases
the need for policy scholars to study scope rules, more attention will need
to be paid to the types of scope rules that exist.
Default Conditions: What Happens if No Rules Exist Related
to Components of an Action Situation?
Rules do not always exist related to all elements of an action situation.
What should a participant—or an observer—deduce about the structure
of a game in the absence of rules affecting all seven parts of an action
situation? This is particularly important due to the configurational nature
of rules. One needs to know the basic rules related to a full rule configuration,
rather than to a single rule, to infer both the structure of the resulting
situation and the likely outcome of any particular rule.
In chapter 5, we identify the default conditions for the internal components
of a rule. A rule that does not list any specific attribute or condition,
for example, applies to all attributes and conditions that the physical
world and the community make possible. It is also useful—and even necessary—to
define the default rules for each of the seven types of rules. The
notion of default conditions allows us to address the question of what
happens if no rules exist related to a specific component of an action
situation (see E. Ostrom, Gardner, and Walker 1994 for an earlier exploration
of this question).
The seven default conditions that we specify in table 7.2 are those that
would be used by a participant or an observer in a general legal system
that presumed general freedom unless a rule specifically prohibited or
mandated an act or event. These are the broadest default conditions that
  
C LA SS IF YI NG RU LE S 211
TABLE 7.2
Default conditions
Default Position Condition One position exists.
Default Boundary Condition Anyone can hold this position.
Default Choice Condition Each player can take any physically possible
action.
Default Aggregation Condition Players act independently. Physical relationships
present in the situation determine
the aggregation of individual moves into
outcomes.*
Default Information Condition Each player can communicate any information
via any channel available to the player.
Default Payoff Condition Any player can retain any outcome that the
player can physically obtain and defend.
Default Scope Condition Each player can affect any state of world
that is physically possible.
* If a rule configuration contains only a default choice condition, the default aggregation
condition must be present.
would be used in a common-law legal system as contrasted to a Roman
Law system that presumes that most things are forbidden unless specifically
permitted. These default conditions are listed in table 7.2.12
If one
were to analyze a situation where there were no rules—and thus the rule
configuration would contain only the default conditions—the resulting
configuration constitutes a Hobbesian state of nature.
One way of representing such a game of nature is the Snatch Game
presented in chapter 2 (figure 2.2). The predicted equilibrium of this game
is grossly inefficient—no exchanges. If the players do adopt internal
norms, as discussed in chapter 4, they may be able to achieve an efficient
outcome of an exchange. But the value of the negative delta parameter
assigned to snatching the offered goods has to be larger than the value of
the goods snatched. Otherwise, even individuals who hold norms against
snatching goods from others may end up doing so when the value of
goods is very high.
Some analysts might respond that it is easy to fix this situation—simply
impose private property rights! But the question is, How does one create
private property? What rules need to be changed to create “private property”?
If all the households in a region were to get together and create a
legal system defining their ownership of goods and the punishment for
stealing, they would need to establish at least four kinds of rules: position,
boundary, choice, and aggregation. A brief summary of one possible set
of these rules is shown in table 7.3.
  
C HA PT ER SE VE N212
TABLE 7.3
Rules changed to create elementary property rights for agricultural commodities
Position Rules There exist two positions: (1) an eligible exchange participant
and (2) a judge.
Boundary Rules (1) All farmer households are permitted to become exchange
participants or else those refusing their entry may
be punished.
(2) The judge must be selected on the basis of merit and
integrity by the households in the community or else the
other rules will not be in effect.
Choice Rules (1) All exchange participants are permitted to offer to exchange
goods they own for goods owned by others or else
those forbidding the exchange may be punished.
(2) If a household’s goods are snatched, the household
can report to a judge or else those preventing the report may
be punished.
(3) If a judge finds that a household has snatched goods
illegally,* the judge must ensure that the illegal household
returns the goods and forfeits its own commodities or else
the judge will be sanctioned.
Aggregation Rules All parties to an exchange must agree before a legal exchange
can occur or else the exchange does not occur.
* Now we could use the term stealing instead of simply snatching.
With these multiple changes of seven rules from their default status
to the simple rules stated above, combined with the remaining default
conditions, a new operational-level game—shown in figure 7.2—is created.
In this game, a new move has been created. If Household 2 snatches
Household 1’s goods, Household 1 has the option of going to a judge
(who is not modeled here as a player since this simple game assumes the
judge complies with the rule that requires him to return the commodities
to their rightful owner and to make the household who stole the goods
forfeit their own commodities as well). Given this option now made available
to Household 1, Household 2 is unlikely to snatch the goods since
Household 1 would be motivated to use this option if they reached this
part of the game. Instead, Household 2 will select the exchange option
where both players are better off.
Of course, this is a highly simplified version of both the action situation
and the rule configuration that could be adopted to avoid the deficient
outcome of the Snatch Game in a State of Nature.13
However, even in this
highly simplified example, several important points have been illustrated.
These are:
  
C LA SS IF YI NG RU LE S 213
(10, 10)
(5, 20)
(15, 15)
Snatch
(10, 10)
Do Not Offer
No Action
Offer
Household 2
Exchange
Household 1
Household 1
Do Nothing Go to Judge
(10, 0)
Figure 7.2 The Snatch Game with minimal property rights.
1. In the absence of any rule directly affecting an element of an action situation,
the relevant rule in place can be described by a default rule.
2. When all rules are in their default, the attributes of the physical world
generate all aspects of the structure of the action situation. This is the Hobbesian
“state of nature.”
3. Rules operate together with the attributes of a physical world to create a
structure.
To change the game from the one shown in figure 2.2 requires changes
in at least seven rules. If the choice rule mandating the action of the judge
had only given the judge permission to return goods to those from whom
they were stolen and to confiscate the goods of an offender, a different
predicted outcome would be likely. It would depend on whether Household
2 thought there were any factors that would affect the likelihood
that the judge would not impose the punishment if they snatched the
goods. If the head of Household 2, for example, was the brother of the
judge, and the choice rule of the judge stated only a permission to impose
a punishment rather than a requirement, a different calculation would
be made by Household 2. And, it needs to be stressed, the default conditions
jointly create the structure of the situation along with the rules (as
well as the relevant characteristics of the biophysical world and relevant
community).
To keep the initial analysis as simple and clear as possible, the Snatch
Game has been represented as a game between households living in the
same region. The problem of gaining a commitment that goods—once
  
C HA PT ER SE VE N214
presented for trade—would not be snatched is a widespread problem underlying
the development of commercial relationships in general and longdistance
trade in particular. When the presumption is made that you solve
this problem by creating property rights and then judges and police to
enforce these rights, the judges and police are viewed as outside the situation
(as shown above by not viewing the judge as a player). These are
automatic agents assigned to do a job reliably and not viewed as strategic
actors in many analyses.
A key problem is that once you have created private property, markets,
and the role of a judge or “ruler,” how is the holder of this position going
to undertake the assigned duties?14
The judge or ruler will need to obtain
funds to organize the court, keep records, maintain a police force to provide
patrols in the marketplace, search the storehouse of merchants suspected
of stealing, and so forth. One way to do this is to impose a fee on
the merchants who want to use a market. Both rulers and merchants have
a mutual interest in the evolution of a safe place for trading goods. The
greater the number of merchants that come to a market and the more
diverse the goods they bring with them, the more valuable the market
becomes both for the merchants and for the ruler. This is especially so if
merchants from far away can be lured who bring goods that are not locally
available. Now the ruler faces a substantial temptation to pledge
security to traders from other regions and get them to come. Once a lucrative
trade is established, however, a ruler also faces a temptation to renege
on that costly promise or even use his or her own military forces to obtain
the valuable goods for the ruler’s household.15
So life goes on. No change
in a set of rules is ever sufficient to solve the next set of problems created
by new opportunities and constraints that continually arise in an evolving
human community.
The Vertical Approach: Operational, Collective-Choice,
and Constitutional-Choice Levels of Analysis
In the IAD framework, authoritative relationships occur in collectivechoice
and constitutional-choice situations. Using the IAD framework,
we focus on operational rules that affect authorized relationships and
on collective-choice rules and constitutional-choice rules (both of which
Commons and Hohfeld considered to be authoritative relationships). This
vertical approach recognizes that rule sets are themselves nested in hierarchical
levels. The participants in operational situations are directly affected
by the operational rules structuring what they must, must not, or
may do. These rules were crafted in a collective-choice situation structured
by collective-choice rules (which participants, in what positions,
  
C LA SS IF YI NG RU LE S 215
chosen how, given information, and an assessment of benefits and costs
can make operational rules). The collective-choice rules were themselves
crafted in a constitutional situation.
Vertical arrangements do not exist just within government. These relationships
also exist in the contractual arrangements among private individuals
as well as within a variety of governmental decision-making arrangements
for determining, enforcing, and altering authorized legal
relationships (operational decisions). Under contractual arrangements,
people may participate in a mutually agreeable arrangement in redefining
and altering legal relationships in order to accomplish the objectives of
some undertaking of mutual interest. A large literature drawing on the
work of Coase (1937), Milgrom and Roberts (1992), Williamson (1985,
2000), and Schmid (2004) has explored a variety of conditions under
which individuals voluntarily move from horizontal to vertical relationships.
Governmental arrangements, inherent in the operations of courts,
executive agencies, and legislative bodies, enable people to sustain efforts
to determine, enforce, and alter legal relationships. Each particular decision-making
structure or decision-making arrangement, then, involves a
complex set of rules regarding the variety of participants who may wish
to pursue their strategic opportunities in order to realize some outcomes
that may be made through those decision structures.
Using Rules as Tools to Change Outcomes
Changing the Snatch Game into a game with property rights demonstrates
how even a relatively simple policy reform requires multiple rule changes.
The AIM classification system introduced here provides a rough mechanism
for simplifying the complex array of rules in any situation. Even
with this system, complexity remains. We cannot just assume that boundary
rules affect only who fills positions, for example, without examining
how restricting or opening entry affect the other components of an action
situation. So, while the tools in chapters 5 and 7 allow us to sort institutional
statements and define default conditions, there is still much to be
learned from the diversity of institutional structures in field settings. What
kinds of institutional structures do participants in field settings use? How
do these participants use rules to change the structure of the situations
that they face? Chapter 8 demonstrates the widely diverse rules that resource
appropriators use to craft institutions for governing and managing
common-pool resources.
  
  
Part III
W O R K I N G W I T H R U L E S
  
  
Eight
Using Rules as Tools to Cope with the Commons
IN CHAPTER 7, we described the seven generic rules that individuals use
when establishing or changing action situations they confront in everyday
life. Chapter 7, hopefully, provided the reader with a useful overview of
the tools that individuals use in creating structure in the multiple action
situations they face in life. Chapters 8 and 9 will apply these tools, and
the framework developed in the earlier chapters of this book, to a focused
study of common-pool resource problems.
Common-pool resource problems are among the core social dilemmas
facing all peoples (see discussion in chapters 1 and 3). Collective action
is required to establish and enforce rules limiting the appropriation of
water, fish, forest products, pasturage, and other resource products. These
problems are portrayed in many contemporary policy textbooks as relatively
simple problems that can be solved analytically to achieve optimal
outcomes. Empirical research over the past decades has led to a realization
of their complexity. Applying the Institutional Analysis and Development
(IAD) framework developed in the first part of this book helps one to
recognize the source of this complexity. The many variables of the biophysical/material
world, the communities involved, and the rules-in-use
combine to affect the structure of appropriation situations, the patterns
of interactions among appropriators from a common-pool resource, and
the outcomes achieved. Those who try to solve these problems have to
cope with complexity as well as coping with the commons.
In the first section of this chapter, we will briefly review the empirical
literature that documents many successes as well as failures of the diverse
ways that resource appropriators have coped with the commons. In light
of a general overview of empirical evidence, we will then apply the classification
system developed in chapter 7 to examine the diversity of rules
that are used in efforts to govern and manage common-pool resources all
over the world.
After explicating the specific rules discovered in extensive field research
by many scholars, the third section of this chapter will examine the disjuncture
that exists between the policy recommendations frequently made
to overcome commons dilemmas and the types of interventions that appropriators
themselves have adopted. Three broad assumptions underlie
  
C HA PT ER EI GH T220
many contemporary policy recommendations that are challenged by evidence
from the field. One is that resources are so interconnected that they
all need to be managed centrally. A second of the challenged beliefs is that
resource appropriators are not themselves capable of designing rules to
sustain resources over time. Somehow, public officials are presumed to
have the capabilities that the resource appropriators themselves lack.
Third, it is presumed that designing rules to improve outcomes is a relatively
simple analytical task that is best done by objective analysts.
In light of the diversity of rules actually used to cope with commonpool
resource problems, one must conclude that conducting an analytical
search for the optimal combination of rules is an impossible task for appropriators,
officials, or policy analysts. No one can do a full analysis of
the combination of rules potentially available and how they might interact
with attributes of the biophysical world and the relevant community. All
efforts to solve resource problems need to be viewed as experiments based
on partial analyses of specific problems. Theory and evidence play a key
role in increasing the probability of selecting rules leading to better as
contrasted to worse outcomes. Theory cannot, however, eliminate the
need to view all policies as ongoing experiments that need to be monitored,
evaluated, and adapted over time.
What factors lead appropriators to think about experimenting with
rules and what calculations do they use in making such decisions? Most
of the policy literature is silent on these questions since the general
presumption is made in this literature that making policies is what government
officials, rather than those who are directly affected by problems,
do. Fortunately, many field researchers have identified variables that are
related to local efforts to self-organize and craft better rules related to
local settings. In light of this research, the last section of the chapter
examines the attributes of resources and of appropriators that affect the
likelihood of collective action and develops a theoretical structure for analyzing
how appropriators may be led to change one or more of the rules
they use in relation to a resource. The analysis does not lead to the conclusion
that if left alone, resource appropriators will always, or even most
frequently, undertake the difficult task of experimenting with rules. Instead,
we will posit a set of conditions of a resource and of the resource
appropriators that are most conducive to self-organization. The absence
of these attributes leads to situations that are the least conducive to self-
organization.
Local self-organization can also be dominated by local elites to use rules
as tools to advantage themselves. That problem leads to a discussion in
the final chapter of the book of the design principles that characterize
robust institutions for governing and managing common-pool resources,
the factors that threaten the long-term sustainability of any such system,
  
U SI NG RU LE S A S T OO LS TO CO PE 221
and the need for polycentric institutions at multiple levels to cope with
poorly performing systems at any level through redundant, multitier governance
systems.
Field Research on Common-Pool Resources
A large number of field studies have found that local groups of resource
users, sometimes by themselves and sometimes with the assistance of
external actors, have managed to create viable institutional arrangements
for coping with common-pool resource problems.1
These empirical
studies document successful self-organized resource governance systems
in diverse sectors in many parts of the world. At the same time, some
commons dilemmas have continued unabated (see, for example, Finlayson
and McCay 1998; Berkes et al. 2001), and some common-property
institutions have experienced drastic negative changes during the last century
(Seixas and Berkes 2003). Further, some common-property institutions
have been converted by local leaders into private property institutions
enforced by governmental officials (Ensminger and Knight 1997;
Mwangi 2003).
Another important set of findings is that national governmental agencies
have been notably unsuccessful in their efforts to design effective and
uniform sets of rules to regulate important common-pool resources across
a broad domain (Ascher 1995). The harmful effects of nationalizing forests
that had earlier been governed by local user-groups have been well
documented for Thailand (Feeny 1988), Africa (Shepherd 1992; Thomson
1977; Thomson, Feeny, and Oakerson 1992), Nepal (Arnold and
Campbell 1986), and India (Gadgil and Iyer 1989; Grafton 2000; Jodha
1990, 1996). Similar results have occurred in regard to inshore fisheries
taken over by state or national agencies from local control by the inshore
fishers themselves (Cordell and McKean 1992; Cruz 1986; Dasgupta
1982; Higgs 1996; Pinkerton 1989).
Many developing countries nationalized their land and water resources
between the 1950s and 1970s. The institutional arrangements that local
resource users had devised to limit entry and use lost their legal standing.
The national governments that assumed these new and difficult tasks
lacked adequate funds and personnel to monitor resource use effectively.
They frequently turned to private forestry firms to gain revenue from these
resources. Governments in these countries wanted to convert commonpool
resources to a de jure government-property regime, but their actions
frequently resulted in de facto open-access regimes (Arnold 1998; Arnold
and Stewart 1991). The incentives of an open-access commons were accentuated
since local users had specifically been told that they would not
  
C HA PT ER EI GH T222
receive the long-term benefits of their own costly stewardship efforts.
Johnson and Forsyth (2002) illustrate the longer-term problem that these
earlier interventions generated by examining the efforts of the Thai government
to establish communal rights to forest access through legislation.
The potential effectiveness of the legislation has, however, been effectively
undermined by the earlier state interventions that supported commercial
interests against local interests.
Tang (1992), Lam (1998), and Joshi et al. (2000) have all found that
large-scale government irrigation systems do not tend to perform at the
same level as smaller-scale, farmer-managed systems (see also Mehra
1981; Levine 1980; Bromley 1982; Hilton 1992). In a study of over one
hundred irrigation systems in Nepal, Lam (1998) found that in terms of
cropping intensity and agricultural yield, crudely constructed irrigation
systems using mud, rock, timber, and sticks significantly outperform systems
built with modern concrete and iron headworks operated by national
agencies.
Extensive empirical research has thus found resource appropriators potentially
capable of self-organizing to manage common-pool resources
sustainably while many centralized government systems have performed
less effectively than presumed according to much of the policy literature.
One cannot assert, however, that all local efforts work well and all largescale
efforts work poorly. From the empirical literature only one conclusion
is tenable: averting the overuse and destruction of common-pool resources
used by many individuals is a challenge. It cannot be assumed
that these problems will be solved by an automatic process. Overcoming
commons dilemmas is always a struggle (Dietz, Ostrom, and Stern 2003).
What Rules Are Found in Self-Organized Common-Pool
Resource Regimes?
A team of researchers at the Workshop in Political Theory and Policy
Analysis has read and archived much of the extensive case study literature
on local common-pool resources written by anthropologists, agricultural
economists, ecologists, historians, political scientists, and sociologists (see
<http://www.indiana.edu/~workshop/wsl/wsl.html>; Hess 1999). Using
the Institutional Analysis and Development (IAD) framework elucidated
in chapters 1 and 2, we developed structured coding forms to help us
identify the specific kinds of action situations faced in the field as well as
the types of rules that users have evolved over time to try to govern and
manage their resource effectively (E. Ostrom et al. 1989; E. Ostrom,
Gardner, and Walker 1994). In order to develop standardized coding
forms, we read hundreds of cases describing how local common-pool re-
  
U SI NG RU LE S A S T OO LS TO CO PE 223
sources were or were not regulated by a government agency, by the users
themselves, or by a nongovernmental organization (NGO).
This section provides an overview of the diversity of specific rules that
we recorded using the general classification described in chapter 7. I do
this for several purposes:
1. To illustrate the immense creativity of individuals coming from all stations
of life living in all parts of the world. Their creativity and entrepreneurship are
frequently unrecognized since many resource appropriators engaged in problem
solving are not well educated and have only local reputations.
2. To show the incredible diversity of rules that individuals have adopted in
one or more settings.
3. To challenge the assumption frequently made by policy analysts that it is
routinely feasible to conduct complete analysis of a problem and develop “the
optimal” set of rules for solving that problem.2
4. To illustrate the benefit of classifying rules by their AIM. While we cannot
point to an optimal set of rules usable in most commons dilemma situations,
we can now use a common language for identifying rules that is functional for
academics as well as those directly involved in solving problems.
To understand the tools that appropriators use in the field, let us examine
the specific boundary, position, choice, and payoff rules used in commonpool
resource situations around the world initially identified in a metaanalysis
conducted by Arun Agrawal, William Blomquist, Edella Schlager,
Shui Yan Tang, and myself (see E. Ostrom et al. 1989). These four clusters
of rules are the major tools used everywhere to affect commons dilemmas.
Information, scope, and aggregation rules are additional tools used to complement
changes induced by these four types of rules.
Affecting the Attributes of Users through Boundary Rules
As discussed in chapter 7, boundary rules define the attributes and conditions
required of those who enter a position in an action situation. In field
settings, many action situations are involved but I will focus attention
on the appropriation situation: Who appropriates (harvests) how many
resource units from which common-pool resource? Boundary rules, thus,
define who has a right to enter and use a resource as an “authorized appropriator”—the
term we will use for this most general position that exists
in multiple settings. Boundary rules affect the types of participants
with whom other participants will be interacting related to a particular
resource.
If contingent cooperation is perceived to be a possibility, then an important
way to enhance the likelihood of using reciprocity norms is to
increase the proportion of appropriators who are well known in a com-
  
C HA PT ER EI GH T224
TABLE 8.1
Attributes and conditions used in boundary rules to define who is authorized to appropriate
from a common-pool resource
Attributes Conditions
Residency or Membership Personal characteristics Relationship with resource
National Ascribed Use of specified technology
Regional Age Continued use of resource
Local community Caste Long-term rights based on
Organization (e.g., co-op) Clan Ownership of a proportion of
Class annual flow of resource
Ethnicity units
Gender Ownership of land
Race Ownership of nonland asset
Acquired (e.g., berth)
Education level Ownership of shares in a
Skill test private organization
Ownership of a share of the
resource system
Temporary use-rights acquired
through
Auction
Per-use fee
Licenses
Lottery
Registration
Seasonal fees
munity. These participants have a long-term stake in that community and
would find it costly to have their reputation for trustworthiness harmed
in that community. Reducing the number of users, but opening the resource
to strangers willing to pay a license fee, as is frequently recommended
in the policy literature, introduces appropriators who lack a longterm
interest in the sustainability of a particular resource. Using licenses
to regulate entry increases the number of strangers using the resource and
may reduce the level of trust among participants and their willingness to
use reciprocity and thus increase enforcement costs substantially.
From our initial reading and our own fieldwork, we expected to find
boundary rules that focused on local residency as a way of increasing the
opportunity for reciprocity, and these rules were used extensively. What
amazed us, however, as we read the extensive number of case studies describing
field settings, was the variety of attributes and conditions used
to define who could be an authorized appropriator from diverse inshore
fisheries, irrigation systems, and forests. As shown in table 8.1, we identified
twenty-three attributes of individuals and thirteen conditions de-
  
U SI NG RU LE S A S T OO LS TO CO PE 225
scribed by case-study authors as having been used in at least one commonpool
resource somewhere in the world (E. Ostrom et al. 1989). While
some systems use only a single attribute or condition, many use two or
three of these rules in combination.
Boundary rules that are used in the field can be broadly grouped in
three general classes related to how individuals gain authority to enter
and appropriate resource units from a common-pool resource. The first
type of boundary rule focuses on residency or membership requirements.
These include an individual’s citizenship, residency, or membership in a
particular organization. Many forestry and fishery user groups require
members to have been born in a particular location. A second broad
group of attributes relates to ascribed or acquired personal attributes.
User groups may require that appropriation depends on age, ethnicity,
clan, caste, and/or education. A third group of boundary rules relates to
the conditions of use relating an appropriator with the resource itself.
Using a particular technology or acquiring appropriation rights through
an auction, a lottery, or purchases of land or livestock are examples of
this type of condition.
For the forty-four case studies of inshore fisheries in many parts of
the world for which sufficient information existed, Schlager (1994, 258)
identified thirty-three user groups as having at least one boundary rule
regarding the use of the resource. All thirty-three groups depended on
some combination of fourteen attributes or conditions. None of these
groups relied on a single attribute or condition. Thirty out of thirty-three
groups limited fishing to those individuals who lived in a nearby community,
while thirteen groups also required membership in a local organization.
Consequently, most inshore fisheries organized by the users themselves
restrict fishing to those individuals who are well known to each
other, have a relatively long-term time horizon, and are connected to one
another in multiple ways (see Taylor 1982; Singleton and Taylor 1992;
Berkes et al. 2001).
After residency, the next most frequent attribute or condition, used in
two-thirds of the organized subgroups, is that appropriators use a particular
type of technology. These rules are often criticized by policy analysts,
since gear restrictions are thought to reduce the “efficiency” of fishing.
Gear restrictions, however, have many other consequences as well. Used
in combination with choice rules that assign fishers that use one type of
gear to one area of the grounds they use, and fishers using a second type
of gear to a separate area, these rules solve conflicts among incompatible
technologies. Many gear restrictions also place a reduced load on the
fishery itself and thus help to sustain longer-term use of the resource
(Acheson 2003; Corson 2002).
  
C HA PT ER EI GH T226
Consider other boundary rules in use. A scattering of groups used ascribed
characteristics (age—two groups; ethnicity—three groups; race—
five groups). Three types of temporary use rights included government
licenses (three groups), lottery (five groups), and registration (four
groups). Seven groups required participants to have purchased an asset
such as a fishing berth, while three groups required ownership of nearby
land as a condition of appropriation. Schlager (1994) did not find that any
particular attribute or condition was correlated with higher performance
levels, but she did find that the thirty-three groups, who had at least one
boundary rule, tended to be able to solve common-pool problems more
effectively than the eleven groups who had not crafted boundary rules.
In a study of forty-three small- to medium-sized irrigation systems managed
by farmers or by government agencies, Tang (1992) found that the
variety of attributes or conditions used in irrigation was smaller than
among inshore fisheries. The single most frequently used boundary rule,
used in thirty-two of the forty-three systems (74 percent), was that an
irrigator must own land in the service area of an irrigation system (84–
85). All of the government-owned and -operated irrigation systems relied
on this attribute and only this attribute. Many of the user-organized systems
relied on other attributes and conditions or land ownership combined
with other rules. Among the other rules used were ownership of a
proportion of the flow of the resource (e.g., fish, water, forest products),
membership in a local organization, and a per-use fee.
Tang found a strong negative relationship between reliance on land as
the sole boundary requirement and performance (87). Over 90 percent of
the systems using other boundary rules or a combination of rules including
land ownership were rated positively in the level of maintenance
achieved and in the level of rule conformance, while fewer than 40 percent
of those systems relying solely on land ownership were rated at a higher
performance level (p = .001).
Many of the boundary rules used by appropriators in the field are a
mechanism to ensure that appropriators will interact with others who live
nearby and have a long-term interest in sustaining the productivity of the
resource. One way that the rules devised by appropriators increase the
sustainability of a resource is to change the composition of the group that
uses a common-pool resource. Their rules tend to increase the proportion
of participants who have a long-term interest in the resource, who are
more likely to use reciprocity, and who can be trusted. Central governments
tend to use a smaller set of rules. Some of these may open up a
resource to strangers without a long-term commitment to the resource,
create too large a geographic domain, generate conflict among users, and
lead to an unwillingness to abide by any rules.3
  
U SI NG RU LE S A S T OO LS TO CO PE 227
Position Rules Creating Monitors
In the discussion above, we focused on how boundary rules create the
general position of authorized appropriator. In some self-organized resource
governance systems, a second position of guard or monitor is also
created. Many different names are used.
Among self-organizing forest governance systems, creating and supporting
a position as guard is frequently essential when resource units
are highly valuable and a few hours of stealth generates substantial illicit
income. Monitoring rule conformance among forest users by officially
designated and paid guards may make the difference between a resource
in good condition and one that has become totally degraded. In a study
of 279 local forest councils in the Kumaon region of India, Agrawal and
Yadama (1997), for example, found that the number of months a guard
was hired was the most important variable affecting forest conditions.
The other variables that affected forest conditions in their study included
the number of meetings held by the forest council (this is usually a time
when infractions are discussed) and the number of residents in the village.
It is evident from the analysis that the capacity of a forest council to monitor
and impose sanctions on rule-breakers is paramount to maintaining the forest
in good condition. Nor should the presence of a guard be taken simply as a
formal mechanism that ensures greater protection. It is also an indication of
the informal commitment of the panchayat and the village community to protect
their forests. Hiring a guard costs money. The funds have to be generated
within the village and earmarked for protection of the resource. If there was
scant interest in protecting the forest, villagers would have little interest in setting
aside the money necessary to hire a guard. (455)
Many self-organized fisheries rely on self-monitoring more than the creation
of a formal position of guard. Most inshore fishers now use shortwave
radios as a routine part of their day-to-day operations allowing a
form of instant monitoring to occur. An official of a West Coast Indian
tribe reports, for example, that “it is not uncommon to hear messages
such as ‘Did you see so-and-so flying all that net?’ over the short-wave
frequency—a clear reference to a violation of specified gear limits” (cited
in Singleton 1998, 134). Given that most fishers will be listening to their
short-wave radios, “such publicity is tantamount to creating a flashing
neon sign over the boat of the offender. Such treatment might be preceded
or followed by a direct approach to the rule violator, advising him to
resolve the problem. In some tribes, a group of fishermen might delegate
themselves to speak to the person” (134).
Whether irrigation systems create a formal position as guard depends
both on the type of governance of the system and on its size. Of the fifteen
  
C HA PT ER EI GH T228
government-owned irrigation systems included in Tang (1992), twelve
(80 percent) have established a position of guard. Stealing water was a
problem on most government-owned systems, but it was endemic on the
three government systems without guards. Of the twenty-eight farmerorganized
systems, seventeen (61 percent) utilize the position of water
distributor or guard. Eleven farmer-organized systems do not employ a
guard. Farmers are vigilant enough in monitoring each other’s activities
on five systems (45 percent) that rule conformance is high. This means,
of course, that self-monitoring is not high enough on the other six systems
to support routine conformance with their own rules.
An earlier study by Romana de los Reyes (1980) of fifty-one communal
irrigation systems in the Philippines illustrates the effect of size. Of the
thirty systems that were less than fifty hectares, only six (20 percent) had
established a position as guard; of the eleven systems that served between
fifty to one hundred hectares, five (45 percent) had established guards;
and of the ten systems over one hundred hectares, seven (70 percent) had
created guards. In a survey of over six hundred farmers served by these
communal irrigation systems, she also found that most farmers also patrolled
their own canals even when they were patrolled by guards accountable
to the farmers for distributing water. Further, the proportion of farmers
who reported patrolling the canals serving their farms increases to 80
percent on the largest self-organized systems compared to 60 percent on
the smallest systems.
Considerable variation thus exists in the kind of monitoring positions
created in self-organized systems. Regardless of the rules creating such
positions, we have consistently found that higher levels of local monitoring
are positively related to resource conditions (see Gibson, Williams,
and Ostrom 2005).
Affecting the Set of Allowable Actions through Choice Rules
Choice rules are also a major type of rule used to regulate common-pool
resources. Some irrigation systems allocate water simply on the basis of
the crops that a farmer grows.4
Some rules involve a simple formula as a
way of devising how many resource units appropriators may obtain. Others
simply choose the resource for a defined period(s) and then allow harvesting
during a particular season. Many forest resources, for example,
are closed to all forms of harvesting during one portion of the year and
open for extraction by all who meet the boundary rules during an open
season. Most choice rules, however, have two components: an allocation
formula and the assets on which the formula is based.
In table 8.2, the eight allocation formulas used in the field are shown
in the left column. A fisher might be assigned to a fixed location (a fishing
  
U SI NG RU LE S A S T OO LS TO CO PE 229
TABLE 8.2
Choice rules used to allocate common-pool resources
Allocation formula for appropriation rights Basis for allocation formula
Percentage of total available units per period Amount of land held
Quantity of resource units per period Amount of historical use
Appropriate only from a specific location Location of appropriator
Appropriate only from a specific time slot Quantity of shares of resource owned
Rotate in time or space Proportion of resource flow owned
Appropriate only during open seasons Purchase of periodic rights at auction
Appropriate only resource units meeting Rights acquired through periodic lottery
criteria
Appropriate whenever and wherever Technology used
License issued by a governmental authority
Equal division to all appropriators
Needs of appropriators (e.g., type of crop)
Ascribed characteristic of appropriator
Membership in organization
Assessment of resource condition
spot) or to a fixed rotational schedule, a member of the founding clan
may be authorized to cut timber anywhere in a forest, while an irrigator
might be assigned to a fixed percentage of the total water available during
a season or to a fixed time slot. In addition to the formula used in a
choice rule, most also attach a condition as a basis for the assignment.
For example, a fisher might be assigned to a fixed location based on a
number drawn in a lottery, on the purchase of that spot in an auction, or
on the basis of his or her historical use.5
An irrigator might be assigned
to a fixed rotation based on the amount of land owned, the amount of
water used historically, or the specific location of the irrigator.
If all of the conditions were equally likely to be combined with all of
the formula, there would be 112 different choice rules (8 allocation formulas
× 14 bases). A further complication is that the rules for one product
may differ from those of another product harvested from the same resource.
In regard to forest resources, for example, children may be authorized
to pick fruit from any tree located in a forest so long as it is for their
own consumption, women may be authorized to collect so many headloads
of dead wood for domestic firewood and certain plants for making
crafts, while shaman are the only ones authorized to collect medicinal
plants from a particular location in a forest (Fortmann and Bruce 1988).
Appropriation rights to fish are frequently related to a specific species. A
still further complication is that the rules may regularly change over the
course of a year depending on resource conditions. Thus, the exact number
of rules that are actually used in the field is difficult to compute.
  
C HA PT ER EI GH T230
Schlager (1994, 259–60) found that user groups included in her study
frequently assigned fishers to fixed locations using a diversity of bases
including technology, lottery, or historical use. Seven groups allocated
fishers to fishing spots using a rotation system, and seven other groups
allowed fishing locations to be used only during a specific season. Four
groups allocated fishing spots for a particular time period (a fishing day
or a fishing season). On the other hand, nine user groups required fishers
to limit their harvest to fish that met a specific size requirement.
An important finding—given the puzzles addressed in this chapter—is
that the choice rule most frequently recommended by policy analysts (see
L. Anderson 1986; Arnason and Gissurarson 1999; Copes 1986) was not
used in any of the coastal fisheries included in Schlager’s study. In none
of the fisheries coded by Schlager was an attempt made to regulate the
quantity of fish harvested per year based on an estimate of the yield. “This
is particularly surprising given that the most frequently recommended
policy prescription made by fishery economists is the use of individual
transferable quotas based on estimates on the economically optimal quantity
of fish to be harvested over the long run” (397).
In an independent study of thirty traditional fishery societies, Acheson,
Wilson, and Steneck also noted the surprising absence of quota rules: “All
of the rules and practices we found in these 30 societies regulate ‘how’
fishing is done. That is, they limit the times fish may be caught, the locations
where fishing is allowed, the technology permitted, and the stage of
the life cycle during which fish may be taken. None of these societies limits
the ‘amount’ of various species that can be caught. Quotas—the single
most important concept and tools of scientific management—is conspicuous
by its absence” (1998, 397; see Wilson et al. 1994). Many local inshore
fishers, when allowed to self-organize, appear to use rules that differ
substantially from those recommended by advocates of scientific management
(Berkes et al. 2001, 177–79). Fishers have to know a great deal
about the ecology of their inshore region including spawning areas,
nursery areas, the migration routes of different species, and seasonable
patterns just in order to succeed as fishers. Those inshore fisheries that
have survived the threat of rapid technological change have learned
how “to maintain these critical life-cycle processes with rules controlling
technology, fishing locations, and fishing times. Such rules in their view
are based on biological reality” (Acheson, Wilson, and Steneck 1998,
405). Lobe and Berkes (2004) also illustrate how a combination of
these three types of rules sustains contemporary coastal shrimp fisheries
in Kerala, India.
In the irrigation systems studied by Tang (1992, 90–91), three types of
choice rules are used most frequently: (1) a fixed time slot is assigned to
each irrigator (nineteen out of the thirty-seven cases for which data are
  
U SI NG RU LE S A S T OO LS TO CO PE 231
available, and in ten out of twelve government-owned systems), (2) a fixed
order for a rotation system among irrigators (thirteen cases), and (3) a
fixed percentage of the total water available during a period of time (five
cases). A variety of conditions were used in these rules such as “amount
of land held, amount of water needed to cultivate existing crops, number
of shares held, location of field, or official discretion” (Tang 1994, 233).
Three poorly performing systems with high levels of conflict had not
crafted any choice rule at all. Farmers also do not use rules that assign a
specific quantity of water to irrigators other than in the rare circumstances
where they control substantial amounts of water in storage (see Maass
and Anderson 1986).
Fixed time slot rules allow farmers considerable certainty as to when
they will receive water without an equivalent certainty about the quantity
of water that will be available in the canal. Fixed time allocation systems
are criticized as inefficient since water is not allocated to the farmers with
the highest productivity. This condition does, however, economize greatly
on the amount of knowledge farmers have to have about the entire system
and on monitoring costs. Spooner (1974) and Netting (1974) described
long-lived irrigation systems in Iran and in Switzerland where there was
full agreement on the order and time allotted to all farmers located on a
segment of the system, but no one knew the entire sequence for the system
as a whole.
In a study of 248 irrigation systems in Nepal, colleagues associated
with the Irrigation Management Systems Study Group (IMSSG) at the
Institute of Agriculture and Animal Sciences in Rampur, Nepal, found a
substantial variety of choice rules used depending largely on the elevation
of the district in which the system was located and the time of year involved.
Thus, not only is a variety of rules used—but the rules chosen
by farmers tend to depend on ecological conditions. As shown in table
8.3, the systems included in the study were located in two districts of
Nepal: (1) Chitwan—a district in the flatlands of Nepal that has larger
river systems with substantially higher volume of water during the
monsoon seasons than (2) Tanahun—a district in the middle hills with
few large rivers.6
Since water is more abundant during the monsoon season
in Chitwan, a larger proportion (43 percent) of the systems in this
district relies on a “free flow rule” that authorizes a continuous supply in
channel than in Tanahun (33 percent). In Tanahun during the water
surplus, monsoon season, the most frequently used rule is “rotational”
(52 percent), but this rule is hardly used in Chitwan during the monsoon
season (6 percent).
Rules are different in both districts during the spring—or the water
deficit time—of the year. The proportion of systems that allow farmers
to appropriate whenever they desire in the dry months is less than 10
  
C HA PT ER EI GH T232
TABLE 8.3
Water allocation rules in Chitwan and Tanahun districts
Water surplus time of year Water deficit time of year
Chitwan Tanahun Chitwan Tanahun
Free flow/continuous supply 38 (43%) 52 (33%) 1 (1%) 14 (9%)
Demand/first in time 43 (49%) 4 (3%) 0 (0%) 4 (3%)
Rotational 5 (6%) 83 (52%) 79 (90%) 103 (64%)
No specific methods (consensus) 2 (2%) 35 (22%) 8 (9%) 39 (24%)
Number of irrigation systems 88 160 88 160
Sources: Adapted from Shukla et al. 1993; Poudel et al. 1994.
percent in both districts (see table 8.3). In the water deficit time of the
year, nine out of ten systems in Chitwan and over six out of ten systems
in Tanahun use some form of a rotation system. This movement back
and forth between rules depending on the season is very typical for selforganized
resource systems around the world.7
Tang (1992) also found that many irrigation systems use different sets
of rules depending on the availability of water. During the most abundant
season, for example, irrigators may be authorized to take water whenever
they need it. During a season when water is moderately available, farmers
may use a rotation system where every farmer is authorized to take water
for a fixed amount of time during the week based on the amount of
land to be irrigated. During scarcity, the irrigation system may employ a
special water distributor who is authorized to allocate water to those
farmers who are growing crops authorized by the irrigation system and
are most in need.
In addition to devising choice rules specifying how resource units may
be harvested, many systems also have to devise rules for how resources
will be mobilized. These types of choice rules specify duties as contrasted
to rights. As discussed in chapter 9, robust common-property regimes
tend to rely on a close match between the formulas used for harvesting
and the formulas used for input requirements. In regard to irrigation,
farmers may even craft different rules related to maintenance according
to the part of the canal needing attention—such as the headworks, the
main canal, secondary canals, or areas that need emergency repair. In
Chitwan, most systems tend to rely on mobilizing labor for repairing the
headworks on an irrigation system on a per household basis (also for
emergency repair anywhere on the system) but use the amount of land
owned and served by a particular part of a canal in cleaning of weeds and
other impediments to the flow of water in the main or secondary canals
(see summary of these findings in Shivakoti and Ostrom 2002, 14–15).
In Tanahun—where the systems tend to be much smaller than in Chitwan,
  
U SI NG RU LE S A S T OO LS TO CO PE 233
and thus smaller differences between farmers located at the head and the
tail—different rules tend to be used. About half of the 160 systems for
which the IMSSG group of researchers gathered data relied on landholding
anywhere in the system and a per household basis for mobilizing regular
repairs on all parts of the system.
The diversity of choice rules devised by users greatly exceeds the few
rules recommended in textbook treatments of this problem. Appropriators
thus cope with the commons by crafting a wide variety of rules affecting
the actions available to participants and thus their basic set of strategies.
Given this wide diversity of rules, it is particularly noteworthy that
rules assigning appropriators a right to a specific quantity of a resource
are used so infrequently in inshore fisheries and irrigation systems. They
are used more frequently when allocating forest products where the quantity
available, as well as the quantity harvested, are much easier to measure
(Agrawal 1994). To assign an appropriator a specific quantity of a
resource unit requires that those making the assignment know the total
available units. In water resources, only when water is stored from one
season to another in a groundwater basin or dam, and reliable information
about the quantity of water is available, are rules that allocate a
quantity of water to an authorized appropriator utilized (Blomquist 1992;
Schlager, Blomquist, and Tang 1994).
Affecting Outcomes through Payoff and Position Rules
One way to reduce or redirect the appropriations made from a commonpool
resource is to change payoff rules so as to add a penalty to actions
that are prohibited. Many user groups also adopt norms that those who
are rule breakers should be socially ostracized or shunned, and individual
appropriators tend to monitor each other’s behavior rather intensively.
Three broad types of payoff rules are used extensively in the field: (1) the
imposition of a fine, (2) the loss of appropriation rights, and (3) incarceration.
The severity of each of these types of sanctions can range from very
low to very high and tends to start out on the low end of the scale.
Inshore fisheries studied by Schlager relied heavily on shunning and
other social norms and less on formal sanctions. Thirty-six of the fortythree
irrigation systems studied by Tang used one of these three rules and
also relied on vigorous monitoring of one another’s behavior and shunning
of rule breakers. The seven systems that did not self-consciously punish
rule infractions were all rated as having poor performance. Fines were
most typically used (in twenty-one cases) and incarceration the least (in
only two cases). Fines tend to be graduated depending on the seriousness
of the infractions and the number of prior infractions. The fines used for
a first or second offense tend to be very low.
  
C HA PT ER EI GH T234
Once a position of a paid guard is created, payoff rules must also
change so as to mobilize resources to remunerate a guard. Several formulas
are used. On government-owned irrigation systems, guards are normally
paid a monthly wage that is not dependent on the performance of
a system or farmers’ satisfaction. Agrawal (2005) describes four different
payment methods for forest guards that users select: by each household
in kind (grains, services), by each household in cash, by the local forest
user organizations out of general funds, and by the local forest organization
out of central government distributed funds. In South India, Wade
(1994) describes self-organized systems where the water distributor-guard
is paid in kind based on a proportion of the yield. As the harvest is reaped,
the guard must go to each farmer to collect his share based on the harvest
level and the amount of land owned by the farmer. Sengupta (1991, 104)
describes another system where immediately after appointment, the
guards “are taken to the temple for oath taking to remain impartial. With
this vow, they break a coconut. They are paid in cash at the rate of Rs
10 per acres . . . per month by the cultivators. The neerpaichys (guards)
themselves collect the money.” Having the farmers pay the guards directly
in kind or in cash enables the farmers to “monitor” the monitor more
effectively and ensure that there is a rough proportionality between benefits
received and the costs of the system.
Boundary and choice rules also affect how easy or difficult it is to monitor
activities and impose sanctions on rule infractions. Closing a forest or
an inshore fishery for a substantial amount of time, for example, has multiple
impacts. It protects particular plants or fish during critical growing
periods and allows the entire system time to regenerate without disturbance.
Further, during the closed season, rule infractions are highly obvious.
Any person in the resource during the closed season is almost certainly
breaking the rules. Similarly, requiring appropriators to use a
particular technology may reduce the pressure on the resource, help to
solve conflicts among users of incompatible technologies, and also make
it very easy to ascertain if rules are being followed. Many irrigation systems
set up rotation systems so that only two persons need to monitor
actions at any one time. The farmers whose “turn” it is watch to be sure
the next farmer does not start a turn early, and the next farmer watches
to be sure the turn-taker stops at the specified time. This keeps monitoring
costs low.
In general, self-organized governance systems need to match the rules
that impose costs in a rough proportion to the likely positive payoffs that
appropriators are likely to obtain over time. If the appropriators do not
view the requirements placed on them to provide funds, time, or materials
to be equitable, they are much less willing to conform. Since self-governed
  
U SI NG RU LE S A S T OO LS TO CO PE 235
systems must rely more on willing consent as contrasted to coerced contributions,
they need to pay more attention to the relationship between positive
and negative payoffs than systems that can easily mobilize police to
extract contributions from participants.
Affecting Outcomes through Changes in Information, Scope,
and Aggregation Rules
Information, scope, and aggregation rules tend to be used in ways that
complement boundary, position, choice, and payoff rules. Individual systems
vary radically in regard to the mandatory information that they require.
Many smaller and informal systems rely entirely on a voluntary
exchange of information and on mutual monitoring. Where resource
units are very valuable and the size of the group is larger, more and more
requirements are added regarding the information that must be kept by
appropriators or their officials. Blomquist (1992, 1994) examined the information
rules for eight groundwater basins in southern California. He
found that the groundwater basins that had overcome severe overdraft
problems had information rules that differed from those used in the one
basin that had not overcome these problems. For example, in the “successful”
basins, all groundwater producers were required to report the
amount of water they produced each year, and various methods were
implemented to verify the accuracy of these records. Producers were entitled
to receive annual reports providing data about the activities of their
monitors, basin conditions, and the production records of others. The
basin that still faced overdraft conditions did not have the same information
rules (Blomquist 1994, 292).
Scope rules are used to limit harvesting activities in some resources by
creating refugia, where actions that are permitted in other areas of a resource
are forbidden in the refugia. By not allowing any appropriation
from these locations, the regenerative capacity of a system can be enhanced.
Sacred groves are probably the most well-known form of refugia.
The ancient sacred groves of India, Africa, and the Mediterranean are
well documented in history. Modern sacred groves are known to exist in
modern times in China, Ghana, India, Mexico, Nepal, Thailand, Uganda,
and Zimbabwe (Gadgil, Hemam, and Reddy 1998, 37; Gombya-Ssembajjwe
1995). Many locally controlled fisheries also protect breeding
grounds as off-limits for harvesting to enable the eggs and small fish to
mature successfully before being subject to harvest (see Folke, Berkes, and
Colding 1998).
Aggregation rules are used extensively in collective-choice processes
and less extensively in appropriation situations. One aggregation rule that
  
C HA PT ER EI GH T236
is found in diverse resource systems is a requirement that harvesting activities
be done in teams. This increases the opportunity for mutual monitoring
and reduces the need to hire special guards. We will return to a discussion
of aggregation rules later in this chapter when we analyze the factors
affecting appropriators’ choice of rules. The aggregation rules used in
making collective choices are crucial rules affecting whose interests are
taken into account when decisions are made to change the operational
rules of an appropriation situation.
It is important to note that we have not yet found any particular rules
to have a statistically positive relationship to performance across resource
types, ecological zones, and communities. On the other hand, the
absence of any boundary or choice rule is consistently associated with
poor performance. Relying on the use of a single type of rule for an entire
set of common-pool resources in a large region is also negatively related
to performance.
Contemporary Approaches to Resource Policy
Instead of studying the literature describing the successful and unsuccessful
efforts of local users or public officials to devise rules for coping with
common-pool resource problems, which is briefly presented above, many
students of public policy read only textbooks that elucidate an approach
that can broadly be called “the scientific management of natural resources.”
This approach teaches future policy analysts to consider fisheries,
forests, pasture lands, and water resources as relatively homogeneous
systems that are closely interrelated across a vast domain (see Sherman
and Laughlin 1992). Irrigation systems are interlinked along watersheds
of major river systems. Fish and wildlife species tend to migrate over a
large range. By implication, uniform systems of rules are usually prescribed
as the best solution. Acheson, Wilson, and Steneck (1998, 391–
92) describe this approach as applied to fisheries management:
For those trained in scientific management, it is also an anathema to manage a
species over only part of its range. From the view of fisheries scientists and
administrators, it is not rational to protect a species in one zone only to have
it migrate into another area where it can be taken by other people due to a
difference in regulations. As a result, the units to be managed range along hundreds
of miles of coast and can only be managed by central governments with
jurisdiction over the entire area. . . . From the point of view of the National
Marine Fisheries Service, it makes sense to have a set of uniform regulations
for the entire US coast rather than one for each state. (italics added)
  
U SI NG RU LE S A S T OO LS TO CO PE 237
Many policy analysts share a belief in the feasibility of central authorities
to design close-to-optimal rules for governing and managing commonpool
resources existing in a large domain.8
Since common-pool resources
are viewed as relatively homogeneous and interlinked and since simple
models have been developed of how they are thought to work (Gordon
1954; Heal 1998), officials acting in the public interest are considered capable
of devising uniform and effective rules for an entire region.
Recommendations calling for central governments to impose uniform
regulations over natural resources within a country’s boundary are frequent
and strident. After reviewing the problems of deforestation in many
countries, Grainger (1993, 224) urges the different departments whose
policies affect forests to come together and “agree on an integrated land
use policy for the whole country.” As his first priority in controlling deforestation,
Grainger recommends directing “more funds into forestry departments
so they have sufficient personnel and technology to monitor
forests and logging operations, prevent illegal deforestation by improving
protection, and ensure that logging operations take place in accordance
with government regulations” (225). As Blomquist and Ingram (2003)
point out, some analysts even call for central administrative control over
transboundary resources. Fortunately, there are now some strong voices
challenging the presumed superiority of central authorities to solve
smaller-scale appropriation problems (see Karkkainen 2001/2; Holling,
Gunderson, and Ludwig 2001).
As an alternative to central control, other policy analysts call for the
imposition of a market system related to resources at various scales. Carson,
Marinova, and Zilberman (1999, 1), in addressing transboundary
water problems in the Middle East, conclude that “the current water allocation
structure has proven inadequate. It should be replaced with some
form of a water market.” Further, they declare that considerations of different
ways to create markets are irrelevant and that the first step for a
dramatic change is: “Replace the current water institutions and allocation
structure with a water market” (1).
Challenging Three Basic Assumptions of Contemporary
Policy Analysis
Based on evidence, it is important to challenge three basic assumptions
underlying the study of contemporary resource policies: (1) the view that
resource appropriators are helpless to overcome their temptations to harvest
excessively from a resource; (2) the assumption that designing rules
to change the incentives of participants is a relatively simple analytical
task; and (3) the view that organization itself requires central direction.
  
C HA PT ER EI GH T238
The first foundation that can be disputed is the model of the human
actor. Resource users are explicitly thought of as rational egoists who
plunder local resources so as to maximize their own short-term benefits.
Government officials are implicitly depicted, on the other hand, as seeking
the more general public interest, having the relevant information at hand,
and the capability of designing optimal policies. As discussed in chapter
4, the rational egoist used in conventional noncooperative game theory is
an appropriate model to use for all participants in open competitive markets
and other settings where participants are relatively anonymous and
have little opportunities to develop norms and longer time horizons. Outside
of these settings, however, one needs to assume a mix of participants
ranging from those with strong norms of reciprocity to those with weak
or few shared intrinsic values. Assuming a multiplicity of orientations is
more appropriate in these settings in which individuals can communicate
and come to know and potentially trust one another. Rational egoists may
come to dominate in any situation in which conflicts are left unresolved
and participants lose trust in one another.
One should not, however, presume that all government officials are
“saints” while assuming that all resource users are “sinners.” Nor should
we presume that officials have all the relevant knowledge to manage complex
dynamic systems while local appropriators are ignorant. The knowledge
base of government officials may not, in reality, be better than that
of local appropriators who have used a particular resource for years and
know its characteristics in considerable detail. Even when the knowledge
base is similar, no guarantee exists that government officials (or the researchers
who advise them) will use available information to make efficient
and/or sustainable decisions.
For example, Moxnes (1998) conducted a series of experiments examining
the capacity of eighty-two subjects who were Norwegian fishers,
officials working for government resource agencies, or researchers familiar
with resource problems to make economically efficient and resourcesustaining
decisions related to a simulated dynamic model of a fishery. All
subjects were assigned the equivalent of private property rights over a
fishery and asked to make decisions about the purchase of vessels and
harvesting rates over a twenty “vertical” year period. On average, all subjects
substantially overinvested in the fishery. Many subjects reported
using heuristics of the type: “Things seem to be going well; I’ll order another
vessel” (1241) similar to the heuristic used in the commons dilemma
experiments with no communication described in chapter 3. Government
officials and researchers did no better than the fishers in solving this
problem.
A second foundational belief of contemporary resource policy is that
designing rules to change the incentives of participants is a relatively sim-
  
U SI NG RU LE S A S T OO LS TO CO PE 239
ple analytical task best done by objective analysts not specifically related
to any specific resource. As Acheson, Wilson, and Steneck (1998) describe
above, analysts frequently view most resources in a particular sector as
relatively similar and sufficiently interrelated that they need to be governed
by the same set of rules.
It should now be obvious that the search for rules that improve the
outcomes obtained in commons dilemmas is an incredibly complex task
involving a potentially infinite combination of specific rules that could be
adopted. To ascertain whether one has found an optimal set of rules to
improve the outcomes achieved in a single situation, one would need to
analyze how diverse rules affect the components of such a situation and
as a result, the likely effect of a reformed structure on incentives, strategies,
and outcomes. Since multiple rules directly or indirectly affect each
of the seven components of action situations, conducting such an analysis
would be an incredibly time- and resource-consuming process.
If only five changes in rules per component were considered, for example,
there would be 57
or 75,525 different situations to analyze. This is a
gross simplification, however, since some of the important rules used in
field settings include more than twenty-five rules (in the case of boundary
rules) and even over one hundred variants (in the case of choice rules).
Further, how these changes affect the outcomes achieved in a particular
location depends on the biophysical characteristics of that location and
the type of community relationships that already exist. No set of policy
analysts (or even all of the game theorists in the world today) would ever
have sufficient time or resources to analyze over 75,000 combinations of
rule changes and resulting situations, let alone all of the variance in these
situations due to biophysical community differences.
Those directly involved would also not be able to do a complete analysis.
They would know a lot about local biophysical processes, but not
necessarily how that resource system might be linked with biophysical
processes occurring at a somewhat larger scale or even how complex biophysical
systems were operating. They would also know a great deal
about local community norms and the distribution of resources and interests
within a community. Given the nonlinearity and complexity of many
action situations, it is challenging to predict the precise effect of a change
in a particular rule.
For example, a change in a boundary rule to restrict who is authorized
to enter and harvest from a resource reduces the number of individuals
who are entitled to appropriate from a resource. It also reduces the number
of individuals who are interested in monitoring what is happening or
contributing funds toward hiring a guard (Agrawal 2000; Agrawal and
Goyal 2001). Thus, the opportunities for rule breaking may increase. Further,
the cost of a rule infraction will be spread over a smaller group of
  
C HA PT ER EI GH T240
appropriators. Thus, the harm to any individual may be greater. Assessing
the overall effects of a change in boundary rules is a nontrivial analytical
task (for examples, see Weissing and Ostrom 1991a, 1991b). Instead of
conducting such a complete analysis, appropriators are more apt to use
their intuitive understanding of the resource and of one another’s norms
and preferences to experiment with different rule changes and assess the
effects of rules with which experiment until they find a combination that
seems to work in their setting.
Local appropriators would receive feedback from their own experiments
over time and could then improve how their rules worked over
time. If they had good communication with other communities who had
experience with multiple rules related to similar resources, they would be
able to make informed judgments about the likely impact of some rules,
but certainly not all of the rules that they might contemplate adopting or
changing. Greater attention to the ways that local experiments are monitored
and the mechanisms available for sharing information could improve
the likelihood of improving performance over time.
The third foundational belief of much contemporary policy analysis is
the view that organization itself requires central direction. Consequently,
the multitudes of self-organized resource governance systems are viewed
as mere collections of individual agents each out to maximize her own
short-term returns. The groups who have actually organized themselves
are invisible to those who cannot imagine organization without rules and
regulations issued by a central authority (see, for example, Lansing 1991).
A more appropriate foundation is to assume that governments at multiple
levels could, but do not always, adopt policies that enhance effective
problem-solving and resource sustainability. Instead of central direction,
what is needed are policies that enhance the accuracy and reliability of
information, that provide low-cost conflict resolution, and that develop
the authority to govern resources at multiple levels.
Useful (but Partial) Analyses of Rule Configurations
I do not wish to argue that analysts as well as participants are unable
or should not examine interesting combinations of rules under specified
conditions. Analyzing specific situations and how diverse rules would affect
the likely incentives, behavior, and outcomes of participants is an
extremely important and useful endeavor that social scientists should and
do perform.
Scholars at the California Institute of Technology, for example, took
on an important assignment for the Civil Aeronautics Board (CAB) to
examine how several rules they were considering for the allocation of
airport slots would work (Grether, Isaac, and Plott 1979, 1981). Not only
  
U SI NG RU LE S A S T OO LS TO CO PE 241
did they develop a formal analysis of how alternative rules would affect
the incentives of airline carriers and the likely resulting behavior, they also
implemented a version of the decision setting in an experimental laboratory.
They were interested in predicting the level of efficiency likely from
alternative decision rules at multiple levels, but also the responsiveness
that was likely due to changed economic conditions and the susceptibility
of alternative rules to collusion. In light of their modeling and their experimentation,
they recommended the establishment of a “one-price sealed
bid auction” as the best method they could recommend for allocating
landing slots. Their recommendations came after they developed a model
of this specific situation and after they had undertaken a trial run process
in an experimental lab to see if their theory did predict behavior in this
particular situation.9
For a paper that I presented at the Annual Bank Conference on Development
Economics (called the ABCDE conference!), I also undertook a
comparative analysis of a set of rules as they would affect the structure
of an action situation—in this case, a formal game—and the net benefits
that would be likely at the equilibria of such a game. In that paper
(E. Ostrom 1996), I assumed an environment of ten farmers who all
owned about the same amount of land on a relatively flat but rich alluvial
plain. The farmers were hypothesized to be interested in constructing and
then managing their own small irrigation system. I explored a series of
games that would be created by using several combinations of rules for
allocating benefits and costs. Given my assumptions about the farmers,
the environment, and the costs of maintenance, I could show that the
farmers would gain the most out of two different combinations of rules.
The ABCDE paper was useful for several reasons. First, it was fun. It
is a delight to work with formal worlds and be able to demonstrate clearly
how rules would affect the structure of a formal game. Second, I was able
to demonstrate overtly the link between changes in the set of rules and
how these affect the structure of the action situation itself (see also
E. Ostrom, Gardner, and Walker 1994, chap. 4, where we undertake similar
efforts). Since I think there is a strong connection between rule configurations
and action situations, it is always helpful to demonstrate the
linkage between rules and games clearly. Third, the paper demonstrates
that in order to undertake such a formal analysis of rules and game structures,
I had to make multiple assumptions about the participants and the
environment they were in. These included:
Assumptions about the participants:
1. the number of farmers (ten);
2. the relative equality of their holdings;
3. the value of the water for farming to all of the farmers;
  
C HA PT ER EI GH T242
4. the engineering knowledge and skills of the farmers;
5. the rights of the participants to use the source of water without contest
by other farmers;
6. the rights of the farmers to organize for joint benefit and create rules
which they themselves would enforce; and
7. the rational egoistic behavior of farmers.
Assumptions about the environments:
1. the existence of a flat plain;
2. the existence of a nearby water source that was not currently used by
other farmers;
3. relatively similar soil conditions across the plain;
4. the layout of the parcels; and
5. the feasibility of two different canal layouts.
Further, I considered the impact of only a total of seven rules—an extremely
small set, when one considers the multiplicity of rules arrayed
above. Any major change in the variables that I assumed could have led
to an entirely different outcome. Thus, the exercise in no way demonstrated
an optimal rule configuration at a more general level. Any effort
to examine the impact of alternative rules on some specific kind of action
situation will always have to make a large number of assumptions as I
did in this effort. We should not fool ourselves into thinking that the
results of these useful exercises are a full analysis showing one rule configuration
producing more net benefits than others.
Coping with Complexity: A General Problem
The complexity we have found in regard to common-pool resources in
the field is not in any way unique to natural resources. For far too long,
social scientists have viewed the physics of static, simple systems as the
model of science we should try to emulate. Those who want to emulate
the science of static, simple systems are grossly out-of-date when it comes
to understanding contemporary science and particularly contemporary
engineering. The engineers responsible for the design of airplanes and
bridges—and now computers—have long coped with complex dynamic
systems. The Boeing 777, for example, has 150,000 distinct subsystems
that are composed, in some instances, of highly complex components.
Design engineers of complex systems long ago gave up hope of ever
doing complete analyses of all combinations of subsystems under all
combinations of external environmental conditions. Obviously, they invest
heavily in trying out diverse design elements under a variety of conditions.
Testing designs by building models, using wind tunnels and com-
  
U SI NG RU LE S A S T OO LS TO CO PE 243
puter simulations, increases the likelihood that engineers can produce a
viable combination of design elements that are robust under many conditions.
They also invest in complex backup systems that enable these designed
systems to achieve a high degree of robustness—meaning the capacity
to maintain some desired system characteristics under changing
circumstances. All such robust systems are, however, fragile to a variety
of small perturbations (Carlson and Doyle 2002). Small, rare disturbances
can cause a disastrous cascade of failure in any highly complex
designed system.
Instead of assuming that designing rules that approach optimality, or
even improve performance, is a relatively simple analytical task that can
be undertaken by distant, objective analysts, we need to understand the
policy design process as involving an effort to tinker with a large number
of component parts (see Jacob 1977). Those who tinker with any tools—
including rules—are trying to find combinations that work together more
effectively than other combinations. Policy changes are experiments based
on more or less informed expectations about potential outcomes and the
distribution of these outcomes for participants across time and space
(Campbell 1969, 1975). Whenever individuals decide to add a rule,
change a rule, or adopt someone else’s proposed rule set, they are conducting
a policy experiment. Further, the complexity of the ever-changing biophysical
and socioeconomic world combined with the complexity of rule
systems means that any proposed rule change faces a nontrivial probability
of error.
Changing Rules as an Adaptive Process
Given the logic of combinatorics, it is not possible for anyone to conduct
a complete analysis of the expected performance of all the potential rule
changes that could be made in an effort to improve the outcomes
achieved. When we study rules used by appropriators from common-pool
resources in the field, we can think of appropriators trying to understand
the biophysical structure of their resource and how they can develop a set
of rules consistent with the time and place exigencies and the norms
shared in their community. Instead of being given a set of instructions
with a fully specified transformation function (as is the case for subjects
in experimental settings discussed in chapter 3), appropriators in the field
have to explore and discover the biophysical structure of a particular resource.
It will usually differ on key parameters from similar resources in
the same region. Further, they have to cope with considerable uncertainty
related to the weather, complicated growth patterns of biological systems
that may be characterized by multiple equilibria, and external price fluc-
  
C HA PT ER EI GH T244
tuations affecting the costs of inputs and value of outcomes (see Baker
2005; Wilson et al. 1994; Wilson 2002). One of their first challenges will
be to convince those who doubt that the resource is limited or that they
will benefit from cooperation, and thus that they need to constrain use in
a manner that they agree is workable and fair (Gibson 2001).
Officials and/or the appropriators themselves may try to improve performance
by changing one or more rules in an adaptive process. Participants
adapt the rules, norms, and strategies of their parents and elders as
well as those who are viewed as highly successful in a particular culture.
They learn about neighboring systems that work better than theirs and
try to discern which rules are helping their neighbors to do better. Human
agents try to use reason and persuasion in their efforts to devise better
rules, but the process of choice from the vast array of rules they might
use always involves experimentation. Self-organized resource governance
systems use many types of decision rules to make collective choices ranging
from deferring to the judgment of one person or elders to using majority
voting to relying on unanimity.
Scholars familiar with the results of field research do substantially
agree on a set of variables that enhance the likelihood of appropriators
organizing themselves to try to avoid the social losses associated with
open access or rules that are not yet working well.10
Considerable consensus
exists that the following attributes of resources and of appropriators
are conducive to an increased likelihood that self-governing associations
will form.
Attributes of the Resource
R1. Feasible improvement: Resource conditions are not at a point of deterioration
such that it is useless to organize or so underutilized that little advantage
results from organizing.
R2. Indicators: Reliable and valid indicators of the condition of the resource
system are frequently available at a relatively low cost.
R3. Predictability: The flow of resource units is relatively predictable.
R4. Spatial extent: The resource system is sufficiently small, given the
transportation and communication technology in use, that appropriators can
develop accurate knowledge of external boundaries and internal micro-
environments.
Attributes of the Appropriators
A1. Salience: Appropriators depend on the resource system for a major
portion of their livelihood or the achievement of important social or religious
values.
A2. Common understanding: Appropriators have a shared image of how the
resource system operates (attributes R1, 2, 3, and 4 above) and how their actions
affect each other and the resource system.
  
U SI NG RU LE S A S T OO LS TO CO PE 245
A3. Low discount rate: Appropriators use a sufficiently low discount rate in
relation to future benefits to be achieved from the resource.
A4. Trust and reciprocity: Appropriators trust one another to keep promises
and relate to one another with reciprocity.
A5. Autonomy: Appropriators are able to determine access and harvesting
rules without external authorities countermanding them.
A6. Prior organizational experience and local leadership: Appropriators
have learned at least minimal skills of organization and leadership through participation
in other local associations or learning about ways that neighboring
groups have organized.
Many of these variables are affected by the larger regime in which a
resource and its appropriators are embedded. Larger regimes can facilitate
local self-organization by providing accurate information about natural
resource systems, providing arenas in which participants can engage in
discovery and conflict-resolution processes, allowing for autonomy, and
providing mechanisms to back up local monitoring and sanctioning efforts.
The probability of participants adapting more effective rules in
macroregimes that facilitate their efforts over time is higher than in regimes
that ignore resource problems entirely or, at the other extreme,
presume that all decisions about governance and management must be
made by central authorities.
A Rule Change Calculus
Now, why are these attributes of a resource and of appropriator likely to
be associated with solving collective-action problems? These attributes
combine to affect the perceived net benefits of a set of appropriators (A)
using a resource under a particular set of rules. The benefits and costs
involved in this calculus involve both extrinsic and intrinsic valuations as
discussed in chapter 4. Each appropriator i (i ∈ A) has to compare his or
her perception of the expected net benefits of harvesting while continuing
to use an old set of rules (Rold) to the benefits he or she expects to achieve
with a new set of rules (Rnew). Each appropriator i must ask whether his
or her incentive to change (Γi) is positive or negative.
Γi = Rnew − Rold.
If Γi is negative for all appropriators, no one has an incentive to change.
If Γi is positive for some appropriators, they then need to estimate three
types of costs:
C1—the up-front costs of time and effort spent devising and agreeing upon
new rules;
  
C HA PT ER EI GH T246
C2—the short-term costs of adopting new appropriation strategies; and
C3—the long-term costs of monitoring and maintaining a self-governed system
over time.
If the sum of these expected costs for each appropriator exceeds the
incentive to change, no appropriator will invest the time and resources
needed to create new institutions. Thus, if
Γi < (C1i + C2i + C3i)
for all i ∈ A, no change occurs.
In field settings, appropriators are not likely to expect the same costs
and benefits from a proposed change. Some may expect positive net benefits.
Others may expect net losses from the same proposed rule change.
Consequently, the collective-choice rules used for changing operational
rules related to appropriation affect whether an institutional change favored
by some and opposed by others will be adopted. For any collectivechoice
rule, such as unanimity, majority, ruling elite, or one-person rule,
a minimum coalition of appropriators, M,A, must agree prior to the
adoption of new rules. If for all coalitions,
Γm ≤ (C1m + C2m + C3m),
no new rules will be adopted. And if for at least one coalition M,A,
Γm > (C1m + C2m + C3m),
for all members of M, a new set of rules may be adopted. If there are
several such coalitions, the question of which coalition will form, and
thus which rules will result, depends on the relative resources of the coalitions
and their bargaining strength (see J. Knight 1992; Ensminger and
Knight 1997). This analysis is applicable to a situation where a set of
appropriators starts with only default conditions—open access—and contemplates
adopting its first rules limiting access. Appropriators considering
changing operational rules over time would also use such a general
calculus.11
The collective-choice rule used in field settings varies from reliance on
the decision of one chief or a few members of an elite, to a reliance on
majority or supermajority vote, all the way to reliance on consensus or
close to unanimity (Buchanan and Tullock 1962). If there are substantial
differences in the perceived benefits and costs of appropriators, it is possible
that M appropriators will impose a new set of rules on the A–M other
  
U SI NG RU LE S A S T OO LS TO CO PE 247
appropriators that strongly favor those in the winning coalition and impose
losses or lower benefits on those in the losing coalition (Thompson,
Mannix, and Bazerman 1988). If expected benefits from a change in institutional
arrangements are not greater than expected costs for many appropriators,
however, the costs of enforcing a change in institutions will
be much higher than when most participants expect to benefit from a
change in rules over time.
Where the enforcement costs are fully borne by the members of M,
operational rules that benefit the A−M other appropriators lower the
long-term costs of monitoring and sanctioning for a governing coalition.
Where external authorities enforce the rules agreed upon by M appropriators,
the distribution of costs and benefits are more likely to benefit M
and may impose costs on the A−M other appropriators (Ensminger and
Knight 1997; Mwangi 2003).
RESOURCE ATTRIBUTES AND CALCULATING COSTS AND BENEFITS
The attributes of a resource (listed above) affect both the benefits and
costs of institutional change. Linking these attributes of the biophysical
world to the perception of appropriators enables one to develop an underlying
metric of perceived net benefits to explain why self-organization
occurs in some locations and not in others. If resource units are relatively
abundant (R1), little reason exists for appropriators to invest costly time
and effort in organizing. If the resource is already substantially destroyed,
the high costs of organizing may not generate sufficient benefits. Self-organization
is likely to occur only after appropriators observe substantial
scarcity. The danger here, however, is that exogenous shocks leading to a
change in relative abundance of the resource units may occur rapidly, and
appropriators may not adapt quickly enough to the new circumstances
(Libecap and Wiggins 1985; Baker 2005).
The presence of reliable indicators about the conditions of a resource
(R2) affects the capacity of appropriators to adapt relatively soon to
changes that could adversely affect their long-term benefit stream (Moxnes
1996). A resource flow that is highly predictable (R3) is much easier
to understand and manage than one that is erratic (Schlager, Blomquist,
and Tang 1994). In the latter case, it is always difficult for appropriators
(or, for that matter, for scientists and government officials) to judge
whether changes in the resource stock or flow are due to overharvesting
or to random exogenous variables.12
Unpredictability of resource units in
smaller locations, such as private pastures, may lead appropriators to
create a larger common-property unit to increase the predictability of
resource availability somewhere in the larger unit (Netting 1972;
  
C HA PT ER EI GH T248
Wilson and Thompson 1993). The spatial extent of a resource (R4) affects
the costs of defining reasonable boundaries and then of monitoring them
over time.
APPROPRIATOR ATTRIBUTES AND CALCULATING COSTS AND BENEFITS
Appropriators’ attributes also affect expected extrinsic and intrinsic benefits
and costs. If appropriators do not obtain a major part of their income
from a resource or value it highly for some other purpose (A1), the high
costs of organizing and maintaining a self-governing system may not be
worth their effort (Lawry 1990; Gibson 2001). If appropriators do not
share a common understanding of how complex resource systems operate
(A2), they will find it extremely difficult to agree on future joint strategies.
As Libecap and Wiggins (1985) argue, asymmetric private information
about heterogeneous assets may adversely affect the willingness of participants
to agree to a reduction in their use patterns before considerable
damage is done to a resource. Given the complexity of many commonpool
resources—especially multispecies or multiproduct resources—understanding
how these systems work may be counterintuitive even for
those who make daily contacts with the resource.
Appropriators with many other viable and attractive options, who thus
discount the importance of future income from a particular resource (A3),
may prefer to “mine” a resource without spending resources to regulate
it. In light of his study of many fisheries, Berkes (1985, 201) noted that
“community control over the fishing effort appears to be very difficult to
achieve in commercial fisheries in general.” He was pessimistic enough
about the likelihood of local organization to reflect that if a “given stock is
not overexploited, this is probably related to insufficient market demand
rather than to community-level controls” (201). We will discuss a more
optimistic picture in chapter 9, however, of the Maine lobster fishery.
Maine lobster fishers are embedded in a polycentric system where smallscale
commercial interests are centered in communities that have considerable
autonomy to craft rules that have enabled the fishery to flourish. In
many other cases, commercial fishing firms simply move on to another
resource once one is destroyed, assuming there will always be other resources
available to them. Berkes and colleagues (2001) also discuss
how comanagement strategies involving organizing at a local level along
with an active role for a larger-scale government are challenging to create
but are an effective way to slowly increase the time horizon of fishers in
these systems.
Appropriators who trust one another (A4) to keep agreements and use
reciprocity in their relationships with one another face lower expected
costs involved in monitoring and sanctioning one another over time. Ap-
  
U SI NG RU LE S A S T OO LS TO CO PE 249
propriators who lack trust at the beginning of a process of organizing
may be able to build this form of social capital (Coleman 1988; E. Ostrom
and Ahn 2003) if they initially adopt small changes that most appropriators
follow before trying to make major institutional changes. Autonomy
(A5) tends to lower the costs of organizing. A group that has little autonomy
may find that those who disagree with locally developed rules seek
contacts with higher-level officials to undo the efforts of appropriators to
achieve regulation.13
Prior experience with other forms of local organization
(A6) greatly enhances the repertoire of rules and strategies known by
local participants as potentially useful to achieve various forms of regulation.
Further, appropriators are more likely to agree upon rules whose
operation they understand from prior experience, rather than rules that
are introduced by external actors and are new to their experience. Given
the complexity of many field settings, appropriators face a difficult task
in evaluating how diverse variables affect expected benefits and costs over
a long time horizon.
Attributes of the resource also affect the attributes of appropriators.
In highly variable resources (R3), for example, it may be particularly difficult
to understand and to sort out outcomes stemming from exogenous
factors and those resulting from the actions of appropriators (McKean
2000). Brander and Taylor (1998) have argued that when the resource
base itself grows very slowly, population growth may exceed the carrying
capacity before participants have achieved a common understanding of
the problem they face (see also Reuveny and Maxwell 2001; Decker and
Reuveny 2005). Rolett and Diamond (2004) identify nine biophysical
variables that are significant predictors of the historical deforestation of
Pacific islands, irrespective of the culture and traditions of the pre-European
settlers.
Many aspects of the macroinstitutional structure surrounding a particular
setting affect the perceived costs and benefits. Thus, external authorities
can do a lot to enhance the likelihood and performance of self-governing
institutions (Shivakumar 2005). The availability of open and fair
courts for resolving conflicts is one important facility that larger governance
units can provide to increase the capability of smaller units. The
actions of external authorities can also seriously impede these developments
as well. Further, when the activities of one set of appropriators, A,
have “spillover effects” on others beyond A, external authorities can either
facilitate processes that allow multiple groups to solve conflicts arising
from negative spillovers or take a more active role in governing particular
resources themselves.
Appropriators in the field rarely face a setting that generates clear-cut,
expected benefit-cost ratios. The collective-choice rules in some settings
give a small elite substantial power to block suggested changes that may
  
C HA PT ER EI GH T250
generate positive gains for most appropriators but some losses for those in
power. Consequently, we cannot conclude that most appropriators using
common-pool resources will undertake self-governed regulation. Many
settings exist where the theoretical expectation should be the opposite:
Appropriators will overuse the resource unless efforts are made to change
one or more of the variables affecting perceived costs or benefits. Given
the number of variables that affect these costs and benefits, many points
of external intervention can enhance or reduce the probability of appropriators’
agreeing upon and following rules that generate higher social
returns. Both social scientists and policymakers have a lot to learn about
how these variables operate interactively in field settings and even how
to measure them so as to conduct well-crafted empirical studies to test
the warrantability and usefulness of this calculus.
Researchers and public officials need to recognize the multiple manifestations
of these theoretical variables in the field. Appropriators may be
highly dependent on a resource (A1), for example, because they are in a
remote location and few roads exist to enable them to leave. Alternatively,
they may be located in a central location, but other opportunities are not
open to them due to lack of training or a discriminatory labor market.
Appropriator’s discount rates (A3) in relation to a particular resource
may be low because they have lived for a long time in a particular location
and expect that they and their grandchildren will remain in that location
or because they possess a secure and well-defined bundle of property
rights to this resource (see Schlager and Ostrom 1992).
Reliable indicators of the condition of a resource (R2) may result from
activities that the appropriators themselves do—such as regularly shearing
the wool from sheep (see Gilles and Jamtgaard 1981) or because of
efforts to gather reliable information by appropriators or by external authorities
(Blomquist 1992). Predictability of resource units (R3) may result
from a clear regularity in the natural environment of the resource or
because storage has been constructed in order to even out the flow of
resource units over both good and bad years. They may have autonomy
to make their own rules (A5) because a national government is weak and
unable to exert authority over resources that it formally owns, or because
national law formally legitimates self-governance—as is the case with Japanese
inshore fisheries.
When the benefits of organizing are commonly understood by participants
to be very high, appropriators lacking many of the attributes conducive
to the development of self-governing institutions may be able to overcome
their liabilities and still develop effective agreements. Libecap
(1995, 166) reflects that “the larger the expected aggregate gains, the
more likely an acceptable share arrangement can be devised.” The crucial
factor is not, however, whether all attributes are favorable but the relative
  
U SI NG RU LE S A S T OO LS TO CO PE 251
size of the expected extrinsic and intrinsic benefits and costs they generate
as perceived by participants. All of the resource and appropriator variables
listed above affect their expected benefits and costs. It is difficult,
however, particularly for outsiders to estimate their specific impact on
expected benefits and costs given the difficulty of making precise measures
of many of these variables and weighing them on a cumulative scale.
Theoretical Puzzles
In addition to the growing consensus concerning the variables most likely
to be associated with self-organization, many unresolved theoretical issues
still exist. Two major theoretical questions relate to the effect of the
number of appropriators involved and their heterogeneity on the likelihood
of self-organization and the type of rules designed.
Size
Many theorists argue that the size of a group is negatively related to
solving collective-action problems in general. Many results from gametheoretical
analysis of repeated games conclude that cooperative strategies
are more likely to emerge and be sustained in smaller rather than
larger groups (see synthesis of this literature in Baland and Platteau 1996).
Scholars who have studied self-organized resource systems in the field
point to the increased transaction costs of larger groups and tend to conclude
that success will more likely happen in smaller groups (see, for example,
Barker et al. 1984; Cernea 1989; Wilson and Thompson 1993;
Meinzen-Dick, Raju, and Gulati 2002). Libecap (1995) reflects that the
common-pool resource experiments without communication, discussed
in chapter 3, are closer to his experience of studying large groups of oil
producers, fishers, and orange producers than the experiments with
communication.
On the other hand, Tang (1992, 68) did not find a statistical relationship
within the 37 farmer-governed systems he studied (which varied from
7 to 300 farmers) between the number of appropriators or the amount
of land being irrigated and performance variables. In Lam’s (1998, 115)
analysis of the performance of a much larger set of irrigation systems in
Nepal ranging in size up to 475 irrigators, he also did not find any significant
relationship between either the number of appropriators or the
amount of land included in the service area with any of the three performance
variables he studied. On the other hand, in a systematic study of
forest institutions, Agrawal (2000) found a curvilinear pattern. Both
smaller and much larger forest user groups were not as able to undertake
  
C HA PT ER EI GH T252
the level of monitoring needed to protect forest resources as moderately
sized groups.
One of the problems with a focus on size of group as a key determining
factor is that many other variables change as group size increases (Chamberlin
1974; R. Hardin 1982). If the costs of providing a public good
related to the use of a common-pool resource, say a sanctioning system,
remain relatively constant as group size increases, then increasing the
number of participants brings additional resources that could be drawn
upon to provide the benefit enjoyed by all (see Isaac, Walker, and Williams
1994). Marwell and Oliver (1993, 45) conclude that when a “good has
pure jointness of supply, group size has a positive effect on the probability
that it will be provided.” On the other hand, if one is analyzing the conflict
levels over a subtractable good and the transaction costs of arriving at
acceptable allocation formulas, group size may well exacerbate the problems
of self-governing systems. Since there are trade-offs among various
impacts of size on other variables, a better working hypothesis is that
group size has a curvilinear relationship to performance.
Heterogeneity
Heterogeneity is also a highly contested variable. For one thing, groups
can differ along a diversity of dimensions including their cultural backgrounds,
interests, and endowments (see Baland and Platteau 1996;
Platteau 2004). Each may operate differently. If groups coming from diverse
cultural backgrounds share access to a common resource, the key
question affecting the likelihood of self-organized solutions is whether the
views of the multiple groups concerning the structure of the resource,
authority, interpretation of rules, trust, and reciprocity differ or are similar.
New settlers to a region may simply learn and accept the rules of the
established group, and their cultural differences on other fronts do not
affect their participation in governing a resource. On the other hand, new
settlers are frequently highly disruptive to the sustenance of a self-governing
enterprise. They may not recognize the legitimacy of the local rules
and may place heavy demands on a resource.
When the interests of appropriators differ, achieving a self-governing
solution to common-pool resource problems is particularly challenging
(Libecap 1995). Appropriators who possess more substantial economic
and political assets may have similar interests to those with fewer assets
or they may differ substantially on multiple attributes. When the more
powerful have similar interests, they may greatly enhance the probability
of successful organization if they invest their resources in organizing a
group and devising rules to govern that group. Those with substantial
economic and political assets are more likely to be a member of a minimal
  
U SI NG RU LE S A S T OO LS TO CO PE 253
winning coalition—and thus have a bigger impact on decisions about institutional
changes. Mancur Olson (1965) long ago recognized the possibility
of a privileged group whereby some of those possessing a large share
of political and economic assets were sufficiently affected to bear a disproportionate
share of the costs of organizing to provide public goods (such
as the organization of a collectivity). On the other hand, if those with
more assets also have low discount rates (A3) related to a particular resource
and lower salience (A1), they may simply be unwilling to expend
inputs or actually impede organizational efforts that might lead to their
having to cut back on their productive activities.
Appropriators may also design institutions that cope effectively with
heterogeneities. In a study of eighteen forestry user groups in Nepal, Varughese
and Ostrom (2001) found that wealth disparities and locational or
sociocultural differences had no impact on the measured level of collective
action and forest conditions. When groups adopted rules that allocate
benefits using the same formulas used to allocate duties and responsibilities,
appropriators who differ significantly in terms of assets will tend to
agree and follow such rules. Poteete and Ostrom (2004) reviewed the
findings from five studies conducted by scholars associated with the International
Forestry Resources and Institutions (IFRI) research network,
who used the same research protocols to measure group and forest characteristics
and the rules crafted by these groups. In these studies, heterogeneity
was not consistently a negative factor affecting forest conditions. In
many cases, appropriators had designed rules that took into account the
diverse forms of heterogeneity found in a user group. These ingenious
rules enabled a group to overcome the potential resentment and injustices
associated with heterogeneity. On the other hand, when heterogeneity is
accentuated by rules rather than counteracted by rules, Platteau (2003)
has documented how inequalities can lead to further inequalities.
Summing Up
Even in a group that differs on many variables, if at least a minimally
winning subset of M appropriators harvesting an endangered but valuable
resource are dependent on it (A1), share a common understanding of their
situations (A2), have a low discount rate (A3), trust one another (A4),
and have autonomy to make their own rules (A5), it is more likely that
they will estimate the expected benefits of governing their resource greater
than the expected costs. Whether the rules agreed upon distribute benefits
and costs fairly depends both on the collective-choice rule used and the
type of heterogeneity existing in the community. Neither size nor heterogeneity
are variables with a uniform effect on the likelihood of organizing
  
C HA PT ER EI GH T254
and sustaining self-governing enterprises. The debate about their effect is
focusing on the wrong variables. Instead of focusing on size or the various
kinds of heterogeneity by themselves, it is important to ask how these
variables interact with other variables as they impact the benefit-cost calculus
of those involved in negotiating and sustaining agreements. Their
impact on costs of producing and distributing information (Scott 1993,
forthcoming) is particularly important.
For appropriators to cope with the complexity of experimenting with
the rules that they could use to sustain a common-pool resource, they
have to conclude that the expected benefits from an institutional change
will exceed the immediate and long-term expected costs. When appropriators
cannot communicate and have no way of gaining trust through their
own efforts or with the help of the macroinstitutional system within
which they are embedded, the prediction of an incapacity to extract themselves
from a pattern of overuse is likely to be empirically supported.
Ocean fisheries, the stratosphere, and other global commons come closest
to the appropriate empirical referents (E. Ostrom et al. 1999).
If appropriators can engage in face-to-face bargaining and have autonomy
to change their rules, they may well attempt to organize themselves.
Whether they organize depends on attributes of the resource system and
the appropriators themselves that affect the benefits to be achieved and
the costs of achieving them. Whether their self-governed enterprise succeeds
over the long term depends on whether they can successfully experiment
with a subset of the rules that are used to govern common-pool
resources and a configuration of rules that are easy to understand and
monitor, keep harvesting levels within bounds, and are considered equitable
by most appropriators.
Once one adopts the view that one cannot create the perfect set of rules
and that all efforts at reform must be viewed as experiments, one recognizes
that policy analysis can never find “the” answer. We can analyze the
effect of rules in highly simplified game-theoretic analyses. We can certainly
expand knowledge about the rich diversity of rules used in practice.
Appropriators in field settings across time and space have already devised
an incredible richness in the rules they use. We need to learn more about
this heritage so as to be better facilitators of building adaptive institutional
designs—in contrast to presuming we are the experts who can devise
the optimal design to solve a complex problem. All analyses of potential
institutional reforms are partial analyses. We can improve their
quality by carefully studying rules-in-use and the incentives, interactions,
and outcomes they generate in light of the biophysical and social world
in which they exist.
  
Nine
Robust Resource Governance in Polycentric
Institutions
THE STUDY OF the rules actually used in many field settings across the
world to regulate the use of common-pool resources leads to an unsettling
conclusion. We must conclude that those making rules in efforts to improve
outcomes in this policy domain can undertake only partial analyses
of a limited set of potential rules and their impact on actions and outcomes
in specific environments. No one can undertake a complete analysis
of all of the potential rules that they might use and analytically determine
which set of rules will be optimal for the outcomes they value in a particular
ecological, economic, social, and political setting. One must recognize
that policies involving rule changes must be viewed as experiments. Further,
since ecological, economic, social, and political settings are always
changing over time, no specific set of rules will produce the same distribution
of benefits and costs over time.
For some readers, this is a depressing lesson. They are looking for “the”
answer of how best to solve commons dilemmas and other policy problems.
We all recognize that some efforts at designing or reforming rules
have had disastrous results. Developing a “sure-fire” method to avoid
all disasters stemming from rules that generate perverse incentives in a
particular environment sounds like a great advance. I am sure that the
designers of modern airplanes would also like to have a sure-fire method
to test out all contingencies before sending planes into the air. Similarly,
the designers of high-speed computers and software would appreciate
having a method that would enable them to produce a “crash-proof”
computer system.
The contemporary levels of knowledge related to designing new institutions
for governing complex resource systems, airplanes to fly through
uncertain weather, and computers subject to diverse exigencies, are substantial,
but not complete (H. Simon 1981). And, I am willing to predict
given the large number of components that combine in a nonadditive
fashion, that our knowledge of how to design these systems will continue
to grow but will never be complete. As soon as one design has proved itself
in one environment, innovations in strategies adopted by participants or
changes in the environment in which a humanly designed system is in
operation will produce unexpected results.
  
C HA PT ER NI NE256
We are not, however, helpless in finding ways to improve the performance
of complex social-ecological systems. We cannot conduct full
analyses of the consequences of changing all possible parts of a complex
system interacting itself with a complex and changing environment. Further,
officials and policy analysts who presume that they have the right
design can be dangerous. They are likely to assume that citizens are
short-sighted and motivated only by extrinsic benefits and costs. Somehow,
the officials and policy analysts assume that they have different motivations
and can find the optimal policy because they are not directly involved
in the problem (Moore 1995). They are indeed isolated from the
problems. This leaves them with little capability to adapt and learn in
light of information about outcomes resulting from their policies.1
All
too often, these “optimal” policies have Leviathan-like characteristics
to them.2
Continuing to presume that complex policy problems are simple problems
that can be solved through the adoption of simple designs that are
given general names, such as private property, government ownership, or
community organization, is a dangerous academic approach. Dichotomizing
the institutional world into “the market” as contrasted to “the
state” is so grossly inadequate and barren that it is surprising how the
dichotomy survives as a basic way of organizing academic studies and
policy advice. Oversimplification of our design options is dangerous since
it hides more of the working parts needed to design effective, sustainable
institutions than it reveals (Seabright 1993). And, it reduces our awareness
of the need to monitor outcomes and improve them over time
through better processes of learning and adaptation.
The language developed in this book to identify the working components
of action arenas that exist everywhere (chapters 1 through 4); to
analyze the similarities and differences in rules, norms, and strategies
(chapter 5); and then to group similar rules together by the component
of an action situation they directly affect (chapters 6 through 8), is undoubtedly
more complex than many contemporary scholars would prefer.
This complexity of language has not been introduced lightly. A scholar
should also keep analysis as simple as possible—given the problems to be
analyzed. Just as important, however, is developing a mode of analysis
that enables scholars, policymakers, and participants in ongoing processes
to grapple with the problems they face by digging through the layers
of nested systems in which these processes exist. When one is analyzing
what is operationally a relatively simple system using a relatively simple
language for analysis, one may not need the full language system developed
in this volume. Most common-pool resources, and many other policy
fields, however, are complex systems and not simple systems. Thus, we
need a consistent, nested set of concepts that can be used in our analysis,
  
R OB US T R ES OU RC E G OV ER NA NC E 257
research, and policy advice in a cumulative manner. The concepts developed
in this book do, I hope, form the foundation for such an endeavor.
They are derived from a commitment to use theory to observe the institutional
world and to learn from that observation and measurement. And,
of course, given the central lesson just discussed, institutional analysts
will improve on these concepts over time as further research and policy
advice uses these concepts.
In chapter 8, after examining the kinds of rules used in the field for
coping with common-pool resources in many settings, I stressed the impossibility
of conducting a full analysis of all options available to officials,
appropriators, scholars, and others interested in improving the performance
of resource governance institutions. The question to be examined
in this final chapter is whether methods exist that can be used to learn
more effectively from the experience of engaging in “reforms as experiments”
(Campbell 1969). Are there ways that we can avoid some of the
disastrous results that have been produced by systems of governance in
the contemporary world? My answer is yes. While no “sure-fire” methods
exist, I will argue that there are approaches to speed up and share the
learning that can result from tinkering with rules and gaining experience
with outcomes. E. Jones (2003), for example, developed a graphical
method for analyzing relationships in these complex systems.
First, instead of trying to search for the single set of rules that is the
optimal set for every type of problem, I will again urge the importance of
studying the underlying designs of those real-world experiments that have
proved to be robust over time as I did in Governing the Commons (see
E. Ostrom 1990). In the first part of this chapter, I will review what
we have learned since 1990 about design principles related to robust,
common-pool resource institutions. In light of still further evidence about
the performance of self-organized systems that are consistent with the
earlier derived design principles, we can conclude that there are ways of
organizing governance that increase the opportunities for adaptation and
learning in a changing and uncertain world with continuing advances in
knowledge and technologies.
The design principles are not blueprints, however! They describe the
broad structural similarities among those self-organized systems that have
been able to adapt and learn so as to be robust to the many social, economic,
and ecological disturbances that occur over time. Threats always
challenge the robustness of any system—no matter how well it fits the
best design principles known for a particular problem Thus, the second
topic of this chapter is a discussion of the threats that exist to any set of
self-organized, resource governance systems. Since one of the important
threats is the effort to impose uniform rules and large boundaries on systems
so they are more comprehensible to academics and policymakers, I
  
C HA PT ER NI NE258
will conclude this chapter by urging readers to think more positively
about the complex, polycentric systems of governance that are created by
individuals who have considerable autonomy to engage in self-governance.
Given the wide variety of ecological problems that individuals face
at diverse scales, an important design principle is getting the boundaries
of any one system roughly to fit the ecological boundaries of the problem
it is designed to address. Since most ecological problems are nested from
very small local ecologies to those of global proportions, following this
principle requires a substantial investment in governance systems at multiple
levels—each with some autonomy but each exposed to information,
sanctioning, and actions from below and above (Low et al. 2003; Folke,
Berkes, and Colding 1998; Moran and Ostrom 2005).
Design Principles and Robust Social-Ecological Systems
The findings briefly reviewed in chapter 8—that self-organized systems
have frequently solved many Commons Dilemmas—have surprised many
scholars. These findings differ from the grim predictions made in the
1970s and 1980s that individuals were trapped in unproductive situations
and could not themselves restructure their perverse incentives. One cannot,
of course, now substitute for the earlier grim predictions a presumption
of a uniformly successful conquering of collective-action problems.
As social scientists, we have to use one of our favorite slogans once
again—it depends!
In my earlier effort to understand the governance systems that had been
identified during the last half of the 1980s as long-surviving systems, I first
tried to identify the specific rules used by the systems that had survived for
a long period of time using Kenneth Shepsle’s (1989) definition of a robust
institution. Shepsle considered a system to be robust if it was long-lasting
and the operational rules had been devised and modified over time according
to a set of collective-choice rules (which themselves might be modified
more slowly over time within a set of constitutional-choice rules,
which were modified, if at all, very infrequently). The contemporary use
of the term robustness in regard to complex systems focuses on adaptability
to disturbances: “the maintenance of some desired system characteristics
despite fluctuations in the behavior of its component parts or its environment”
(Carlson and Doyle 2002, 2538; see also Anderies, Janssen,
and Ostrom 2004).
Among the many governance systems that met Shepsle’s criteria for
robustness—as well as the criteria specified by Carlson and Doyle—the
specific operational and collective-choice rules that were observed varied
dramatically from one system to the next. It was frustrating that I could
not identify any particular rules consistently associated with robust gover-
  
R OB US T R ES OU RC E G OV ER NA NC E 259
nance of common-pool resources. Instead of focusing on specific rules,
my effort turned to identifying eight underlying design principles that
characterized robust common-property institutions. No assertion was
made that those crafting these institutions self-consciously used the design
principles. Rather, it was my thought that robust systems had simply met
most of these principles and that those systems that had collapsed or were
performing ineffectively were not so structured.
Design principles derived from studies of long-enduring institutions for governing
sustainable resources:
1. Clearly defined boundaries. The boundaries of the resource system (e.g.,
irrigation system or fishery) and the individuals or households with rights to
harvest resource units are clearly defined.
2. Proportional equivalence between benefits and costs. Rules specifying the
amount of resource products that a user is allocated are related to local conditions
and to rules requiring labor, materials, and/or money inputs.
3. Collective-choice arrangements. Many of the individuals affected by
harvesting and protection rules are included in the group who can modify
these rules.
4. Monitoring. Monitors, who actively audit biophysical conditions and user
behavior, are at least partially accountable to the users and/or are the users
themselves.
5. Graduated sanctions. Users who violate rules-in-use are likely to receive
graduated sanctions (depending on the seriousness and context of the offense)
from other users, from officials accountable to these users, or from both.
6. Conflict-resolution mechanisms. Users and their officials have rapid access
to low-cost, local arenas to resolve conflict among users or between users
and officials.
7. Minimal recognition of rights to organize. The rights of users to devise
their own institutions are not challenged by external governmental authorities,
and users have long-term tenure rights to the resource.
For resources that are parts of larger systems:
8. Nestled enterprises. Appropriation, provision, monitoring, enforcement,
conflict resolution, and governance activities are organized in multiple layers
of nested enterprises (based on E. Ostrom 1990, 90).
When I first speculated about these design principles (E. Ostrom 1990),
I stressed the speculative nature of my efforts and urged others to test out
these tentative conclusions through further empirical research that would
help ascertain if these principles distinguished between robust and failed
systems. Since Governing the Commons was published, other scholars
have responded to the challenge and examined the relevance of these principles
for helping to explain the performance of resource governance systems
(such as fisheries, irrigation systems, pastures, and forests) throughout
the world (see Dietz, Ostrom, and Stern 2003).3
  
C HA PT ER NI NE260
Martin S. Weinstein (2000), for example, examined indigenous inshore
fishery institutions in Canada and Japan and found that these extremely
long-lived institutions were characterized to a large extent by the design
principles that I had earlier proposed. Abernathy and Sally (2000) studied
nine small but long-surviving irrigation systems in the dry areas of Burkina
Faso and Niger. They measured system performance using both physical
and nonphysical factors and did not find a single indicator that could
systematically be used to measure system performance. They found that
an average measure of performance based on six indicators, on the other
hand, was highly correlated with governance arrangements conforming
to the design principles listed above. Other studies of irrigation systems
that found systems that were characterized by the design principles to
exhibit robustness include works by Crook and Jones (1999), Guillet
(1992a, 1992b), Gupta and Tiwari (2002), and Merrey (1996).
Haley (2002) examined a somewhat different question using the design
principles: the performance of the same private oil corporation—Arco—
related to the exploitation of two different oil fields located in indigenous
territories. Arco discovered oil in 1992 in the Pastaza Province in Eastern
Ecuador near a Quichua indigenous community (the Villano field) and in
1994 in the Colville Delta in northern Alaska on land owned by an Inupiat
Eskimo community (the Alpine field). Both communities have received
benefits from the oil revenues generated. Haley estimated that the indigenous
community received around 3.2 percent of the total government and
resource owner take for the Alpine field, but less than 1 percent of the
government share (and not even the total return) related to the Villano
field. Haley appraised each case for the presence or absence of the design
principles and found that the governance arrangements for the Alpine
field were consistent with all of the design principles. For the first decade,
the arrangements for the Villano field were not consistent with any of the
design principles. Recently, the Villano field has developed some ad hoc
arrangements that are consistent with two of the principles.
In light of the positive reaction to these design principles, let us briefly
review some of the research that has focused specifically on each of them.
Well-Defined Boundaries
The first principle is that the boundaries of the resource system, as well
as the individuals or households with rights to harvest resource units, are
clearly defined. The problem that is addressed by systems that do define
their boundaries is clearly free-riding. If a group of users can determine
their own membership—including those who agree to use the resource
according to their agreed-upon rules and excluding those who do not
agree to these rules—the group has made an important first step toward
  
R OB US T R ES OU RC E G OV ER NA NC E 261
limiting access and developing greater trust and reciprocity. Using this
principle enables participants to know who is in and who is out of a
defined set of relationships and, thus, with whom to cooperate. Smaller
resource governance systems do not always have extensively developed
rule systems, but those that are robust do demark their boundaries (see,
for example, Schlager 1994; Berkes et al. 2001).
Group boundaries are frequently marked by well-understood attributes,
such as residing in a particular community or joining a specific
local cooperative (as shown in table 8.1 in chapter 8). Further, membership
may be marked by various “tags”—symbolic boundaries—and involve
complex rituals and beliefs that help solidify individual beliefs about
the trustworthiness of others. Contemporary developments in evolutionary
theory applied to cultural systems and processes of adaptation help
to explain how these design principles work to help groups sustain and
build their cooperation over long periods of time (Janssen and Ostrom
forthcoming b).
Just defining the resource boundaries carefully, however, may not be
sufficient in and of itself! In a study of irrigation systems in Nepal, Shukla
(2002) found that almost all of these systems have well-demarked boundaries.
A substantial difference exists, however, between the systems that
have been designed, built, and maintained by farmers as contrasted to
the systems designed by government engineers. On the farmer-designed
systems, the farmers themselves determine how large the area to be served
should be. The farmers who demark the boundary will also have to participate
in the construction of the system and its maintenance by contributing
time, materials, and potentially some funds. Thus, the boundaries of
irrigation systems developed by farmers tend to be conservative so that
those who make the system work are more assured of getting water. Farmers
on these systems—even at the tail end—tend to receive water in the
dry season as a consequence of keeping their systems small and the other
elements of the physical and institutional structures they build.
The boundaries of those systems constructed by government agencies,
by contrast, are frequently demarked as part of donor-funded projects.
Irrigation engineers are strongly motivated to show a positive benefit-cost
ratio. The more farmers placed within the service boundary of a system,
the higher the benefits that can be reported in the plans submitted to donors
for funding (Palanisami 1982; R. Repetto 1986). Once funding is
granted, few efforts are made to check the reliability of earlier estimates.
Farmers in the larger service area are promised water but may not receive
a reliable supply. Farmers on these systems are more likely to steal water
and less likely to contribute resources to maintenance. Thus, appropriatordefined
boundaries tend to include a clear set of participants who know
that they have mutual responsibilities as well as benefits.
  
C HA PT ER NI NE262
Externally imposed boundaries may not be viewed as legitimate by
those who have cared for a resource for long periods of time. If imposed
boundaries are enforced, they generate substantial costs for local peoples
(Ghate 2003). On the other hand, the boundaries may not even be known
to local appropriators. Paper parks have been created in the capitals of
many countries that look clear on the official maps but are not demarked
or enforced locally (Hayes 2004; Dietz, Ostrom, and Stern 2003).
Some governmental reserves have had remarkable stability over time.
Vogt and colleagues (2005) have dug into archives and talked with older
residents and officials to determine why remotely sensed images of forest
reserves in one region of Uganda show remarkably stable boundaries
when so many forest reserves have failed to reduce deforestation within
their boundaries. In this case, Vogt and colleagues determined that the
boundaries were negotiated as part of the 1900 agreement between the
Regents of the Buganda Kingdom and the British colonial government—
and have thus had legitimacy for more than a century. Further, the boundaries
have been clearly demarked with stone-covered cairns, and specific
tree species were planted in the cairns. Local clan elders and traditional
administrators participated in the original demarcation, and local residents
continue to participate in the renewal of these boundaries every two
decades or so. Strict enforcement of the boundaries backed by legitimate
and well-known boundaries have led to a remarkable stability in a region
of Africa where many government forests have been extensively deforested
by local residents (see also Dietz, Ostrom, and Stern 2003).
In a thoughtful analysis of the usefulness of the design principles for
analyzing why some donor-sponsored conservation projects have failed
while others have succeeded, Morrow and Hull (1996) pointed out that
many donor projects formally met the first design principle. Formal congruence
with the first principle is not enough, however, to enable appropriators
to defend their borders from free riders. Morrow and Hull suggested
a rephrasing for the first design principle to be: “The resource itself
and the users of the resources are clearly defined, and the appropriators
are able to effectively defend the resource from outsiders” (1643). Given
our own research on the importance of defending the boundaries that are
demarked, this rephrasing is a positive step forward (Dietz, Ostrom, and
Stern 2003; Gibson, Williams, and Ostrom 2005).
Proportional Equivalence between Benefits and Costs
The second design principle is that the rules-in-use allocate benefits proportional
to inputs that are required. If a group of users is going to harvest
from a resource over the long run, they must devise rules related to how
much, when, and how different products are to be harvested. They also
  
R OB US T R ES OU RC E G OV ER NA NC E 263
need to assess the costs of operating a system on users. When the rules
related to the distribution of benefits are made broadly consistent with
the distribution of costs, participants are more willing to pitch in to keep
a resource well-maintained and sustainable (see, for example, Nemarundwe
and Kozanayi 2003). Relating user inputs to the benefits they
obtain is a crucial element of establishing a fair system (Trawick 2001).
If some users get all the benefits and pay few of the costs, others are not
willing to follow rules over time (Ensminger 2000). Thus, fairness is a
crucial attribute of the rules of robust systems (Chakraborty 2004).
The Chisasibi Cree have devised a complex set of entry and authority
rules related to the coastal fish stocks of James Bay, as well as the beaver
stock located in their defined hunting territory. Fikret Berkes (1987, 87)
explains that these resource systems and the rules used to regulate them
have survived and prospered for so long because effective “social mechanisms
ensure adherence to rules which exist by virtue of mutual consent
within the community. People who violate these rules suffer not only a
loss of favour from the animals (important in the Cree ideology of hunting)
but also social disgrace.” Fair rules of distribution help to build trusting
relationships since more individuals are willing to abide by these rules
because they participated in their design and also because they meet
shared concepts of fairness (Bowles 1998; Trosper 2002).
In long-surviving irrigation systems, for example, subtly different rules
are used in each system for assessing water fees used to pay for maintenance
activities, but water tends to be allocated proportional to fees or
other required inputs (Bardhan 2000; Bardhan and Dayton-Johnson
2002). Sometimes water and responsibilities for resource inputs are distributed
on a share basis, sometimes on the order in which water is taken,
and sometimes strictly on the amount of land irrigated. No single set of
rules defined for all irrigation systems in a region would satisfy the particular
problems in managing each of these broadly similar, but distinctly
different, systems (Tang 1992; Lam 1998).
Collective-Choice Arrangements
The third design principle is that most of the individuals affected by a
resource regime are authorized to participate in making and modifying
their rules. Resource regimes that use this principle are both better able to
tailor rules to local circumstances and to devise rules that are considered
fair by participants. As environments change over time, being able to craft
local rules is particularly important as officials located far away do not
know of the change. When a local elite is empowered at the collectivechoice
level, policies that primarily benefit them can be expected (Platteau
2003, 2004; Ensminger 1990).
  
C HA PT ER NI NE264
In a study of forty-eight irrigation systems in India, Bardhan (2000)
finds that the quality of maintenance of irrigation canals is significantly
lower on those systems where farmers perceive the rules to have been
made by a local elite. On the other hand, those farmers (of the 480 interviewed)
who responded that the rules for their system have been crafted
by most of the farmers, as contrasted to the elite or the government, have
a more positive attitude about the water allocation rules and the rule
compliance of other farmers. Further, in all of the villages where a government
agency decides how water is to be allocated and distributed, frequent
rule violations are reported, and farmers tend to contribute less to
the local village fund. Consistent with this is the finding by Ray and Williams
(1999) that the deadweight loss from upstream farmers stealing
water on government-owned irrigation systems in Maharashtra, India,
approaches one-fourth of the revenues that could be earned in an efficient
water allocation and pricing regime.
Knox and Meinzen-Dick (2001, 22) note that property rights “are significantly
more likely to address the interests and needs of local people
when they are not imposed from the outside but rather are based on existing
rights and reflect local values and norms.” As they point out, these
rules take time and effort to develop, try out, modify, and then experiment
with again. Users who have been engaged in this process for some time
understand the rules that they have crafted, agree on why they are using
one rule rather than another, and tend to follow their own rules to a
greater extent than those that are imposed on them. Sekher (2000) conducted
a study of villages in Orissa, India, that varied in regard to the
extent of participation of local villagers in making rules related to nearby
forests that they used. He found that the “wider the representation of the
community in the organization, the better are its chances of securing local
cooperation and rule confirmation for managing and preserving the resource”
(8).
In a comparative study of farmer-designed and governed irrigation
systems (FMIS), as contrasted to those designed and operated by engineers
without involvement of the farmers in making rules to govern these
systems, Shukla (2002, 83), a water engineer himself, is relatively critical
of the “unrealistic planning and design, incomplete development, a nonsystematic
and inadequate maintenance program, deficit operation, and
lack of participation of the users” that characterized many of these systems
in Nepal. Drawing on the earlier research of Pant and Lohani (1983),
Yoder (1994), Lam (1998), and Pradhan (1989), Shukla identifies the following
as the strengths of the farmer-designed systems: “(1) Their technical
deficiencies are compensated by management inputs; (2) they are low
cost and based on local resources; (3) effective irrigation organizations
  
R OB US T R ES OU RC E G OV ER NA NC E 265
exist in most FMIS; (4) most FMIS have well-defined rules and roles for
water allocation, distribution, resource mobilization, and conflict resolution;
and (5) the leaders of these systems are accountable to the users”
(2002, 83).
Monitoring
Few long-surviving resource regimes rely primarily on endogenous levels
of trust and reciprocity among appropriators to keep rule breaking levels
down. Evidence of the consequence of inadequate monitoring is convincingly
presented by Schweik (2000). It is obvious to most institutional analysts
that rules must be enforced in some manner to achieve robust governance;
the question of how rules will actually be enforced is frequently
ignored when proposed institutional changes are analyzed and a reform
is proposed. All too many “comanaged paper parks” have been drafted
in the home office of an overseas donor or even in a country’s capital city
only to be destroyed by illegal harvesting in the specified territory. While
many agree that rule enforcement is necessary for creating a sustainable
resource over time, a vigorous debate is raging about who should be the
monitors (Bruner et al. 2001; Igoe 2004; Hockings and Phillips 1999;
Stevens 1997; Wells and Brandon 1992).
Most long-surviving resource regimes select their own monitors, who
are accountable to the appropriators or are appropriators themselves and
who keep an eye on resource conditions as well as on harvesting activities
(design principle 4). By creating official positions for local monitors, a
resource regime does not have to rely only on the norms of local appropriators
to impose personal costs on those who break a rule. The community
creates an official position. In some systems, appropriators rotate into this
position so everyone has a duty to be a monitor. In other systems, all
participants contribute resources and they jointly hire monitors. With
local monitors, conditional cooperators are assured that someone is generally
checking on the conformance of others to local rules. Thus, they
can continue their own cooperation without constant fear that others are
taking advantage of them.
Some government-owned forests have successfully adopted monitoring
arrangements similar to those of self-organized systems. Banana and
Gombya-Ssembajjwe (2000) compare the Echuya Forest in Uganda with
three other government-owned forests and one private forest. In the three
other government-owned forests, they found extensive illegal harvesting—charcoal
burning, pit-sawing, grazing, and cutting commercial
firewood—in the forest plots they had randomly selected for observation
and measurement. Over 70 percent of the sample forest plots in the three
  
C HA PT ER NI NE266
other government forests contained evidence of illegal harvesting (90). In
the Echuya Forest and the private forest, the level of illegal harvest was
relatively minor—only 20 percent of the sample plots showed any evidence
of illegal harvesting (90). In Echuya, members of an Abayanda
pygmy community, who live in the forest itself, have been asked by the
government to monitor local harvesting from the predominately bamboo
forest that is officially limited to one day a week. Even though the Echuya
Forest is quite large, local monitors have made a significant difference in
the level of illegal harvesting. A similar system drawing on local appropriators
to monitor government reserves has evolved the State of Rondoˆnia
in Brazil. In a series of forest reserves located near to a large area devoted
to colonist settlements, Batistella (2001) has documented the positive impact
of the government using local rubber tappers to monitor forest use
(see online supplement to Dietz, Ostrom, and Stern 2003).
In a study of the forest conditions used by 178 forest user groups located
in twelve countries studied by the International Forestry Resources
and Institutions (IFRI) research network, Gibson, Williams, and Ostrom
(2005) found that the level of local monitoring varies substantially across
groups. One of the measures obtained in this study is the frequency with
which a local group monitors and sanctions rule breaking behavior in
the forest.4
We examined the impact of this variable on appropriators’
assessment of forest conditions (as well as on a forester’s assessment). We
also examined the impact of a group’s social capital, the group’s dependence
of forest resources, and whether the group was formally organized
or not. The result of the analysis is that regular monitoring by a local
group is more important than the other three variables in enhancing forest
conditions. Regardless of the levels of social capital, forest dependence
and formal organization, regular monitoring and sanctioning is strongly
and statistically associated with better forest conditions (Gibson, Williams,
and Ostrom 2005).
Graduated Sanctions
The fifth design principle identified earlier was the use of graduated sanctions
by robust governance arrangements. In many self-organized systems,
the first sanction imposed by a local monitor is so low as to have
no impact on the expected benefit-cost ratio of breaking local rules (given
the substantial temptations frequently involved). Rather, the initial sanction
needs to be considered more as information to the person who is
“caught” as well as to others in the community. Everyone can make an
error or can face difficult problems leading them to break a rule. A few
rule infractions, however, can generate a downward cascade of cooperation
in a group that relies only on conditional cooperation and has no
  
R OB US T R ES OU RC E G OV ER NA NC E 267
capacity to sanction (see, for example, Kikuchi et al. 1998). In a regime
that uses graduated punishments, however, a person who purposely or by
error breaks a rule is notified that others notice the infraction (thereby
increasing the individual’s confidence that others would also be caught).
Further, the individual learns that others basically continue to extend their
trust and want only a small token to convey a recognition that the mishap
occurred. Self-organized regimes rely more on what Margaret Levi (1988)
calls “quasi-voluntary” cooperation than either strictly voluntary or coerced
cooperation. A real threat to the continuance of self-organized regimes
occurs, however, if some participants break rules repeatedly. The
capability to escalate sanctions enables such a regime to warn members
that if they do not conform, they will have to pay ever higher sanctions
and may eventually be forced to leave the community.
Let me summarize the argument to this point. When the users of a
resource design their own rules (design principle 3) that are enforced by
local users or accountable to them (design principle 4) using graduated
sanctions (design principle 5) that clearly define who has rights to withdraw
from a well-defined resource (design principle 1) and that effectively
assign costs proportionate to benefits (design principle 2), collective action
and monitoring problems tend to be solved in a reinforcing manner.
Individuals who think a set of rules will be effective in producing higher
joint benefits and that monitoring (including their own) will protect them
against being a sucker, are willing to undertake conditional cooperation.
Once some users have made contingent self-commitments, they are then
motivated to monitor other people’s behavior, at least from time to time,
in order to assure themselves that others are following the rules most of
the time. Conditional cooperation and mutual monitoring reinforce one
another especially in regimes where the rules are designed to reduce monitoring
costs (Kameda, Takezawa, and Hastie 2003). Over time, further
adherence to shared norms evolves and high levels of cooperation are
achieved without the need to engage in extensive monitoring and the imposition
of costly sanctions in all cases of observed infractions in order
to achieve rule conformance.
Conflict-Resolution Mechanisms
The operation of the above principles is bolstered by the sixth principle,
which points to the importance of access to rapid, low-cost, local arenas
to resolve conflict among users or between users and officials. Rules, unlike
physical constraints, have to be understood in order to be effective.
There are always situations in which participants can interpret a rule that
they have jointly made in different ways. By devising simple, local mechanisms
to get conflicts aired immediately and resolutions that are generally
  
C HA PT ER NI NE268
known in the community, the number of conflicts that reduce trust can
be reduced. If individuals are going to follow rules over a long period of
time, some mechanism for discussing and resolving what is or is not a
rule infraction is quite necessary to the continuance of rule conformance
itself. Further, one way of reducing the problem of elite capture of a local
resource is the availability of arenas for conflict resolution at levels above
that of a local resource.
Minimal Recognition of Rights
The capability of local users to develop an ever more effective regime over
time is affected by whether they have at least minimal recognition of the
right to organize by a national or local government (design principle 7).
While some resource regimes have operated for relatively long times without
such rights (see Ghate 2000), participants have had to rely almost
entirely on unanimity as the rule used to change rules. Otherwise, any
temporarily disgruntled participant who voted against a rule change
could go to the external authorities to threaten the regime itself! Unanimity
as a decision rule for changing rules imposes high transaction costs
and prevents a group from searching for better matched rules at relatively
lower costs. Lobe and Berkes (2004) do describe a remarkable system,
however, designed by a set of fishers in Bengal who have no legal rights
and who call themselves illicit fishers, that is enforced locally without any
recognition by government officials.
Some users do devise their own rules without creating formal, governmental
jurisdictions for this purpose. In many inshore fisheries, for example,
local fishers devise extensive rules defining who can use a fishing
ground and what kind of equipment can be used (Acheson 2003; Schlager
2004). So long as external governmental officials give at least minimal
recognition to the legitimacy of such rules, the fishers themselves may be
effective enforcers of these rules when government agencies do not have
the staff to enforce them. When external governmental officials presume
that only they can make authoritative rules, then it is difficult, but not
impossible, for local users to sustain a self-organized regime (Johnson and
Libecap 1982).
Communities that have the authority to craft their own rules, however,
are frequently able to overcome the lack of local scientific knowledge if
and when reliable information is made available about complex relationships
and is understood locally. When a study was first conducted by
Workshop colleagues in the Loma Alta comuna in 1995, they found that
the comuna owned almost seven thousand hectares of land in western
Ecuador (Gibson and Becker 2000). Further, the comuna had full authority
to allocate land—including land in a high altitude fog forest—to mem-
  
R OB US T R ES OU RC E G OV ER NA NC E 269
bers of the community for their use. At the time of the initial study, our
team found that the comuna had allocated a substantial portion of their
land to community members who cut down the indigenous trees in order
to grow paja toquilla to generate fiber to make hats and other handicrafts
for income.
Members of the comuna had not recognized that the indigenous trees
growing in this fog forest were extremely efficient in capturing water from
the fog and allowing water to precipitate into the ground. The water so
captured eventually emerged as a stream many miles away. Members of
the community relied heavily on this stream to water their agricultural
plots. Given the distance involved, it would be a challenge to discover
that connection even though farmers in the community had increasing
concerns about the reduced flow of water in this stream. In other words,
the community did not recognize the “ecosystem services” that the indigenous
forest generated for their own agricultural lands.
Fortunately, one of the coleaders of our initial research team, Dusty
Becker, was able to return to Loma Alta in 1996 with a group of Earthwatch
volunteers. They spent a summer carefully monitoring the amount
of water captured by the indigenous trees in the high altitude forest as
compared to the trees planted for commercial purposes. They also recruited
some of the high school students from Loma Alta to work with
them collecting data on a daily basis. By the end of the summer, it was
clear from the data they had collected that the indigenous trees collected
substantially more water than the trees planted for commercial purposes.
Becker (2003) estimated that the community lost 2 million liters of water
per hectare per year in those sections of the forest converted to agroforestry.
The students from Loma Alta were so motivated by participating in
this summer project that they proposed making a video for their parents
that demonstrated the value of keeping the indigenous trees in their own
forest. Once the knowledge was generally available to the citizens of
Loma Alta, they voted to create a forest reserve of one thousand hectares
in the highlands of their own forest (ibid.).
Nested Enterprises
When common-pool resources are larger, an eighth design principle tends
to characterize robust systems—the presence of governance activities organized
in multiple layers of nested enterprises. The rules appropriate
for allocating water among major branches of an irrigation system, for
example, may not be appropriate for allocating water among farmers
along a single distributory channel (Yoder 1994). Consequently, among
long-enduring self-governed regimes, smaller-scale organizations tend to
be nested in ever larger organizations. O. Choe (2004) provides an excel-
  
C HA PT ER NI NE270
lent overview of how nested enterprises have successfully been used to
overcome the weakness of relying only on large-scale or small-scale units
to govern complex resource systems. We will return to this design principle
in our discussion of polycentricity in the last section of this chapter.
Using the Design Principles in Practice
It is reassuring to review the research conducted on self-organized,
common-pool resource governance systems since 1990 and find that
many scholars have agreed with my earlier speculations about the design
principles characterizing robust systems. Tucker (1999) uses the design
principles to examine the evolution of a common-property forest owned
by a community in Honduras. She finds that examining the congruence
of this system with the design principles helps to identify underlying weaknesses
in the regime that made this system more vulnerable to forest degradation
stimulated by rapid economic development. Her findings are consistent
with those of scholars who have identified multiple threats to the
sustainability of self-organized governance systems over time—the topic
of the next section of this chapter—but her analysis illustrates the connection
between the lack of an adequate response to external threats and
systems that are weak in regard to several of the design principles.
There is a danger, however, that project planners searching for the
“right” design will try to build a one-size-fits-all project supposedly based
on the design principles. Such an effort is entirely inconsistent with the
theoretical argument presented in this book concerning the importance
of matching the rules of a system to the underlying biophysical world and
type of human community involved. The question is often raised, however:
How can the design principles be used in practice in addition to their
use in research?
At a recent colloquium where the design principles were discussed,
Mike McGinnis made an interesting observation drawing broadly on the
work of Herbert Simon. He noted that Simon has repeatedly stressed the
complexity of designing humanly engineered systems whether they be
computers, road networks, or institutional arrangements. In The Sciences
of the Artificial (1981), Simon specially argues that no humanly designed,
complex system can be fully planned to achieve optimal performance.
Rather, he stressed that all complex systems must be built up from simpler
components. Simon does point out that where one begins a search to
improve the performance of a complex system, however, makes a substantial
difference in the quality and speed of the search process (see also
H. Simon, 1972, 1995, 1999).
My own conclusion related to the impossibility of doing a complete
analysis of a complex, adaptive system is, of course, strongly influenced
  
R OB US T R ES OU RC E G OV ER NA NC E 271
by the work of Simon, as well as our research on coupled social-ecological
systems. So, one way of thinking about the practical implications of the
design principles is as a beginning point for conducting a broad search
for appropriate means of solving problems. One can translate the design
principles into a series of questions that could be asked when thinking
about improving the sustainability of a common-pool resource system.
For local appropriators, a rough translation of the first six design principles
into a set of initial questions would be:
1. How can we better define the boundaries of this resource, and of the individuals
who are using it, so as to make clear who is authorized to harvest and
where harvesting is authorized?
2. How can we clarify the relationship between the benefits received and the
contributions to the costs of sustaining this system?
3. How can we enhance the participation of those involved in making key
decisions about this system?
4. Who is monitoring this system and do they face appropriate incentives
given the challenge of monitoring?
5. What are the sanctions we are authorizing and can they be adjusted so that
someone who makes an error or a small rule infraction is sufficiently warned so
as to ensure longer-term compliance without our trying to impose unrealistic
sanctions?
6. What local and regional mechanisms exist to resolve conflicts arising over
the use of this resource?
The seventh and eighth principles are targeted at a higher level of governance.
They could be translated as:
7. Are there functional and creative efforts by local appropriators to
craft effective stewardship mechanisms for local resources that should be
recognized?
8. How do we create a multiple-layer, polycentric system that can be
dynamic, adaptive, and effective over time?
These are not, of course, the only questions appropriators and officials
should ask in an effective design process, but they can be thought of as a
good beginning.
Threats to Robust Governance of Common-Pool Resources
No matter how well a governance system is initially designed, however,
all humanly designed systems are vulnerable to threats. Self-organized,
resource-governance regimes are no exception. Both exogenous and endogenous
factors challenge their long-term viability. Robust institutions
  
C HA PT ER NI NE272
may survive many threats for long periods of time. New threats may,
however, unravel systems that have survived for multiple generations.5
Major migration (out of or into an area) is always a threat that may or
may not be countered effectively (Baker 2005). Out-migration may simply
change the economic viability of a regime due to loss of those who
contribute needed resources. In-migration may bring new participants
who do not trust others and do not rapidly learn social norms that have
been established over a long period of time. Since collective action is
largely based on mutual trust and reciprocity, some self-organized resource
regimes that are in areas of rapid settlement have disintegrated
within relatively short times (Baland and Platteau 1996).
Even institutions that are characterized by the design principles fail.
Thus, we need to speculate about other threats to community governance
that arise from observations in the field, theoretical conjectures, and empirical
findings of scholars studying small-scale resource governance systems.
Here is a list of five threats to sustainable community governance
of small-scale resource governance systems that I have come across in
different contexts:
1. rapid exogenous changes;
2. transmission failures from one generation to the next of the operational
principles on which community governance is based;
3. programs relying on blueprint thinking and easy access to external funds;
4. corruption and other forms of opportunistic behavior; and
5. lack of large-scale institutional arrangements related to reliable information
collection, aggregation, and dissemination; fair and low cost conflictresolution
mechanisms; educational and extension facilities; and facilities for
helping when natural disasters or other major problems occur at a local level.
Let us briefly discuss each of these.
Rapid Exogenous Changes
All rapid changes in technology, in human, animal, or plant populations;
in factor availability; in substitution of relative importance of monetary
transactions in the national governance system; or in the heterogeneity of
participants are a threat to the continuance of any self-organized system,
whether it is a firm in a competitive market or a community-governed
resource. Individuals who have adapted an effective way of coping with
a particular technological, economic, or social environment may be able
to adjust to slow changes in one or several variables if substantial feedback
is provided about the consequences of these changes for the longterm
sustainability of the resource and/or the set of institutions used for
governing that resource (Gupta and Tiwari 2002). They may even be able
to adjust to changes in these variables that occur at a moderate rate. The
  
R OB US T R ES OU RC E G OV ER NA NC E 273
faster that key variables change and the more variables that change at
the same time, the more demanding is the problem of adaptation to new
circumstances. These kinds of threats are difficult for all organizations.
Those that rely to a greater extent on quasi-voluntary compliance are,
however, more threatened than those who are able to coerce contributions
(Bromley and Chapagain 1984; Goodland, Ledec, and Webb 1989).
Ottar Brox (1990) provides a vivid illustration of what happened in
the northern regions of Norway when technology, population density,
and other factors changed rapidly. As he points out, traditional northern
Norwegian fisheries were seasonal fisheries. “Large oceanic fish populations
migrate during phases in their life or yearly cycles, and occur within
reach of coastal fishermen only during short seasons” (231). Using traditional
harvesting techniques, “coastal fishermen did not have the boats,
gear and preservational techniques necessary to follow the fish populations
continually” (231). This had the consequence that it was almost
impossible to destroy the fishery.
Nor were the part-time farmers and part-time fishers able to reap most
of the resource rent from fishing until the Norwegian Raw Fish Act of
1938, which empowered fishers with the right to negotiate legally enforceable
landing prices. Fishers, who for many centuries could not themselves
reap the rents from a migratory fishery, now could do so, and could do
so in an era of fast-changing technology making it possible to capture and
store ever-greater quantities of fish. Further, other fishermen from other
countries after the Second World War had the technology and capital to
substantially increase effort dramatically above that which could be devoted
prior to this era. A fishery that had survived, and even flourished,
during many centuries of part-time fishing rapidly changed to a threatened
resource without adequate institutional means to respond to the new
incentives facing the fishers.
Transmission Failures
Rapid change of population or culture may lead to a circumstance in
which the general principles involved in the design of effective communitygoverned
institutions are not transmitted from one generation to another.
When individuals substitute rote reliance on formal rules for an understanding
of why particular formal rules are used, they can argue for
interpretations of the formal rules that undercut the viability of community
organization. For example, the charter or constitution of a community
organization may specify that simple majority rule will be used in making
decisions about future projects and how the costs and benefits of these
projects will be allocated. If the founders of such an organization recognize
the importance of gaining general agreement, they will rarely push
forward on a large project supported only by a minimal winning
  
C HA PT ER NI NE274
coalition. When there is a bare majority, almost as many community
members oppose a project as those who support it. If over time, however,
the principle of gaining general agreement to future projects is not conveyed
and accepted by those who accept leadership responsibilities, then
decisions barely receiving a majority may be pushed forward. Leaders
of communities who rely for too many decisions on minimal winning
coalitions may find themselves having to use patronage, coercion, and/
or corruption, rather than a foundation of general agreement, to keep
themselves in power.
Similarly, if participants view their own rules as obstacles to be overcome,
rather than as the written representation of general underlying
principles of organization, they may push for interpretations of rules
that lead to their general weakening. If each household tries to find every
legal way to minimize the amount of labor contributed to the maintenance
of a farmer-governed irrigation system, for example, eventually the cumulative
effect is an insufficient maintenance effort and the unraveling of
the contingent contributions of all. If one family tries to make a favorable
interpretation of how much labor they should contribute, given the
land they own, others come to know that this family is interpreting rules
in a manner that is highly favorable to them. Others who would be favored
by such an interpretation begin to use it as well. The total quantity
of labor contributed declines. Unless there is a community discussion
about the underlying principles that can be used in interpreting rules,
practices may evolve that cannot be sustained over time. The danger exists
that the unraveling continues unabated until the community organization
falls apart.
If one family tries to make a favorable interpretation of how much
labor they should contribute, given the land they own, others come to
know that this family is interpreting rules in a manner that is highly favorable
to them. Others, who would be favored by such an interpretation,
begin to use it as well. The total quantity of labor contributed declines.
Unless there is a community discussion about the underlying principles
that can be used in interpreting rules, practices may evolve that cannot be
sustained over time. Then, the danger exists that the unraveling continues
unabated until the community organization falls apart.
Programs Relying on Blueprint Thinking and Easy Access
to External Funds
Blueprint thinking occurs whenever policymakers, donors, citizens, or
scholars propose uniform solutions to a wide variety of problems that are
clustered under a single name based on one or more successful exemplars.
David Korten (1980) called this the “blueprint approach” and made a
  
R OB US T R ES OU RC E G OV ER NA NC E 275
devastating critique of its prevalence in development work at the end of
the 1970s. Projects or programs rely on some formula—the design of
another project, imposition of a particular voting rule in all settings, or
the way the project is initiated—rather than learning the specifics of a
particular setting and enabling participants to experiment and learn from
their own experience and that of others.
Even with all of the lessons learned in the last three decades about the
dangers of blueprint thinking, the temptation to fall into this trap continues
unabated. Fabia´n Repetto (2002), for example, describes an ambitious,
but failed, antipoverty program adopted in Argentina, the Plan Solidaridad,
which was an imitation of the supposedly successful antipoverty
program, Progresa, implemented in Mexico. Roconi (2002) describes another
Argentine program that was modeled on a blueprint, Plan Trabajar,
which also ended up stymied by clientelistic networks that resorted to
rent seeking and other mutually beneficial actions that undermined the
program. In searching for the “holy grail,” efforts to design homegrown
solutions to unique ecological conditions are stymied while policymakers
switch policies rapidly trying to copy whatever is considered the latest
and best (Mukand and Rodrik 2002; Acuna and Tommasi 2000). Pritchett
and Woolcock (2003) bemoan the problem of trying to find solutions
when “the” problem is actually the blueprint solution recommended
by donors and national governments for solving a problem.
Tragically, advocates of community governance have sometimes fallen
into this trap. A major program at the World Bank, Community Driven
Development (CDD), sounds as if it should support effective local development.
A requirement of this program is that the initial proposal for a
World Bank project must come from local officials or communities. Many
of the projects that are called “community driven,” however, turn out to
be quick investments in infrastructure such as local schools or roads. They
are indeed recommended by local officials. What school principal would
not actively lobby for a new school building once the possibility of gaining
World Bank funds is announced? The principal only needs to make a good
plea for the importance of a new school without any requirement for
financing the repayment of the loan. One evaluation of such CDD projects
found few other improvements beyond the infrastructure (e.g., few books
in the new schools and little impact on children’s educational achievement)
(World Bank 2002). Negative evaluations have not, however, had
much impact on the fervent advocates of CDD (Platteau 2004; Mansuri
and Rao 2003). CDD projects at the Bank have increased from an annual
expenditure of $325 million in 1996 to over $2 billion in 2003 (Mansuri
and Rao 2003) based more on enthusiasm than on objective evaluations
(Conning and Kevane 2002; Tendler 2000). Major risks of elite capture
and fraud exist in such programs (Platteau and Gaspart 2003).
  
C HA PT ER NI NE276
The availability of funds from donors or from national governmental
budgets that make no requirements for contributions from recipients can
also undermine local institutions.6
This is particularly salient in regard to
local infrastructure.7
Monetary resources for constructing, operating, and
maintaining infrastructure is frequently contributed by the taxpayers of
the nation in which the infrastructure is located or the taxpayers of those
nations providing economic assistance funds. When these external funds
are used, the financial connection between supply and use is nonexistent.
Whether the resources so mobilized are directly invested in the construction
and operation of the infrastructure or are diverted for individual
use by politicians or contractors depends on the professionalism of those
involved and on active efforts to monitor and sanction diversions of resources
and on the incentives built into the disbursement rules used
(Platteau and Gaspart 2003). Consequently, a considerable portion of the
mobilized resources is diverted to purposes other than those for which it
was intended.
Further, the design of projects is oriented more toward capturing the
approval of those who fund new construction than toward providing systems
that solve the problems facing present and future users. To convince
politicians that large chunks of a national budget should be devoted to
the construction of local infrastructure, planners attempt to design projects
that are “politically attractive.” This means that politicians who support
such expenditures can claim that the voters’ funds are being used to
invest in projects that will greatly expand the amount of food available
and lower the cost of living. Development projects need to have considerable
local involvement to be successful (Shivakumar 2005).
To convince external funding agencies that major infrastructure projects
should be funded through loans or grants, the evaluative criteria used
by these agencies in selecting projects has to play a prominent role in the
design of projects (Gibson, Andersson, et al. 2005). Projects designed by
engineers, who lack on-the-ground experience or training as institutional
analysts, are frequently oriented toward winning political support or international
funding. This orientation does not lead to the construction of
projects that serve most users effectively or encourage the investment of
users in their long-term sustenance. Inefficiencies occur at almost every
stage. At the same time, this inefficient process leads to the construction of
projects that generate substantial profits for large landholders and strong
political support for a government.
Processes that encourage looking to external sources of funding make
it difficult to build upon indigenous knowledge and institutions (Haller
2001, 2002). A central part of the message asking for external funds is
that what has been accomplished locally has failed and massive external
technical knowledge and funds are needed to achieve “development.” In
  
R OB US T R ES OU RC E G OV ER NA NC E 277
some cases, no recognition is made at all of prior institutional arrangements.
This has three adverse consequences: (1) property rights that resource
users had slowly achieved under earlier regimes are swept away
and the poor lose substantial assets; (2) those who have lost prior investments
are less willing to venture further investments; and (3) there is a
general downgrading of the status of indigenous knowledge and institutions.
In light of their own analysis of a failed effort to use external funds
to create an effective community forestry project, Morrow and Hull
(1996) provide a good summary of the problems resulting from externally
driven funding and priorities: “This case, along with the experience of
other community forestry enterprises in Latin America, suggests that
donor-driven projects often fail to analyze in sufficient depth the factors
outlined by the design principles, particularly the issues of institutional
and technological appropriateness and the impact of the larger political
economy” (1655).
Corruption and Rent-Seeking
All types of opportunistic behavior are encouraged, rather than discouraged,
by the availability of massive funds to subsidize the construction
and operation of large-scale infrastructure projects (Gibson, Andersson,
et al., 2005). Corrupt exchanges between officials and private contractors
are a notorious and widespread form of opportunism; corrupt payments
by citizens to government officials are less publicized, but probably no
less widespread. Free-riding on the part of those receiving benefits and
the lack of trust between citizens and officials, as well as among citizens,
are also endemic. Further, the potential rents that can be derived from
free electricity and free water by large-scale landowners stimulate efforts
to influence public decision making as to where projects should be located
and how they should be financed. Politicians, for their part, win political
support by strategic decisions concerning who will receive or continue to
receive artificially created economic rents.
Robert Bates (1987, 128) explains many of the characteristics of African
agricultural policies by arguing that major “inefficiencies persist because
they are politically useful; economic inefficiencies afford governments
means of retaining political power.” Part of Bates’s argument
relates to the artificial control exercised over the prices paid for agricultural
products, a topic that is not addressed in this study. The other part
of Bates’s argument relates to the artificial lowering of input prices.
When they lower the price of inputs, private sources furnish lesser quantities,
users demand greater quantities, and the result is excess demand. One consequence
is that the inputs acquire new value; the administratively created short-
  
C HA PT ER NI NE278
age creates an economic premium for those who acquire them. Another is that,
at the mandated price, the market cannot allocate the inputs; they are in short
supply. Rather than being allocated through a pricing system, they must be
rationed. Those in charge of the regulated market thereby acquire the capacity
to exercise discretion and to confer the resources upon those whose favor they
desire. (128)
Public programs that distribute farm credit, tractor-hire services, seeds,
and fertilizers, and which bestow access to government-managed irrigation
schemes and public land, thus become instruments of political organization
in the countryside of Africa (130).
There is an added dimension to rent seeking. The losses that the general
consumer and taxpayer accrue from rent-seeking activities are one dimension.
The second aspect of rent seeking in highly centralized economies is
the acquisition of resources needed to accumulate and retain political
power. All forms of opportunistic behavior, therefore, are exacerbated in
an environment in which an abundance of funds are available for the
construction of new and frequently large-scale infrastructure projects that
provide subsidized electricity, local roads, schools, and water.
Lack of Large-Scale Supportive Institutions
While smaller-scale, community-governed resource institutions may be
more effective than centralized government in achieving many aspects of
sustainable development, the absence of supportive, large-scale institutional
arrangements may be just as much a threat to long-term sustenance
as the presence of preemptive large-scale governmental agencies. Obtaining
reliable information about the effects of different uses of resource
systems and resource conditions is an activity that is essential to longterm
sustainability. If all local communities were to have to develop all of
their own scientific information about the physical settings in which they
were located, few would have the resources to accomplish this.
Let me use the example of the important role that the U.S. Geological
Survey has played in the development of more effective local groundwater
institutions in some parts of the United States. What is important to stress
is that the Geological Survey does not construct engineering works or do
anything other than obtain and disseminate accurate information about
hydrologic and geologic structures within the United States. When a local
set of water users wants to obtain better information about a local
groundwater basin, they can contract with the Geological Survey to conduct
an intensive study in their area. Water producers would pay a portion
of the cost of such a survey. The Geological Survey would pay the other
portion. The information contained in such a survey is then public infor-
  
R OB US T R ES OU RC E G OV ER NA NC E 279
mation available to all interested parties. The Geological Survey employs
a highly professional staff who rely on the most recent scientific techniques
for determining the structure and condition of groundwater basins.
Local water producers obtain the very best available information
from an agency that is not trying to push any particular future project
that the agency is interested in conducting. Many countries, such as India,
that do have large and sometimes dominating state agencies do not have
agencies that provide public access to high-quality information about resource
conditions and consequences. Recent efforts to open up groundwater
exploration in India may lead to the massive destruction of groundwater
basins rather than a firm basis for long-term growth.
Similarly, the lack of a low-cost, fair method for resolving those conflicts
that spill out beyond the bounds of a local community is also a threat
to long-run sustainability. All groups face internal conflicts or intergroup
conflicts that can destroy the fundamental trust and reciprocity on which
so much effective governance is based. If the only kind of conflict-resolution
mechanisms available are either so costly or so biased that most selfgoverned
common-pool resources cannot make use of them, these conflicts
can themselves destroy even very robust institutional arrangements.
Modest Coping Methods for Dealing with Threats to Sustainability
No surefire mechanisms exist for addressing all of the above threats.
There are three modest methods that I would like to discuss here before
turning to the concept of polycentricity—which is the major coping
method discussed in the last section of this chapter. I do so because these
methods are not frequently mentioned as being important ways of increasing
the effectiveness of self-governed institutions. However, they frequently
have high payoffs. They are: (1) the creation of associations of
community-governed entities, (2) comparative institutional research that
provides a more effective knowledge base about design and operating
principles, and (3) the development of more effective high school and
college courses on local governance.
Creating Associations of Community-Governance Entities
Those who think local participation is important in the process of developing
sustainable resources and more effective governance of resources
are frequently committed to doing a good deal of “community organization.”
All too frequently, this type of organization is conceptualized as
fostering a large number of community groups at the same level. If community
organization is fostered by nongovernmental organizations
  
C HA PT ER NI NE280
(NGOs) who then provide staff assistance and some external resources,
the organizations may flourish as long as the NGOs remain interested,
but wither on the vine when the NGOs turn to other types of projects. A
technique that draws on our knowledge of how self-governed institutions
operate is helping to create associations of community organizations.
When community organizations are brought together in federations,
they can provide one another some of the backup that NGOs may provide
to single-layer community organizations. While no single communitygoverned
organization may be able to fund information collection that
is unbiased and of real value to the organization, a federation of such
organizations may be able to amass the funds to do so. Simply having a
newsletter that shares information about what has worked and why it
has worked in some settings helps others learn from each other’s trialand-error
methods. Having an annual meeting that brings people together
to discuss their common problems and ways of tackling them greatly expands
that repertoire of techniques for coping with threats that any one
group can muster on its own.8
Rigorous Institutional Research
In addition to the type of exchange of information that those involved in
self-governing entities can undertake on their own, it is important to find
ways of undertaking rigorous, comparative research that controls for the
many confounding variables that simultaneously affect performance
(Hayes 2004; Gibson, Williams, and Ostrom 2005). In the field of medicine,
folk medicine has frequently been based on unknown foundations
that turned out to be relatively sound, but some folk medicine continued
for centuries, doing more harm to patients than good. The commons that
are governed by users and the institutions they use are complex and sometimes
difficult to understand. It is important to blend knowledge and information
obtained in many different ways as we try to build a more effective
knowledge base about what works and why. The recent study of Theesfeld
(2004) is an outstanding example of such a blend. It is a rigorous study
drawing on theory, in-depth fieldwork, and quantitative survey research
to understand the constraints facing Bulgarian farmers in their efforts to
engage in collective action in the Bulgarian transitional economy.
Developing Better Curricula on Local Governance
Textbooks on governance used to focus as much on local as they did on
national governance arrangements. During the past half-century, introductory
textbooks on American government have moved from a fifty-fifty
split between national and local government, to a ninety-five to five split.
  
R OB US T R ES OU RC E G OV ER NA NC E 281
The textbooks used in the West have strongly influenced the textbooks
used in developing countries. Consequently, many public officials learn
nothing in high school and college about how local communities can govern
themselves effectively or about the threats to local self-governance.
Instead, a presumption is made that governance is what is done in national
capitals and what goes on in villages is outmoded if not completely
useless.9
The Advantage and Limits of Polycentric Systems in Coping with
Design and Long-Term Sustainability of Systems
The last major task to be undertaken in this chapter is to discuss why
autonomous, self-organized resource governance systems may be more
effective in learning from experimentation than a single central authority.
I will first discuss the advantages and limits of a fully decentralized system
where all responsibility for making decisions related to smaller-scale
common-pool resources is localized. Then, I will discuss why a polycentric
governance system involving higher levels of government as well as
local systems is better able to cope more effectively with tragedies of the
commons (V. Ostrom 1999).
Among the advantages of authorizing the users of smaller-scale
common-pool resources to adopt policies regulating the use of these resources
are:
• Local knowledge. Appropriators who have lived and appropriated from a
resource system over a long period of time have developed relatively accurate
mental models of how the biophysical system itself operates, since the very
success of their appropriation efforts depends on such knowledge. They also
know others living in the area well and what norms of behavior are considered
appropriate.
• Inclusion of trustworthy participants. Appropriators can devise rules that
increase the probability that others are trustworthy and will use reciprocity.
This lowers the cost of relying entirely on formal sanctions and paying for
extensive guarding.
• Reliance on disaggregated knowledge. Feedback about how the resource system
responds to changes in actions of appropriators is provided in a disaggregated
way. Fishers are quite aware, for example, if the size and species distribution
of their catch is changing over time. Irrigators learn whether a
particular rotation system allows most farmers to grow the crops they most
prefer by examining the resulting productivity of specific fields.
• Better adapted rules. Given the above, appropriators are more likely to craft
rules over time that are better adapted to each of the local common-pool
resources than any general system of rules.
  
C HA PT ER NI NE282
• Lower enforcement costs. Since local appropriators have to bear the cost of
monitoring, they are apt to craft rules that make infractions highly obvious
so that monitoring costs are lower. Further, by creating rules that are seen as
legitimate, rule conformance will tend to be higher.
• Parallel autonomous systems. The probability of failure throughout a large
region is greatly reduced by the establishment of parallel systems of rule
making, interpretation, and enforcement.
There are, of course, limits to all ways of organizing the governance of
common-pool resources. Among the limits of a highly decentralized system
are:
• Some appropriators will not organize. While the evidence from the field is
that many local appropriators do invest considerable time and energy into
their own regulatory efforts, other groups of appropriators will not do so.
• Some self-organized efforts will fail. Given the complexity of the task involved
in designing rules, some groups will select combinations of rules that
generate failure instead of success. They may be unable to adapt rapidly
enough to avoid the collapse of a resource system.
• Local tyrannies. Not all self-organized resource governance systems will be
organized democratically or rely on the input of most appropriators. Some
will be dominated by a local leader or a power elite who only change rules
that they think will advantage them still further. This problem is accentuated
in locations where the cost of exit is particularly high and reduced where
appropriators can exercise choice over submitting to a local regime or not.
• Stagnation. Where local ecological systems are characterized by considerable
variance, experimentation can produce severe and unexpected results leading
appropriators to cling to systems that have worked relatively well in the
past and to stop innovating.
• Inappropriate discrimination. The use of identity tags is frequently an essential
method for increasing the level of trust and rule conformance. Tags based
on ascribed characteristics that have nothing to do with their trustworthiness
can, however, be the basis of excluding some individuals from access to
sources of productive endeavor.
• Limited access to scientific information. While time and place information
may be extensively developed and used, local groups may not have access to
scientific knowledge concerning the type of resource system involved.
• Conflict among appropriators. Without access to an external set of conflictresolution
mechanisms, conflict within and across common-pool resource
systems can escalate and provoke physical violence. Two or more groups
may claim the same territory and may continue to make raids on one another
over a very long period of time.
• Inability to cope with larger-scale common-pool resources. Without access
to some larger-scale jurisdiction, local appropriators may have substantial
  
R OB US T R ES OU RC E G OV ER NA NC E 283
difficulties regulating only a part of a larger-scale common-pool resource.
They may not be able to exclude others who refused to abide by the rules
that a local group would prefer to use. Given this, local appropriators have
no incentives to restrict their own use and watch others take away all of the
valued resource units that they have not appropriated.
The Capabilities of Polycentric Systems in Coping with Tragedies
of the Commons
Many of the capabilities of a parallel adaptive system can be retained in
a polycentric governance system. By polycentric I mean a system where
citizens are able to organize not just one but multiple governing authorities
at differing scales (see V. Ostrom, Tiebout, and Warren 1961; V. Ostrom
1997, 1999). Each unit exercises considerable independence to
make and enforce rules within a circumscribed domain of authority for a
specified geographical area. In a polycentric system, some units are
general-purpose governments while others may be highly specialized. Selforganized
resource governance systems in such a system may be special
districts, private associations, or parts of a local government. These are
nested in several levels of general-purpose governments that also provide
civil, equity, as well as criminal courts.10
In a polycentric system, the users of each common-pool resource would
have some authority to make at least some of the rules related to how
that particular resource will be utilized. Thus, they would achieve most
of the advantages of utilizing local knowledge as well as the potential to
learn from others who are also engaged in a similar trial-and-error learning
process in parallel systems (Folke, Berkes, and Colding 1998). On the
other hand, problems associated with local tyrannies and inappropriate
discrimination can be addressed by larger, general-purpose governmental
units who are responsible for protecting the rights of all citizens and for
the oversight of appropriate exercises of authority within smaller units of
government. It is also possible to make a more effective blend of scientific
information with local knowledge where major universities and research
stations are located in larger units but have a responsibility to relate recent
scientific findings to multiple smaller units within their region. Because
polycentric systems have overlapping units, information about what has
worked well in one setting can be transmitted to others who may try it
out in their settings. Associations of local resource governance units can
be encouraged to speed up the exchange of information about relevant
local conditions and about policy experiments that have proved particularly
successful. And, when small systems fail, there are larger systems to
call upon—and vice versa.
  
C HA PT ER NI NE284
When there is only a single governing authority, policymakers have to
experiment simultaneously with all of the common-pool resources within
their jurisdiction with each policy change. And, once a major change has
been made and implemented, further changes will not be made rapidly.
The process of experimentation will usually be slow, and information
about results may be contradictory and difficult to interpret. Thus, an
experiment that is based on erroneous data about one key structural variable
or one false assumption about how actors will react can lead to a
very large disaster (see Wilson, Low, et al. 1999). In any design process
where there is substantial probability of error, having redundant teams of
designers has repeatedly been shown to have considerable advantage (see
Landau 1969, 1973; Bendor 1985). The important point is: If the systems
are relatively separable, allocating responsibility for experimenting with
rules will not avoid failure, but will drastically reduce the probability of
immense failures for an entire region.
While the theoretical argument in support of polycentric systems of
governance to cope with multitier ecological systems is strong, the resistance
to these forms of organization is also strong. As discussed in chapter
8, the very concept of organization is closely tied for many scholars to the
presence of a central director who has designed a system to operate in a
particular way. Consequently, the mechanisms used by polycentric systems
are not well understood in many cases. Polycentric systems are themselves
complex, adaptive systems without one central authority dominating
all of the others. Thus, no guarantee exists that such systems will
find the combination of rules at diverse levels that are optimal for any
particular environment. In fact, one should expect that all governance
systems will be operating at less than optimal levels given the immense
difficulty of fine-tuning any complex, multitiered system.
Alcorn and Toledo (1998) stress the complementary institutional systems
at the national level in Mexico, supportive of local communities, as
generating a more sustainable governance system than exists in similar
ecological conditions. In the United States, there are many examples of
dynamic polycentric resource governance systems where there is strong
evidence of high performance.
One example from the United States is the Maine lobster fishery. This
system is noteworthy because of the long-term, complementary roles
adopted by both local and state governance systems. Maine is organized
into riparian territories along most of the coast. Boundary rules and many
of the day-to-day fishing regulations are organized by harbor gangs
(Acheson 1988, 2003).
In order to go fishing at all, one must become a member of a “harbor gang,”
the group of fishermen who go lobstering from a single harbor. Once one has
  
R OB US T R ES OU RC E G OV ER NA NC E 285
gained admittance into such a group, one can only set traps in the traditional
territory of that particular harbor gang. Members of harbor gangs are expected
to obey the rules of their gang concerning fishing practices, which vary somewhat
from one part of the coast to another. In all areas a person who gains a
reputation for molesting others’ gear or for violating conservation laws will be
severely sanctioned. Incursions into the territory of one gang by fishers from
another are ordinarily punished by surreptitious destruction of lobster gear.
There is strong statistical evidence that the territorial system, which operates
to limit the number of fishers exploiting lobsters in each territory, helps to conserve
the lobster resource. (Acheson, Wilson, and Steneck 1998, 400)
At the same time, the state of Maine has long established formal laws
that protect the breeding stock and increase the likelihood that regeneration
rates will be high. “At present, the most important conservation laws
are minimum and maximum size measures, a prohibition against catching
lobsters with eggs, and a law to prohibit the taking of lobsters which once
had eggs and were marked—i.e. the ‘V-notch’ law” (Acheson, Wilson,
and Steneck 1998, 400). Neither the state nor any of the harbor gangs
has tried to limit the quantity of lobster captured. The state does not try
to limit the number of lobster fishers, since this is already done at a local
level. However, the state has been willing to intercede when issues exceed
the scope of control of local groups. In the late 1920s, for example, when
lobster stocks were at very low levels and many local areas appear to have
had substantial compliance problems, the state took a number of steps—
including threats to close the fishery—that supported informal local enforcement
efforts. By the late 1930s, compliance problems were largely
resolved and stocks had rebounded (although it can’t be shown that these
two results are related, just correlated).
Recently, in response to changes that were breaking down the informal
harbor gang system, the state has formalized the system by dividing the
state into zones with democratically elected councils. Each council has
been given authority over rules that have principally local impacts—trap
limits, days and times fished, and so forth. Interestingly, the formalization
of local zones was followed almost immediately by the creation of an
informal council of councils to address problems at a greater than local
scale (J. Wilson 1997).
The system of comanagement of the Pacific salmon fisheries in the state
of Washington is another noteworthy example of an evolving polycentric
system that appears to be working much better than an earlier system
that was dominated primarily by state and federal agencies (see Singleton
1998). The change in the system came as a result of a major court decision
in the mid-1970s.11
The court held that the twenty-one Indian tribes, who
had signed treaties more than a century before, had protected rights to
  
C HA PT ER NI NE286
50 percent of the fish that passed through the normal fishing areas of the
tribes. The decision has required the state to develop a “comanagement”
system that involves both the state of Washington and the twentyone
Indian tribes in diverse policy roles related to salmon. The salmon
fisheries are a large, transboundary resource utilized by major commercial
firms as well as by the Indian tribes. Having the state strongly involved
means that it is “safe” for any local group to agree to follow strong conservation
practices because they know that other local groups are also
involved in the same conservation practices. At the same time, the earlier
centrally regulated system had focused on the ocean fishery and spent
little time on the fresh-water habitats that are essential to maintain the
viability of salmon fisheries over the long term. Individual tribal authorities
have concentrated their attention on the specific stocks and how to
manage these better.
Polycentric systems can generate considerable conflict among the various
units at multiple levels due to their interdependence. Conflicts that
escalate from misunderstandings to ever more serious charges and countercharges
that turn to violence are certainly negative processes. Conflict
may, on the other hand, generate more information that is useful to participants
in their efforts to solve challenging problems. Ebbin (2002, 2004)
has traced the evolution of conflict in the comanaged salmon fisheries
along the coast of Washington both as a fishery biologist working with
several of the tribal organizations and as a researcher conducting extensive
interviews with participants at all levels. While the early conflict was
framed as a technical problem regarding the knowledge to be used in
managing the system, later conflict “focused on questions of equity and
conservation as well as authority and jurisdiction” (Ebbin 2004, 82). The
redefinition of conflict in the court system and in other arenas helped to
create new institutional mechanisms that “changed the rules of the game
and the processes in which new conflicts are addressed” (82). After some
experience with the new institutions, even the government officials recognized
that new information was being generated that initially led to more
conflict but eventually led to better management of the stock.12
Coping with potential tragedies of the commons is never easy and never
finished. Now that we know that those dependent on these resources are
not forever trapped in situations that will only get worse over time, we
need to recognize that governance is frequently an adaptive process involving
multiple actors at diverse levels. Such systems look terribly messy
and hard to understand. The scholars’ love of tidiness needs to be resisted.
Instead, we need to develop better theories of complex adaptive systems
focused on overcoming social dilemmas, particularly those that have
proved themselves able to utilize renewable natural resources sustainably
over time.
  
R OB US T R ES OU RC E G OV ER NA NC E 287
Conclusion
Contextual variables are essential for understanding the initial growth
and sustainability of collective action as well as the challenges that longsurviving,
self-organized regimes must try to overcome. Simply saying
that “context matters” is not, however, a satisfactory theoretical approach.
In particular, we need to address how context affects the recruitment
of individuals who may become conditional cooperators and willing
punishers into an ongoing system and the likelihood that the norms held
by these participants are adopted and strengthened by others in a relevant
population.
Trying to understand why so many diverse institutions are created in
the first place, and then the consequences that are engendered in diverse
ecological, social, and economic settings, is a big challenge. Considerable
empirical and theoretical progress has been made and, I hope, adequately
summarized in this volume.
Empirical and theoretical work in the future needs to ask how a large
array of contextual variables affects the processes of teaching and evoking
social norms, of informing participants about the behavior of others and
their adherence to social norms, and of rewarding those who use social
norms, such as reciprocity, trust, and fairness. We need to understand
how institutional, cultural, and biophysical contexts affect the types of
individuals who are recruited into and leave particular types of collectiveaction
situations, the kind of information that is made available about
past actions, and how individuals can themselves change structural variables
so as to enhance the probabilities of norm-using types being involved
and growing in strength over time.
Further developments along these lines are essential for the development
of public policies that enhance socially beneficial, cooperative behavior
based in part on social norms. It is possible that past policy initiatives
that attempted to solve collective-action problems primarily by
changing extrinsic payoff structures may have been misdirected. Imposing
sanctions and inducements can crowd out the formation of social norms
that can enhance cooperative behavior in their own way. Increasing the
authority of individuals to devise their own rules may well result in processes
that allow social norms to evolve and thereby increase the probability
of individuals actually solving collective-action problems. There is a
real role for legislators and government agency officials in solving problems
at all levels but they need to follow Mark Moore’s (1995, 20) advice
to become “explorers who, with others, seek to discover, define, and produce
public value.”
Norms of reciprocity and trust are necessary for the long-term sustenance
of self-governing regimes. Norms alone, however, are not sufficient
  
C HA PT ER NI NE288
to support individuals facing the temptations of social dilemmas. Rules
that are fair, effective, and legitimate are necessary complements to shared
norms for sustaining self-governing institutions over time. And, in turn,
self-organizing arrangements enable people to learn more about one another’s
needs and the ecology around them. Learning problem-solving
skills in a local context generates citizens with more general problemsolving
skills that enables them to reach out and more effectively examine
far-reaching problems that affect all peoples living on this earth.
  
Notes
Chapter One
Understanding the Diversity of Structured Human Interactions
1. The Workshop in Political Theory and Policy Analysis was established at
Indiana University in the 1973–74 academic year. A central interest of all Workshop
activities has always been understanding institutions at all scales. Early research
focused on polycentric systems in metropolitan areas drawing on Vincent
Ostrom’s earlier work with Charles Tiebout and Robert Warren (1961), “The
Organization of Government in Metropolitan Areas” (see E. Ostrom, Parks, and
Whitaker 1974; McGinnis 1999b). Studies conducted at a macrolevel (see Kaminski
1992; V. Ostrom 1997; Loveman 1993; Sawyer 1992) have focused more on
national-level constitutional and collective-choice decisions as these eventually
impinge on the day-to-day decisions of citizens and/or subjects. Studies conducted
at a microlevel (Firmin-Sellers 1996; Gibson 1999; Agrawal 1999; E. Ostrom,
Gardner, and Walker 1994; E. Ostrom 1990) have tended to focus more on
operational-level decisions as they are in turn affected by collective-choice and
constitutional-choice rules—some of which have been crafted at the local level.
2. For earlier discussions see Kiser and Ostrom 1982; E. Ostrom 1986; Oakerson
1992; Gardner and Ostrom 1991; E. Ostrom, Gardner, and Walker 1994;
Crawford and Ostrom 2000.
3. See Diermeier and Krehbiel (2003), who develop an institutional methodology
using these four steps in an iterative process.
4. Elements of the framework have been used in teaching both graduate and
undergraduate courses at Indiana University since the mid-1970s.
5. See Oakerson 1992; E. Ostrom 1986, 1999; and E. Ostrom, Gardner, and
Walker 1994, chap. 2.
6. See discussion in chapter 5 on doing fieldwork related to studying rules and
norms.
7. In an interesting paper, Ahn et al. (2004) examine the impact of an increase
in the number of referees in the National Hockey League. Other studies had found
that during the two seasons of a gradual transition from one to two referees,
where part of the games were played with one referee, players derived more penalty
minutes when two referees were on the ice. This is in contradiction to the
economic theory of crime that predicts a deterrence effect. Ahn et al. found that
over a period of seasons, a deterrence effect can be found at the team level since
the composition of the team was adjusted. At the team level, the amount of penalty
minutes dropped, and salaries of aggressive players, the “goons,” decreased relative
to the average.
8. In a sympathetic critique of E. Ostrom, Gardner, and Walker’s (1994) use
of the IAD framework, Jonathan Bendor (1995, 189) reflects on the confusion
that exists regarding the concept of a framework. “Hard core social scientists,
those believing that even frameworks must generate hypotheses if they are to have
  
N OT ES TO CH AP TE R T WO290
any scientific value, might therefore dismiss the IAD approach as constituting a
mere list of important variables.” Bendor then argues that it would be premature
to dismiss the IAD approach, but does worry about the problem of the number
of possible rule configurations that it generates.
9. Chapter 2 of Thomas Dye’s (1981) extensively used textbook, Understanding
Public Policy, is devoted to “models of politics.” Included among these “models”
are organization charts (22), a framework for analyzing policy processes (24),
group theory (26–28), elite theory (29–31), criteria for evaluating policy (31–35),
incrementalism (36–37), game theory (36–40), and systems theory (41–43). No
wonder students find it challenging to sort out the difference between frameworks,
theories, and models.
10. As I have indicated elsewhere, I will forever be grateful to Larry Kiser and
Sue Crawford for our long, long conversations, outlines, and rough drafts of chapters
that were part of these earlier efforts and to Roger Parks for the many memos
he has sent us to clarify key points and conversations we have had about these
central issues. This book would not have been possible without those extensive
efforts and many others by my wonderful Workshop colleagues.
Chapter Two
Zooming In and Linking Action Situations
1. When game theorists first formally describe the structure of a game, they
make a distinction between the normal form and the extensive form. An n-player
normal form game is usually specified as consisting of: (1) a set of players
i = 1, . . . , n; (2) a set Si of strategies for player i = 1, . . . , n. A strategy profile
for the game would be: s = (s1, . . . , sn), where si ∈ Si for i = 1, . . . , n; (3) function
πi : S → R for player i = 1, . . . , n, where S is the set of strategy profiles, so πi (s)
is player i’s payoff when strategy profile s is chosen (Gintis 2000b, 12). A normal
form game is usually represented in matrix format with the strategies represented
by the rows and columns and the payoffs as the cells.
2. There are also major debates over whether Deep Blue is simply a fast processor
or whether it represents a form of “real intelligence.”
3. This game was inspired by a game of STEAL proposed by Plott and Meyer
(1975).
4. Hamburger (1979) pointed out that participants in a two-person Prisoner’s
Dilemma are more likely to select cooperative strategies than those in a threeperson
PD Game.
5. Scharpf (1997) also examines two-level games.
6. Peter Hupe and Michael Hill (2004) develop a somewhat modified version
of these three levels that they call constitutive, directive, and operation. They use
these three levels to integrate many of the efforts in the policy science to lay out
the stages of the policy process.
7. The four levels presented here are broadly similar to the four levels of social
analysis presented by Oliver Williamson (2000). His top level is called the embedded
level, where informal custom and tradition are located. The next level in his
  
N OT ES TO CH AP TE R T HR EE 291
system is “the institutional environment,” where the formal rules establishing a
judiciary, executive, etc., are found in his L2 (our constitutional-choice level). The
governance level is his L3 (our collective-choice level), and resource allocation
and employment is his L4 (our operational level).
8. Ciriacy-Wantrup and Bishop (1975) also recommended three levels when
they described the nested multilevel aspect of rules within organizations. Their
three levels of decision making were: (1) the “operating level,” where decisions
about daily activities were made; (2) the “institutional level,” which regulates the
decisions made at an operational level; and (3) the “policy level,” which focuses
on the structure of an organization by creating a charter that has standing with
an external government.
9. It is important for the reader to recognize that not all constitutional rules are
embodied in a written national constitution. Of course, the constitutions of some
countries are not contained at all in a written constitution. Further, each house of
a bicameral legislature will have its own rules related to the decision rules to be
followed in making collective choices. Each private corporation will have a set of
constitutional rules specifying the rights and duties of corporate officials in making
policies for the firm. Every family will have made some basic rules for how future
family policies will be made. These will rarely be made in a formal arena.
10. The decisions reached by members of a legislature will not be stable if opponents
of these decisions are able (1) to win a majority of seats in future elections
on promises made to the electorate that they will reverse the earlier decision or
(2) to change the constitutions so as to make the decisions unconstitutional. The
stability of decisions in complex modern institutions is dependent not only upon
the preferences and procedures used to organize decision making in one arena,
but upon the entire nested set of arenas (Shepsle 1989).
11. This problem is particularly acute if one were to adopt the “institutionsas-equilibria”
approach described briefly in the first part of chapter 4. The complexity
of the statements that would be required to specify an equilibrium if one
took the institutions-as-equilibria approach are literally beyond comprehension.
12. As Hamilton wrote in The Federalist, no. 37: “All new laws though
penned with the greatest technical skills and passed on the fullest and most mature
deliberation, are considered as more or less obscure and equivocal, until their
meaning be liquidated and ascertained by a series of particular discussions and
adjudications.”
Chapter Three
Studying Action Situations in the Lab
1. In some versions of this game, the Trustee also receives an endowment.
2. Another major effort to replicate experiments across cultures is the research
conducted in fifteen small-scale and relatively isolated societies (see Henrich et al.
2004). In this case, the experiments included the Dictator and Ultimatum Games
(and, in some locations, Public Good Games) where the predictions from traditional
game theory using an assumption of selfish participants were very clear. No
  
N OT ES TO CH AP TE R F OU R292
cooperation should be seen in any of these experiments. The predictions, however,
were not supported in this major effort. As one of the scholars responsible for
organizing this study, Herbert Gintis (2004, 65) reports: “We found, first, that the
self-interested actor model is not supported in any society studied. Second, there
is considerably more behavioral variability across groups than had been found in
previous cross-cultural research. . . . Third, group-level differences in the degree
of cooperation in production and the extent of contact with market economies
explain a substantial portion of the behavioral variations across societies.”
3. Still further experiments with the basic Trust Game have been reported by
Bolle (1998); Fehr, Ga¨chter, and Kirchsteiger (1996); and Fehr and Ga¨chter (forthcoming).
Rieskamp and Gigerenzer (2003) report on a very interesting effort to
program the heuristics that individuals tend to use when faced with situations
involving the structure of the basic Trust Game. For an overview of this research
tradition, see E. Ostrom and Walker 2003 and the special issue “Trust and Trustworthiness”
edited by Bohnet and Croson 2004.
4. James Cox and colleagues at the University of Arizona have embarked on
an ambitious and rigorous program to sort out behavior motivated by norms such
as trust and reciprocity, as contrasted to other-regarding preferences. See Cox
2004; Cox and Deck in press; and Cox and Sadiraj 2004.
5. In the experiments reported herein, we used a return rate of 1 cent per unit.
6. Gu¨rerk, Irlenbusch, and Rockenbach (2004) have examined cooperation
levels in a public good experimental environment. They find cooperation reaches
maximum levels when subjects can themselves adopt and use a costly sanctioning
system. Not only is cooperation at a maximum; sanctioning costs are very low at
the end of thirty periods. See also Maier-Rigaud and Apesteguia 2003.
7. Casari and Plott (2003) also changed the instructions given to players so as
to make the differences between the three conditions very distinct. In general, the
broad pattern of behavior in the base condition was very similar to our own earlier
experiments.
8. A form of retribution, called a grim trigger strategy, was occasionally discussed
in communication rounds but never deployed. Grim trigger strategies are
those that cooperate until someone defects and then never cooperate again. They
have been posited to be one of the strategies that, if known to be followed by some
participants, will lead others to cooperate and thereby solve a social dilemma (see
E. Ostrom, Gardner, and Walker 1994, chap. 7).
Chapter Four
Animating Institutional Analysis
1. It is also necessary to adopt a “solution concept,” such as the core that is
used in cooperative game theory (Shepsle 1975) or the Nash equilibrium that is
used in noncooperative game theory (Calvert 1995). Both are used extensively in
institutional theory to generate predictions (see Diermeier and Krehbiel 2003 for
a comparison of these two solution concepts as used in institutional theories).
Skyrms (1997) makes a powerful critique of how equilibrium concepts are casually
used in the social sciences without positing a dynamic process that would lead
  
N OT ES TO CH AP TE R F OU R 293
to a particular equilibrium. Animating institution theories, however, comes prior
to choosing solution concepts.
2. Jerome Bruner (2004) provides an excellent overview of the diverse learning
theories—molecular associationism, molar configurationism, associative bonds,
conditioned reflex, gestalt theory, behavioralism, psycholoinquistics—of the
last century and a half that have loosely led to a better understanding of how
humans learn.
3. Mike McGinnis has pointed out to me that many of the situations that are
modeled as if they were simple situations are actually very complex when one
studies them in the field. Thus, the notion of a simple situation is one that we can
use in analyzing models of a situation, but rarely in analyzing field settings. I
deeply appreciate the exchange of memos and discussions I have had with Mike
regarding this issue.
4. Classical utility theory did not make this assumption. In fact, little effort
was devoted to where utility came from until researchers tried to test propositions
derived from theory and have had to make a specific assumption as to the direct,
monotonic linkage between external, objective payoffs and internal utility valuations
(see V. Smith 1982). It is well known that without the second assumption,
it is difficult to derive hypotheses about human behavior that can be empirically
tested. In fact, it is impossible to test the proposition that individuals maximize
utility without obtaining some objective measure of utility. A variety of recent
theoretical efforts have explored different valuation assumptions by linking utility
overtly to the distribution of objective payoffs to self and others (see Fehr and
Ga¨chter 1998, 2000a, for examples).
5. See Rabin 1998 and Kahneman and Tversky 2000 for recent reviews;
Kahneman, Knetsch, and Thaler 1990 for a description of the endowment anomaly;
and Grether and Plott 1979 for a discussion of the preference reversal anomaly.
Hodgson 2004b also makes a cogent argument that the model of opportunistic
behavior, which is an extension of the rational egoist to behavior within firms
(Williamson 1985), places far too much emphasis on a single motivation as the
primary source of contractual difficulties leading to the creation of firms.
6. In an ingenious experiment with monkeys as subjects, Sugrue, Corrado, and
Newsome (2004) examined how the history of past choices and rewards affects
the internal representation of the expected value of alternative potential moves.
By training monkeys in a dynamic foraging environment, their eye movements
and behavior provided a window into their subjective valuation. The researchers
provided strong support that their subjects were using cognitive mechanisms leading
them to match time spent in foraging at a particular site in proportion to the
abundance of resources available at that site.
7. Thanks to John Schiemann for further clarification of this point via an
e-mail communication.
8. In a thoughtful article on the development of shared norms in a community,
Gibbard (1990, 798) places considerable emphasis on the importance of ritual as
a vivid symbol. “A vivid symbol, after all, is a stimulus that commands attention
and generalizes of itself. . . . Why are ritual and symbols so pervasive in human
life? Perhaps because psychic mechanisms that respond to them stabilize cooperative
arrangements and guard an individual’s place in those arrangements.”
  
N OT ES TO CH AP TE R F OU R294
9. Recent research using brain-imaging technology (Hasson et al. 2004; Pessoa
2004) investigated whether the brains of different subjects, who watched a vivid
thirty-minute segment of the movie “The Good, the Bad, and the Ugly” had
similar activation patterns. They found a relatively substantial correlation between
the signals generated by one brain with those of another subject watching
the same segment. Hasson et al. (2004) found that a major component of brain
activation occurred when emotionally salient segments of the movie—e.g., scenes
containing gunshots and explosions—occurred. They also recorded strong responses
when subjects viewed faces. While Hasson et al. did find considerable
evidence that regions of individual brains “ticked together,” they also found that
substantial portions of cortex activation could not be predicted from another subject’s
responses.
10. Vernon Smith (2001, 21) has reflected that we all function in at least two
worlds: “one of personal exchange governed by self-policed norms of reciprocity
(positive and negative) in which there is much intentional cooperation generating
gains from such trade; another of impersonal exchange through markets governed
by constantly evolving cultural rules invented by no one person, in which unintentionally,
and without awareness, we cooperate by pursuing our own self-defined
interests. Each world is a complex self-ordering system to which we, our parents,
our parents’ parents, etc., have adapted and contributed to its evolution.”
11. Although the discussion in chapter 3 focused on the Trust Game and Commons
Dilemma, the large number of experiments on the Ultimatum Game and the
Dictator Game also challenge the predictions derived from using rational egoist
assumptions about strategies and outcomes (Camerer 2003; Cox 2004).
12. Recent research by Brandts, Saijo, and Schram (2004) conduct the same
experimental social dilemma (in this case, a linear, voluntary contribution, public
goods game) in Japan, the Netherlands, Spain, and the United States. They found
only minor differences in the level of cooperation in all four countries.
13. Cameron (1999), for example, conducted ultimatum experiments in Indonesia
and was able to use sums that amount to three months’ wages. In this extremely
tempting situation, she still found that 56 percent of the Proposers allocated
between 40 and 50 percent of this very substantial sum to the Responder.
14. See Rothstein 1998, 2005; Levi 1997a; Frohlich and Oppenheimer 1996.
Many scholars are also engaged in serious theoretical efforts to try to understand
how humans use norms (see, in particular, Falk, Fehr, and Fischbacher 2002;
McCabe and Smith 2003; Frey 1997a; Ben-Ner and Putterman 2000a, 2000b;
Casari and Plott 2003; Rothstein 2005).
15. Orbell and Dawes (1991) made a cogent theoretical argument that individuals
project their own normative preferences onto others. Glaeser et al. (2000)
observed a high correlation between those who were trustworthy in sequential
trust settings and those who were also more likely to be trusting. Esarey and Ahn
(2004) find that when no information is provided about the prior behavior of
subjects in the second mover’s position, that subjects who are themselves trustworthy
tend to trust others at a significantly higher rate. Brosig (2002) finds that
communication enables subjects to signal intentions accurately.
16. It is always a challenge to understand which heuristic is being used in a
particular setting. Jager and Janssen (2003) assume that cognitive costs affect this
  
N OT ES TO CH AP TE R F OU R 295
choice. In their analysis, an individual is likely to evaluate heuristics based on their
satisfaction with outcomes achieved and the certainty with which a heuristic leads
to satisfactory results.
17. The intrinsic cost or anguish that an individual suffers from failing to use
a norm, such as telling the truth or keeping a promise, is usually referred to as
guilt, if primarily self-inflicted, or as shame, when the knowledge of the failure is
well known by others (Posner and Rasmusen 1999). While guilt may be selfinflicted,
extensive research by psychologists have led to a recognition that it arises
out of interpersonal transitions and varies by interpersonal context (Baumeister,
Stillwell, and Heatherton 1994). The same action generates guilt in one situation
(e.g., the immediate family), but not in another (e.g., the workplace) (Millar and
Tesser 1988).
18. In chapter 5, where we define delta parameters in more detail, we make
a further distinction between those parameters that are invoked by internal
mechanisms (e.g., through guilt) and those that are invoked by external mechanisms
(e.g., gossip leading to shame, etc.). For simplicity, I do not make a distinction
here.
19. Gangs in urban areas can be thought of as a clique of participants who
obtain positive rewards from doing acts that others think are prohibited. The
positive internal reward is magnified by the increase in the esteem of other gang
members for the perpetrator of these acts.
20. Riolo, Cohen, and Axelrod (2001) have shown how individuals who engage
in image scoring (Nowak and Sigmund 1998) can contribute to others’ welfare
without any need for internal norms of reciprocity or even repeat encounters
with the same individual. What is required is that all individuals carry arbitrary
tags (Holland 1995), and that individuals at the beginning of an evolutionary
process tend to donate resources to others who are “similar” to them—meaning
they carry some or all of the same tags. They show that within a few hundred
generations, a cluster of cooperative individuals reaching up to 80 percent of the
population can emerge, be challenged by a new cluster who have inherited a different
set of tags, and then be replaced by the new cluster as it grows to about 80
percent of the population (442). Such models are a useful starting point for thinking
about competition and relative survival rates as among different strategies
over a long time period.
21. An excellent example of the internal accounts that individuals develop
without much self-conscious thought was conducted among a small group of sugarcane
cultivators in Ecuador (Price, in press). The tradition in this region is to
organize regular mingas where attendance is required and all are expected to work
toward achieving a collective outcome such as clearing the weeds out of a sugarcane
field with machetes. Since mingas were held once or twice a week, workers
could become familiar with the effort that others regularly expended. Price asked
a set of workers to rank the effort that others regularly expended and compared
these individual perceptions with data obtained from systematic measurement of
the same workers’ effort in six observed mingas. Price found a high correlation
between these perceptions of work effort and objective measures of effort.
22. The Journal of Economic Theory devoted a special issue in 2001 to the
evolution of preferences (vol. 97, no. 2). See also Orbell et al. 2004 for a rigorous
  
N OT ES TO CH AP TE R F IV E296
and imaginative use of simulation of evolutionary processes to address how cooperative
dispositions may have evolved. As Werner Gu¨th had kindly pointed out to
me in correspondence, preference evolution is only one of the possibilities open
to models that use an indirect evolutionary approach. One can make other aspects
of participant decision making endogenous as well, such as beliefs and timing of
decisions.
23. Eshel, Samuelson, and Shaked (1998) develop a learning model where a
population of Altruists who adopt a strategy of providing a local public good
interacts in a local circular neighborhood with a population of Egoists who
free ride. In this local interaction setting, Altruists’ strategies are imitated sufficiently
often in a Markovian learning process to become one of the absorbing
states. Altruists interacting with Egoists in a larger environment are not so likely
to survive.
24. This implies that in a game where players know only their own payoffs
and not the payoffs of others, that they are more likely to behave like rational
egoists. McCabe and Smith (2003) show that players tend to evolve toward the
predicted, subgame perfect outcomes in experiments where they have only private
information of their own payoffs and to cooperative outcomes when they have
information about payoffs and the moves made by other players (see also
McCabe, Rassenti, and Smith 1996).
25. Further, Kikuchi, Watanabe, and Yamagishi (1996) have found that those
who express a high degree of trust are able to predict others’ behavior more accurately
than those with low levels of trust.
26. To examine the frequency of nonrational egoist preferences, a group of 181
undergraduates were given a questionnaire containing a similar payoff structure
on the first day of classes at Indiana University in January 1999 (Ahn et al. 2003).
They were asked to rank their preferences. In this nondecision setting, 52 percent
reflected preferences that were not consistent with being rational egoists (27 percent
ranked the outcome [C,C] over [D,C] and 25 percent were indifferent).
27. See also Esarey and Ahn (2004), who report on a critical experiment designed
to test the capability of traditional game-theoretic predictions as contrasted
to indirect evolutionary theory predictions for a repeated, sequential PD game.
They found more cooperation than predicted by traditional game theory and that
subjects enter the experiment with relatively stable preferences for reciprocity or
being a rational egoist. The behavior of second movers was broadly consistent
with the predictions of indirect evolutionary theory.
Chapter Five
A Grammar of Institutions
1. The grammar of institutions and much of the analysis in this chapter first
appeared in print in a 1995 American Political Science Review article. This chapter
builds on and adds to our APSR article. In particular, we add several figures
that had to be cut from the original article due to space limitations and that existed
in an appendix form until they were reprinted in Crawford and Ostrom 2000. In
this chapter, we provide more information on how to use the grammar of institu-
  
N OT ES TO CH AP TE R F IV E 297
tions, including new discussions of how the grammar relates to the other institutional
analysis tools and frameworks described in this book. The grammar in this
chapter retains the same logic and basic components as the grammar outlined in
the 1995 article; however, we have refined the grammar a bit based on what we
have learned from scholars using the grammar and what we learned as we worked
through the grammar in the larger context of this book. We have tried in particular
to improve the clarity of the distinction between rules and norms. To use a software
analogy, this chapter is Institutional Grammar 2.0 with an expanded instruction
book written specifically for institutional analysts.
2. All of these statements use the notion of a “holder”—a part of the rule that
will be filled by concepts that are either defined in general everyday language, such
as age, drive a car, or being in a particular location, or in legal documents, such
as legislation, court decisions, administrative decisions. The fact that all of our
rules involved “holders” is a recognition that ATTRIBUTES, AIMs, and CONDITIONS
all refer to concepts that are either defined in everyday language or are
created by a generative rule. The three deontic operators are defined in a formal
language—deontic logic.
3. We can also compare our syntax with that of other scholars. The Dalhousie
logic of rules, for example, uses a similar syntax without an OR ELSE and requires
that prescriptive statements be recast to use the forbidden DEONTIC operator
(see Braybrooke 1996).
4. Our concept of ATTRIBUTES is the same as the wenn component of the
Dalhousie syntax (Braybrooke 1996).
5. The largest group to which a prescriptive statement could apply is the folk
component in the Dalhousie system (Braybrooke 1996).
6. John R. Commons ([1924] 1968) stressed the correlative nature of rights.
To state that someone has a right, someone must have a duty to observe that right.
The person with the right, then, is permitted to do something, while those with
the duty are forbidden or required to do something.
7. See V. Ostrom and E. Ostrom (1970) for further development of these ideas.
8. The deontic “may” is still meaningful in statements that do not contain an
OR ELSE. Consider a legislative body that shares an institutional statement like
the following: [All junior members] [P] [contest senior members] [in committee
hearings]. This is the equivalent to: [All junior members] [∼F] [contest senior members]
[in committee hearings]. This prescription implies a prescription on senior
members not to reprimand or castigate junior members who challenge them in
committee hearings. Of course, the existence of such a norm does not ensure that
senior members will follow it in all instances. However, there will be a shared
notion that a rebuke based on seniority alone is inappropriate or unacceptable. If
a senior member reprimands a junior member, then we might expect the junior
member to use the grant of permission to defend against the senior member’s
actions. It means something for the junior member to say “everyone here knows
that I am permitted to challenge senior members in committee hearings.”
9. Societies undergoing substantial “liberalization” could be thought of as developing
shared understandings that individuals who in earlier times had been
forbidden to take certain actions are now permitted to do so. When the new norm
is shared, individuals who still attempt to obstruct the previously restricted actors
  
N OT ES TO CH AP TE R F IV E298
now face a cost for breaking the norm. Thus, changes in norms over time will be
reflected both in the particular deontic assigned to an action and potentially to
whom a cost of breaking the norm is assigned.
10. Kreps et al. (1982, 247) do not assume that the basis for one actor playing
tit-for-tat is necessarily the acceptance of a norm. They simply assume that either
some players have available to them only a tit-for-tat strategy or that there is some
probability that one player’s payoffs are such that tit-for-tat is strongly dominant.
The latter condition would be the case if some players in a population have some
combination of delta parameters associated with playing tit-for-tat whose values
are high enough to make tit-for-tat the dominant strategy for these players. Game
theorists have frequently assumed that such players were somehow “irrational,”
but an institutional explanation would be that such players had accepted a norm
that the obligatory way to play a repeated Prisoner’s Dilemma game was to follow
a tit-for-tat strategy.
11. We are, of course, aware that all actions have consequences as pointed out
to us by many of our students—Ryan Adams, in particular. The difference that
the OR ELSE makes is that the consequences specified by a rule would NOT have
occurred without the rule being in place and being enforced. Thus, a specific rule
adds further consequences to those that would occur as a result of individuals
interacting in a situation without that specific rule.
12. Tsebelis (1989, 1991) argues that in a game with only mixed-strategy equilibria,
increasing the size of the OR ELSE does not reduce the level of rule infraction
but rather reduces the level of monitoring. Weissing and Ostrom (1991b)
have shown that Tsebelis’s results hold in many but not all cases.
13. Hans Albert (1986, 25) bemoaned the difficulty of making any universally
binding demarcation between norms and rules of law. He opted to call prescriptions
“law only where they meet certain minimum requirements such as the existence
of secondary rules that regulate the identification, modification and adjudication
of the various primary rules in a society” (see also Hart 1961).
14. Or, as we indicated above, the rule or norm backing a rule may reward the
sanctioner for taking positive actions rather than punishing the sanctioner for
shirking.
15. It would, of course, be possible to include discount rates in the analysis,
but we assume they are zero here to keep the focus on other questions and not
those related to the size of the discount rate. For a discussion of the importance
of discount rates in the analysis of cooperation see Axelrod 1981, 1986.
16. A mixed strategy is a probability distribution over the pure strategies for
a player. In a static game, one may view the mixed strategy as the probability of
choosing one of the other pure strategies. One can also interpret mixed strategies
as behavioral tendencies in a repeated context where the probability of choosing
a pure strategy, say C, is viewed as a cooperation rate.
17. In many farmer-governed irrigation systems, for example, farmers devise
simple rules that are easy to monitor themselves and do not employ any formal
guards. In others, where a guard is employed by the farmers, a frequent payment
to the guard is in a proportion of the yield obtained by the irrigators. Thus, the
guard participates in the increased productivity of a system that reduces the rate
  
N OT ES TO CH AP TE R S IX 299
of stealing water or free-riding on the provision of needed resource inputs to maintain
the system (see E. Ostrom 1992a).
18. The delta parameters for player 4 could be disaggregated into their internal
and external components for an analysis that wished to focus on questions that
distinguished between internal and external sources of normative constraints related
to the Sanctioner.
19. One could argue that when prescriptions are rules, individuals will place
higher values on the deltas than when the prescriptions are norms (see Braybrooke
1987). If one assumes that the presence of rules influences the internal deltas, that δoi
are higher in the Rules Game (figure 5.6) than in the Norms Game (figure 5.4), then
the rule would influence the structure of the game even when the Monitor fails.
20. In this game, player 3 always correctly detects whether defection has occurred
and player 4 only has the option of sanctioning players who have defected.
If players 1 and 2 cooperate, player 4 does not have a choice of whether or not
to sanction. This eliminates issues of false detection and corrupt sanctioners from
the current analysis but not from future efforts that assumed only a probability
of correct detection and honest sanctioning.
21. Offe and Wiesenthal 1980; Hirschman 1985; Etzioni 1988; Coleman
1988; Ellickson 1991; Elster 1989b; Knack 1992; Ude´hn 1993; Margolis 1991;
E. Ostrom 1990; V. Ostrom 1997; Mansbridge 1990, 1994; Schmid 2004.
22. See E. Ostrom 1986 for a discussion of the configurational aspect of rules.
23. This logic of the delta parameter is similar to Etzioni’s (1988) discussion
of deontology.
Chapter Six
Why Classify Generic Rules?
1. And, it turns out, for other evolving sciences as well. Grimm and Wissel
(1997) note a babbling equilibrium problem in ecology, for example, when they
identify 163 definitions of “stability” in the ecological literature.
2. The latter problem assumed a substantial urgency during the summer of
2001 as I was writing an early draft of this chapter. I had taken a new laptop
computer to our Canadian writing cabin on the Manitoulin Island and had not
used the new computer with the old printer that I have used there for years. When
I first started to print, out came gibberish that occasionally had a recognizable
word but was generally a mess. After many hours of trying to fix this myself, I took
it to the island’s “Computer Doc,” Roland Panamick. He had to work through a
dozen hypotheses as to what was the problem. The first was a mechanical problem—the
printer might have needed to be cleaned. This was eminently reasonable,
given that I had used it for some time without a cleaning. However, in this case,
cleaning was not the source of the problem. After that, he began working on the
various levels of software that drive a computer. It turned out that there was a
major conflict occurring in the Windows software commands for a printer and
the software needed to run my old-fashioned dot matrix printer. After digging
down layer after layer of software commands, he was finally able to make my
printer work. If he had not understood which software commands did what and
  
N OT ES TO CH AP TE R S EV EN300
how they fit (or did not fit) together, he could not have fixed this system for me.
I will always remember the several hours I spent in his shop, both for the terror
in my heart that he would not be able to fix it (meaning I would be stuck without
a printer for several more weeks) and for the lesson he gave me in the importance
of understanding the multiple languages that create the structure that I use all the
time and take pretty much for granted.
3. See, for example, the seven market rules that Gode and Sunder (1997) use
in their analysis “What Makes Markets Allocationally Efficient?” and the fifteen
types of rules that Libecap (1996, 44) examines to understand the evolution of
mining law in the American West.
4. Readers who would like to pursue the impact of the “lack of agreement” in
these earlier studies are invited to look at E. Ostrom 1986, where I do discuss the
impact of these rules in some depth. It has also been reprinted in McGinnis 2000,
chap. 3.
Chapter Seven
Classifying Rules
1. See Benda-Beckmann 2000 for critique of the simple dichotomy of public
versus private as an adequate conceptual foundation for comparative legal research.
Schmid (1999) makes a strong critique of the tendency to talk about government
versus markets.
2. Since strategies and norms also have AIMs, we could also use the same way
of classifying them. Given the focus of this volume, however, we are most interested
in understanding the prescriptions that create the structure of an action situation
and will not focus here on the way that individuals adopt normative values
(internal delta components) in relation to actions they feel that must, must not,
or may take or outcomes to which similar prescriptions have been attached.
3. The classes of rules proposed in this chapter are somewhat similar to the five
“dimensions” of rules that Koremenos, Lipson, and Snidal (2001, 763) propose as
being basis for the rational design of institutional institutions:
Membership rules (MEMBERSHIP)
Scope of issues covered (SCOPE)
Centralization of tasks (CENTRALIZATION)
Rules for controlling the institution (CONTROL)
Flexibility of arrangements (FLEXIBILITY)
The classification used in this chapter does have a distinct theoretical foundation—that
rules do directly, and indirectly, affect the elements of an action situation,
and thus we focus first on the seven rules that have this direct impact.
4. Anthony Giddens (1979, 65) expressed the configural nature of rules when
he stated: “There is not a singular relation between ‘an activity’ and ‘a rule,’ as is
sometimes suggested or implied by statements like ‘the rule governing the Queen’s
move’ in chess. Activities or practices are brought into being in the context of
overlapping and connected sets of rules, given coherence by their involvement in
the constitution of social systems in the movement of time.”
  
N OT ES TO CH AP TE R S EV EN 301
5. Anthony Giddens (1979, 117) defined a position in a similar manner, as “a
social identity that carries with it a certain range . . . of prerogatives and obligations
that an actor who is accorded that identity (or is an ‘incumbent’ of that
position) may activate or carry out: these prerogatives and obligations constitute
the role-prescriptions associated with that position.”
6. When some persons can be forced into a decision situation, unscrupulous participants
may initiate formal actions in order to obtain illegal payoffs to let the person
withdraw. Whenever formal rules are written in an overly strict, unenforceable manner,
they can be used by street-level bureaucrats to force a shakedown. The New
York “bite” occurs when residents pay building inspectors not to enforce overly strict
building codes. Hobbes characterized bad laws as “traps for money.”
7. The four sets of position rules are similar to a set of rules used by Shepsle
(1979b) to define a committee system in a legislature. However, the form of the
statements differs. Shepsle’s formulations state the result of the operation of the
rule in a situation. Shepsle shows that equilibria can be structurally induced by a
particular combination of rules when the distribution of preferences and a simple
majority-rule institution could not lead to an equilibrium.
8. In earlier papers, this type of rule has been called “authority rule.” Since all
rules assign some form of authority to some participant in regard to one of the
components of an action situation, we thought it would be clearer to use the term
“choice” for this kind of rule.
9. Lack of agreement rules strongly affects the bargaining strength and strategies
adopted in labor management disputes. For an interesting empirical study of
the effect of three different rules specifying what actions can be taken by either
labor or management, see Dannin and Singh forthcoming.
10. President Clinton issued Executive Order No. 12866, which was retained
by President Bush. The order directs regulatory agencies to specify performance
objectives whenever possible in new regulations rather than requiring the acts that
must, or must not, be done.
11. The use of scope rules do date back, however, to one of the earliest-known
government codes—the Hammurabi Code (at <http://www.yale.edu/lawweb/
avalon/medieval/hammenu.htm/>). See Foliente 2000 for a discussion of this code,
as well as recent efforts to devise building codes relying more on scope rules.
12. This is a modified version originally presented in Gardner and Ostrom 1991.
13. Gode and Sunder (1997, 610–11) analyze the set of rules that they posit is
needed to transform a simple market into an efficient market even with “zero
intelligence traders.” They specify the following seven rules as needed to accomplish
this task:
1. Voluntary trading rule: traders are free to accept or reject offers.
2. Binding contract rule: bids and asks are binding, i.e., buyers must pay
what they bid; and sellers must sell at what they ask.
3. Price priority rule: higher bids dominate lower bids, and lower asks dominate
higher asks.
4. Accumulation rule: the highest bid (and the lowest ask if it is a double
auction) are picked only after all bids (and tasks) have been collected.
  
N OT ES TO CH AP TE R E IG HT302
5. Double auction rule: buyers can bid as well as sellers can ask.
6. Multiple rounds rule: multiple rounds of bids and asks are allowed; i.e.,
if the highest bid is less than the lowest ask, then there are further rounds of
bids and asks.
7. Public bids and asks with bid-ask improvement rule: a bid must be greater
than previous bids, and an ask must be less than previous asks.
One would also need a judge or some other enforcer to ensure these rules are
followed.
14. As we discuss in chapters 2 and 5, rules at one level are operational only
when these rules are backed by monitoring and sanctioning rules (or norms) at
another level. To undertake analysis at one level, one may assume that the rules
needed to change a situation at that level will be monitored and enforced. This
helps to simplify the analysis, but the assumption must be thoroughly examined
when doing applied work. All too many formal laws have been passed to change
behavior at an operational level without asking how the police and judges at that
level will themselves be monitored and sanctioned if they do not apply the rules
appropriately.
15. The extensive historical and analytical review of Greif, Milgrom, and Weingast
(1994) of the history of merchant guilds in the medieval era demonstrate that
local rulers frequently did renege on their promises and either let local merchants
steal the goods of an alien trader or confiscated their goods for themselves. Greif,
Milgrom, and Weingast argue that to obtain stable, long-term commitment from
rulers required that the merchants themselves organize so as to be able to threaten
(and enforce the threat so it was credible) a boycott of all merchants from the
organizing city or region if rulers did not keep their pledges. Such an organization—a
merchant guild—could coordinate the actions of its own members so that
rulers who failed to keep their commitments would face serious expected costs.
“To permit an efficient expansion of trade in the medieval environment, there was
a need for an organization that would supplement the operation of a multilateral
reputation mechanism by coordinating the responses of a large fraction of the
merchants. Only when a coordinating organization exists can the multilateral reputation
mechanism potential overcome the commitment problem” (753). To implement
such a system would require still a further set of rules beyond those specified
above.
Chapter Eight
Using Rules as Tools to Cope with the Commons
1. McCay and Acheson 1987; Fortmann and Bruce 1988; Berkes 1989; Berkes
et al. 2001; V. Ostrom, Feeny, and Picht 1993; Netting 1993; Bromley et al. 1992;
National Research Council 2002.
2. On the other hand, one can use simplified models of a policy process as
useful benchmarks to examine deviations observed in the field (see Scharpf 1997).
One can then ask: What is it about a setting that leads officials and/or appropriators
to deviate a little or a lot from a presumed solution?
  
N OT ES TO CH AP TE R E IG HT 303
3. Boundary rules are also important in more urban and industrial settings. In
a recent study of the biophysical, social, and institutional factors associated with
the establishment of conjunctive surface and groundwater arrangements, Heikkila
(2004) randomly selected seventy groundwater basins in California. Her statistical
analysis demonstrated that “institutional settings devised around the boundaries
of the resource, which allow water providers to control the resource through
basin adjudication or special groundwater management districts, can facilitate
conjunctive water management” (111).
4. Faysse (2003) provides a good analysis of these types of rules on the efficiency
of the allocation process.
5. Lobe and Berkes (2004) provide a detailed study of the padu fishing system
on an island off the coast of the state of Kerala in India. Each local shrimp fishery
governance system creates a set of fishing locations. All of these systems “attempt
to redistribute the catch fairly among the fishers by rotating access to the fishing
location” (276). They rotate according to a lottery system run each year that assures
an equal opportunity to fish in the best locations.
6. The water-holding capacity of paddy fields in Chitwan is also considerably
better than the sloped terraces that characterize Tanahun and other districts in
the middle hills (Shivakoti personal communication).
7. As a field researcher, I have witnessed a wide variety of systems that change
their rule structure regularly during a year depending on the seasons. Many such
systems have been developed locally and without knowledge of what others have
done. When I visited several sites in Tarija, Bolivia (near Cochabamba), for example,
in the mid-1990s, the local officials proudly told me that they had developed
a complex system of land allocation that existed nowhere else on earth! They then
described their system of small plots of land allocated as private property during
the good agricultural season of the year. All of the land in the valley was then
converted to a common-property system during the dry period of each year. This
enabled all of the local farmers to graze animals on the stubble of their agricultural
fields without concern for who owned which plot. It was indeed an ingenious
system well tailored to the local environment. It was, however, not unique given
the long history of the “open field system” of Europe (see Dahlman 1980; de
Moor, Shaw-Taylor, and Warde 2002).
8. See, for example, Weimer and Vining 1992. Other, more critical approaches
taken to the study of policy include scholars’ focus on the deliberative processes
involved in efforts to solve problems. See, for example, Dryzek 2000; Elster 1998;
Habermas 1996; Rawls 1971; and Fisher 2003.
9. Plott and Porter (1996) undertook a similar analysis including experiments
examining diverse ways of pricing the use of space stations. They point out the
importance of using lab experiments as a method for avoiding some of the costs
of experimenting later in the world using only models as a basis for choosing
policies.
10. See McKean 1992, 2000; Wade 1994; Schlager 1990; Baland and Platteau
1996; Tang 1992; E. Ostrom 1990, 1992b; E. Ostrom, Gardner, and Walker 1994;
Meinzen-Dick, Raju, and Gulati 2002. In addition, see Agrawal 2002 for an excellent
discussion of the over thirty variables that one or more scholar has speculated
affect the likelihood of collective action being undertaken.
  
N OT ES TO CH AP TE R N IN E304
11. Theesfeld (2004) provides an analysis of these calculations in regard to
irrigation in a transitional economy—Bulgaria.
12. See Feeny, Hanna, and McEvoy 1996 for a discussion of these issues related
to the collapse of the California sardine industry.
13. See Libecap 1995 for a discussion of the efforts to use the courts to challenge
the validity of de facto governance of inshore fisheries in the U.S.; see also
P. Alexander (1982).
Chapter Nine
Robust Resource Governance in Polycentric Institutions
1. As Jared Diamond (2005) illustrates so dramatically, some ecological disasters
might have been avoided if those making major decisions were not isolated
from the impact of their own decisions.
2. On the other hand, this also means that corrupt leaders searching for rules
that will maximize the returns they obtain over time are not fully able to do a
complete analysis either. Firmin-Sellers (1996) provides a well-documented study
of the failed effort of a strong African leader to gain the benefits he predicted
would come his way when the constitution he strongly supported for the Gold
Coast was approved and put into operation.
3. Recently, considerable attention has also been devoted to the relevance of
these principles for the governance of digital resources (Mueller 2002; Schweik
and Semenov 2003; Schweik, Evans, and Grove forthcoming; Hess and Ostrom
2003; Van Wendel de Joode 2004; Less 2002).
4. Four levels of monitoring were recorded: never, occasionally, seasonally, or
year-round. We dichotomized this variable and recoded seasonal or year-round
monitoring as “regular” and never or sporadic monitoring as “sporadic.”
5. Lansing (1991) graphically describes how changes in agricultural practices
induced by external agencies supposedly to help farmers in Bali increase their
productivity almost destroyed irrigation systems that had flourished for centuries.
6. The problem of local units becoming dependent on external funding is not
limited to the funding provided by international aid agencies. Two decades ago,
Sieber (1981) reviewed some of the reverse effects created by domestic U.S. policy.
The supposed aim of Nixon’s “New Federalism” reform was to increase the autonomy
of local units and strengthen the overall federal system. A study by Hudson
(1980) revealed that the policy has an opposite effect in some cities such as
El Paso. “El Paso is now more dependent, politically and economically, on federal
grants than it was prior to the New Federalism and local autonomy is significantly
reduced” (Hudson 1980, 900, quoted in Sieber 1981, 186).
7. Douglas Vermillion (2001) provides a cogent story about this process. In
discussing the Small-Scale Irrigation Turnover Program in Indonesia, he heard
about an earlier effort to increase “farmer participation” in the project. A member
of the project staff
informed farmers in a small scheme located in a hilly area of West Sumatra that a
certain, but undisclosed amount of funds were available to make small repairs to their
  
N OT ES TO CH AP TE R N IN E 305
scheme before full management responsibility would be turned over to them. Farmers
were invited to make a list of priorities for repairs. This was seen as a form of farmer
participation. The farmers responded by generating a long wish list. High on the list
was a curious request for the government to raise the masonry embankment along a
300-meter reach of the upper main canal by about 25 cms. When asked why they requested
this, some farmer representatives answered that about 11 years before, the
government had first installed the masonry embankment. By now the calan had accumulated
about 20 cms of silt. They said that if the government raised the embankment
another 25 cms they would probably not have to do any desiltation for another 11
years! (187)
8. See Yoder 1991, Pradhan and Yoder 1989, and Water and Energy Commission
Secretariat 1990 for descriptions of a highly innovative and successful program
of assisting farmers to design their own institutional rules rather than imposing
a set of model bylaws on them.
9. For a refreshing and different view of the importance of village governance,
see Ayo 2002.
10. See Blomquist, Schlager, and Heikkila 2004 for an insightful study of polycentric
institutions in the American West. Wagner (in press) provides an excellent
overview of the study of polycentricity. Hong and Page (2004) provide strong
evidence of the superiority of having diverse problem solvers involved in making
decisions about complex systems. While their study focuses on individual decision
makers rather than on the diversity of decision-making units in a polycentric system,
the core point of their study is that diversity of skills and backgrounds enhance
decision-making performance. This provides another foundation for why
polycentric systems that draw on diversely organized units will outperform singlelayer
systems solving the same set of problems over time.
11. This decision was rendered by Judge Boldt of the Western District Court
in 1974 and upheld in 1979 by the Supreme Court (443 U.S. 658 1979).
12. For good analyses of other “co-management” systems on the Pacific Coast
of North America, see Pinkerton 1989 and Poffenberger and McGean 1996. Wilson,
Nielsen, and Degnbol (2003) provide an excellent collection of empirical
studies focusing on relative successes and failures of systems that are called “comanagement,”
in many parts of the world. Unfortunately, some such systems have
fallen into the trap of “blueprint” thinking described above and have lost the
vitality of evolving polycentric systems, while others have continued to adapt and
develop productive institutional structures.
  
  
References
Abernathy, Charles L., and Hilmy Sally. 2000. “Experiments of Some Government-Sponsored
Organizations of Irrigators in Niger and Burkina Faso, West
Africa.” Journal of Applied Irrigation Studies 35(2):177–205.
Acheson, James M. 1988. The Lobster Gangs of Maine. Hanover, NH: University
Press of New England.
. 2003. Capturing the Commons: Devising Institutions to Manage the
Maine Lobster Industry. Hanover, NH: University Press of New England.
Acheson, James M., James A. Wilson, and Robert S. Steneck. 1998. “Managing
Chaotic Fisheries.” In Linking Social and Ecological Systems: Management
Practices and Social Mechanisms for Building Resilience, ed. Fikret Berkes and
Carl Folke, 390–413. New York: Cambridge University Press.
Acuna, Carlos H., and Mariano Tommasi. 2000. “Some Reflections on the Institutional
Reforms Required for Latin America.” In Institutional Reforms, Growth
and Human Development in Latin America, 357–400. New Haven: Yale Center
for International and Area Studies.
Aggarwal, Vinod K., and Ce´dric Dupont. 1999. “Goods, Games, and Institutions.”
International Political Science Review 20(4):393–409.
Agrawal, Arun. 1994. “Rules, Rule Making, and Rule Breaking: Examining the
Fit between Rule Systems and Resource Use.” In Rules, Games, and CommonPool
Resources, ed. Elinor Ostrom, Roy Gardner, and James Walker, 267–82.
Ann Arbor: University of Michigan Press.
. 1999. Greener Pastures: Politics, Markets, and Community among a
Migrant Pastoral People. Durham, NC: Duke University Press.
. 2000. “Small Is Beautiful, but Is Larger Better? Forest-Management Institutions
in the Kumaon Himalaya, India.” In People and Forests: Communities,
Institutions, and Governance, ed. Clark Gibson, Margaret McKean, and Elinor
Ostrom, 57–85. Cambridge: MIT Press.
. 2002. “Common Resources and Institutional Sustainability.” In The
Drama of the Commons, National Research Council, Committee on the
Human Dimensions of Global Change, ed. Elinor Ostrom, Thomas Dietz, Nives
Dolsˇak, Paul C. Stern, Susan Stonich, and Elke Weber, 41–85. Washington, DC:
National Academy Press.
. 2005. Environmentality: Technologies of Government and the Making
of Subjects. Durham, NC: Duke University Press.
Agrawal, Arun, and Clark Gibson, eds. 2001. Communities and the Environment:
Ethnicity, Gender, and the State in Community-Based Conservation. New
Brunswick, NJ: Rutgers University Press.
Agrawal Arun, and Sanjeev Goyal. 2001. “Group Size and Collective Action:
Third-Party Monitoring in Common-Pool Resources.” Comparative Political
Studies 34(1):63–93.
  
R EF ER EN CE S308
Agrawal, Arun, and Gautam N. Yadama. 1997. “How Do Local Institutions
Mediate Market and Population Pressures on Resources? Forest Panchayats in
Kumaon, India.” Development and Change 28(3):435–65.
Ahn, T. K., Marco A. Janssen, Derek S. Reiners, and Jeffrey E. Stake. 2004.
“Learning to Play Nice: Strategy Evolution in the National Hockey League.”
Working Paper. Bloomington: Indiana University, Workshop in Political Theory
and Policy Analysis.
Ahn, T. K., Elinor Ostrom, David Schmidt, and James Walker. 2003. “Trust in
Two-Person Games: Game Structures and Linkages.” In Trust and Reciprocity:
Interdisciplinary Lessons from Experimental Research, ed. Elinor Ostrom and
James Walker, 323–51. New York: Russell Sage Foundation.
Ahn, T. K., Elinor Ostrom, and James Walker. 2003. “Heterogeneous Preferences
and Collective Action.” Public Choice 117(3–4) (December): 295–314.
Alba, Joseph W., and Howard Marmorstein. 1987. “The Effects of Frequency
Knowledge on Consumer Decision Making.” Journal of Consumer Research
14 (June): 14–25.
Albert, Hans. 1986. “Law as an Instrument of Rational Practice.” In Contract
and Organization: Legal Analysis in the Light of Economic and Social Theory,
ed. Terence Daintith and Gunther Teubner, 25–51. Berlin: Walter de Gruyter.
Alchian, Armen A. 1950. “Uncertainty, Evolution, and Economic Theory.” Journal
of Political Economy 58(3):211–21.
Alcorn, Janis B., and Victor Toledo. 1998. “Resilient Resource Management in
Mexico’s Forest Ecosystems: The Contribution of Property Rights.” In Linking
Social and Ecological Systems: Management Practices and Social Mechanisms
for Building Resilience, ed. Fikret Berkes and Carl Folke, 216–49. New York:
Cambridge University Press.
Aldrich, John. 1995. Why Parties? Chicago: University of Chicago Press.
Alexander, Christopher. 1964. Notes on the Synthesis of Form. Cambridge:
Harvard University Press.
Alexander, Paul. 1982. Sri Lankan Fishermen: Rural Capitalism and Peasant
Society. Canberra: Australian National University.
Allen, Barbara. 2005. Tocqueville, Covenant, and the Democratic Revolution:
Harmonizing Earth with Heaven. Lanham, MD: Lexington Books.
Allport, Floyd H. 1962. “A Structuronomic Conception of Behavior: Individual
and Collective.” Journal of Abnormal and Social Psychology 64:3–30.
Amaro de Matos, Joa˜o, and Pedro P. Barros. 2004. “Social Norms and the Paradox
of Elections’ Turnout.” Public Choice 121(1–2):239–55.
Anderies, John M., Marco A. Janssen, and Elinor Ostrom. 2004. “A Framework
to Analyze the Robustness of Social-Ecological Systems from an Institutional
Perspective.” Ecology and Society 9(1):18. <http://www.ecologyandsociety.org/
vol9/iss1/art18>
Anderson, John R. 1991. “Is Human Cognition Adaptive?” Behavioural and
Brain Sciences 14(3):471–517.
Anderson, Lee G. 1986. The Economics of Fisheries Management. Rev. ed.
Baltimore: Johns Hopkins University Press.
. 1995. “Privatizing Open Access Fisheries: Individual Transferable
Quotas.” In The Handbook of Environmental Economics, ed. Daniel W.
Bromley, 453–74. Oxford: Blackwell.
  
R EF ER EN CE S 309
Andersson, Krister P. 2002. “Can Decentralization Save Bolivia’s Forests? An
Institutional Analysis of Municipal Forest Governance.” Ph.D. diss., Indiana
University.
. 2004. “Who Talks with Whom? The Role of Repeated Interactions in
Decentralized Forest Governance.” World Development 32(2):233–49.
Andreoni, James. 1988. “Why Free Ride? Strategies and Learning in Public Goods
Experiments.” Journal of Public Economics 37 (December): 291–304.
. 1989. “Giving with Impure Altruism: Applications to Charity and
Ricardian Equivalence.” Journal of Political Economy 97:1447–58.
Appleyard, Roger. 1987. “Events and Acts: The Structure of their Effects.” Behavioral
Science 32:92–105.
Arnason, Ragnar, and Hannes H. Gissurarson, eds. 1999. Individual Transferable
Quotas in Theory and Practice. Reykjavik: University of Iceland Press.
Arnold, J. E. M. 1998. Managing Forests as Common Property. FAO Forestry
paper no. 136. Rome, Italy: Food and Agriculture Organization of the United
Nations.
Arnold, J. E. M., and J. Gabriel Campbell. 1986. “Collective Management of Hill
Forests in Nepal: The Community Forestry Development Project.” In Proceedings
of the Conference on Common Property Resource Management, National
Research Council, 425–54. Washington, DC: National Academy Press.
Arnold, J. E. M., and William C. Stewart. 1991. Common Property Resource
Management in India. Tropical Forestry papers no. 24. Oxford: Oxford
Forestry Institute.
Arrow, Kenneth J. 1951. Social Choice and Individual Values. 2d ed. New York:
Wiley.
. 1963. “Uncertainty and the Welfare Economics of Medical Care.” American
Economic Review 53:91–6.
. 1974. The Limits of Organization. New York: W. W. Norton.
Ascher, William. 1995. Communities and Sustainable Forestry in Developing
Countries. San Francisco: ICS Press.
Axelrod, Robert. 1981. “The Emergence of Cooperation among Egoists.” American
Political Science Review 75:306–18.
. 1984. The Evolution of Cooperation. New York: Basic Books.
. 1986. “An Evolutionary Approach to Norms.” American Political
Science Review 80:1095–111.
. 1997. The Complexity of Cooperation: Agent-Based Models of Competition
and Collaboration. Princeton: Princeton University Press.
Ayers, Ian, and John Braithwaite. 1992. Responsive Regulation: Transcending the
Regulation Debate. Oxford: Oxford University Press.
Ayo, S. Bamidele. 2002. Public Administration and the Conduct of Community
Affairs among the Yoruba in Nigeria. Oakland, CA: ICS Press.
Azar, Ofer H. 2004. “What Sustains Social Norms and How They Evolve?
The Case of Tipping.” Journal of Economic Behavior and Organization
54(1):49–64.
Bacharach, Michael, and Diego Gambetta. 2001. “Trust in Signs.” In Trust in
Society, ed. Karen S. Cook, 148–84. New York: Russell Sage Foundation.
Baker, Mark. 2005. The Kuhls of Kangra: Community Managed Irrigation in the
Western Himalaya. Seattle: University of Washington Press.
  
R EF ER EN CE S310
Baland, Jean-Marie, and Jean-Phillipe Platteau. 1996. Halting Degradation of
Natural Resources: Is There a Role for Rural Communities? Oxford: Clarendon
Press.
Banana, Abwoli Y., and William Gombya-Ssembajjwe. 2000. “Successful Forest
Management: The Importance of Security of Tenure and Rule Enforcement
in Ugandan Forests.” In People and Forests: Communities, Institutions, and
Governance, ed. Clark Gibson, Margaret McKean, and Elinor Ostrom, 87–98.
Cambridge: MIT Press.
Bardhan, Pranab K. 2000. “Irrigation and Cooperation: An Empirical Analysis
of Forty-eight Irrigation Communities in South India.” Economic Development
and Cultural Change 48:847–65.
Bardhan, Pranab K., and Jeff Dayton-Johnson. 2002. “Unequal Irrigators: Heterogeneity
and Commons Management in Large-Scale Multivariate Research.”
In The Drama of the Commons, National Research Council, Committee on
the Human Dimensions of Global Change, ed. Elinor Ostrom, Thomas Dietz,
Nives Dolsˇak, Paul Stern, Susan Stonich, and Elke Weber, 87–112. Washington,
DC: National Academy Press.
Barker, Randolph, E. Walter Coward, Jr., Gilbert Levine, and Leslie E. Small.
1984. Irrigation Development in Asia: Past Trends and Future Directions.
Ithaca: Cornell University Press.
Barkow, Jerome H., Leda Cosmides, and John Tooby, eds. 1992. The Adapted
Mind: Evolutionary Psychology and the Generation of Culture. Oxford:
Oxford University Press.
Barnard, Chester. 1938. The Functions of the Executive. Cambridge: Harvard
University Press.
Barry, Brian, and Russell Hardin. 1982. Rational Man and Irrational Society? An
Introduction and Sourcebook. Beverly Hills, CA: Sage.
Barwise, Jon, and John Perry. 1983. Situations and Attitudes. Cambridge: MIT
Press.
Bates, Robert H. 1987. Essays on the Political Economy of Rural Africa. Berkeley
and Los Angeles: University of California Press.
Bates, Robert H., Avner Greif, Margaret Levi, Jean-Laurent Rosenthal, and Barry
Weingast. 1998. Analytical Narratives. Princeton: Princeton University Press.
. 2000. “The Analytic Narrative Project.” American Political Science
Review 94 (September): 696–702.
Batistella, Mateus. 2001. “Landscape Change and Land-Use/Land-Cover Dynamics
in Rondoˆnia, Brazilian Amazon.” Ph.D. diss., Indiana University.
Baumeister, Roy F., Arlene M. Stillwell, and Todd F. Heatherton. 1994. “Guilt:
An Interpersonal Approach.” Psychological Bulletin 115(2):243–67.
Becker, C. Dustin. 2003. “Grassroots to Grassroots: Why Forest Preservation Was
Rapid at Loma Alta, Ecuador.” World Development 31(1):163–76.
Benda-Beckmann, Franz von. 1995. “Anthropological Approaches to Property
Law and Economics.” European Journal of Law and Economics 2:309–36.
. 1997. “Citizens, Strangers and Indigenous Peoples: Conceptual Politics
and Legal Pluralism.” Law and Anthropology 9:1–42.
  
R EF ER EN CE S 311
. 2000. “Relative Publics and Property Rights: A Cross-Cultural Perspective.”
In Property and Values: Alternatives to Public and Private Ownership,
ed. Charles Geisler and Gail Daneker, 151–74. Washington, DC: Island Press.
. 2001. “Between Free Riders and Free Raiders: Property Rights and Soil Degradation
in Context.” In Economic Policy and Sustainable Land Use, ed. Nico Heerink,
Herman van Keulen, and Marijke Kuiper, 293–316. New York: Physica-Verlag.
Bendor, Jonathan. 1985. Parallel Systems: Redundancy in Government. Berkeley
and Los Angeles: University of California Press.
. 1995. “Book Review of Rules, Games, and Common-Pool Resources.”
American Political Science Review 89(1) (March): 188–89.
Bendor, Jonathan, and Dilip Mookherjee. 1990. “Norms, Third-Party Sanctions,
and Cooperation.” Journal of Law, Economics, and Organization 6:33–63.
Benedict, Ruth. 1934. Patterns of Culture. Boston: Houghton Mifflin.
Ben-Ner, Avner, and Louis Putterman. 2000a. “On Some Implications of Evolutionary
Psychology for the Study of Preferences and Institutions.” Journal of
Economic Behavior and Organization 43(1):91–99.
. 2000b. “Values Matter.” World Economics 1(1):39–60.
Benz, Matthias, and Alois Stutzer. 2004. “Are Voters Better Informed When They
Have a Larger Say in Politics? Evidence for the European Union and Switzerland.”
Public Choice 119(1–2):31–59.
Berg, Joyce, John Dickhaut, and Kevin McCabe. 1995. “Trust, Reciprocity, and
Social History.” Games and Economic Behavior 10(1):122–42.
Berkes, Fikret. 1985. “Fishermen and ‘The Tragedy of the Commons.’ ” Environmental
Conservation 12:199–206.
. 1987. “Common Property Resource Management and Cree Indian Fisheries
in Subarctic Canada.” In The Question of the Commons: The Culture and
Ecology of Communal Resources, ed. Bonnie J. McCay and James Acheson,
66–91. Tucson: University of Arizona Press.
, ed. 1989. Common Property Resources: Ecology and Community-Based
Sustainable Development. London: Belhaven Press.
Berkes, Fikret, Robin Mahon, Patrick McConney, Richard Pollnac, and Robert
Pomeroy. 2001. Managing Small-Scale Fisheries: Alternative Directions and
Methods. Ottawa: International Development Research Centre.
Bicchieri, Cristina. 1997. “Learning to Cooperate.” In The Dynamics of Norms,
ed. Cristina Bicchieri, Richard Jeffrey, and Brian Skyrms, 17–46. Cambridge:
Cambridge University Press.
Bickers, Kenneth N., and John T. Williams. 2001. Public Policy Analysis: A Political
Economy Approach. Boston: Houghton Mifflin.
Black, Max. 1962. Models and Metaphors. Ithaca: Cornell University Press.
Blomquist, William. 1992. Dividing the Waters: Governing Groundwater in
Southern California. Oakland, CA: ICS Press.
. 1994. “Changing Rules, Changing Games: Evidence from Groundwater
Systems in Southern California.” In Rules, Games, and Common-Pool Resources,
ed. Elinor Ostrom, Roy Gardner, and James Walker, 283–300. Ann
Arbor: University of Michigan Press.
  
R EF ER EN CE S312
Blomquist, William, and Helen M. Ingram. 2003. “Boundaries Seen and Unseen:
Resolving Transboundary Groundwater Problems.” Water International
28(2):162–69.
Blomquist, William, Edella Schlager, and Tanya Heikkila. 2004. Common Waters,
Diverging Streams: Linking Institutions and Water Management in Arizona,
California, and Colorado. Washington, DC: Resources for the Future.
Bloom, Paul. 2000. How Children Learn the Meanings of Words. Cambridge:
MIT Press.
Bohnet, Iris, and Rachel Croson. 2004. “Introduction to Special Issue on Trust
and Trustworthiness.” Journal of Economic Behavior and Organization
55(4):443–45.
Bohnet, Iris, and Bruno S. Frey. 1999. “The Sound of Silence in Prisoner’s Dilemma
and Dictator Games.” Journal of Economic Behavior and Organization
38(1) (January): 43–58.
Bolle, Friedel. 1998. “Rewarding Trust: An Experimental Study.” Theory and
Decision 45(1):83–98.
Bolton, Gary E., and Axel Ockenfels. 2000. “ERC: A Theory of Equity, Reciprocity,
and Competition.” American Economic Review 90:166–93.
Bo¨rgers, Tilman, and Rajiv Sarin. 1997. “Learning through Reinforcement and
Replicator Dynamics.” Journal of Economic Theory 77:1–14.
Boulding, Kenneth E. 1963. “Towards a Pure Theory of Threat Systems.” American
Economic Review 53:424–34.
Bourdieu, Pierre. 1977. Outline of a Theory of Practice. New York: Cambridge
University Press.
Bowles, Samuel. 1998. “Endogenous Preferences: The Cultural Consequences of
Markets and Other Institutions.” Journal of Economic Literature 36 (March):
75–111.
Bowles, Samuel, and Herbert Gintis. 2004. “The Evolution of Strong Reciprocity:
Cooperation in Heterogeneous Populations.” Theoretical Population Biology
65(1):17–28.
Boyd, Robert, Herbert Gintis, Samuel Bowles, and Peter J. Richerson. 2003. “The
Evolution of Altruistic Punishment.” Proceedings of the National Academy of
Science 100:3531–35.
Boyd, Robert, and Peter J. Richerson. 1985. Culture and the Evolutionary Process.
Chicago: University of Chicago Press.
Brander, James A., and M. Scott Taylor. 1998. “The Simple Economics of Easter
Island: A Recardo-Malthus Model of Renewable Resource Use.” American
Economic Review 88(1):119–38.
Brandts, Jordi, Tatsuyoshi Saijo, and Arthur Schram. 2004. “How Universal Is
Behavior? A Four Country Comparison of Spite and Cooperation in Voluntary
Contribution Mechanisms.” Public Choice 119(3–4):381–424.
Braybrooke, David. 1987. Philosophy of Social Science. Englewood Cliffs, NJ:
Prentice Hall.
. 1996. “The Representation of Rules in Logic and Their Definition.”
In Social Rules: Origin; Character; Logic; Change, ed. David Braybrooke.
Boulder, CO: Westview Press.
  
R EF ER EN CE S 313
Brennan, Geoffrey, and James Buchanan. 1985. The Reason of Rules: Constitutional
Political Economy. New York: Cambridge University Press.
Bromley, Daniel W. 1982. Improving Irrigated Agriculture: Institutional Reform
and the Small Farmer. Working Paper no. 531. Washington, DC: World Bank.
Bromley, Daniel W., and Devendra P. Chapagain. 1984. “The Village against the
Center: Resource Depletion in South Asia.” American Journal of Agricultural
Economics 66:868–73.
Bromley, Daniel W., David Feeny, Margaret McKean, Pauline Peters, Jere Gilles,
Ronald Oakerson, C. Ford Runge, and James Thomson, eds. 1992. Making the
Commons Work: Theory, Practice, and Policy. San Francisco: ICS Press.
Brosig, Jeannette. 2002. “Identifying Cooperative Behavior.” Journal of Economic
Behavior and Organization 47(3):275–90.
Brown, William M., and Christopher Moore. 2002. “Smile Asymmetries and
Reputation as Reliable Indicators of Likelihood to Cooperate: An Evolutionary
Analysis.” In Advances in Psychology Research, ed. Serge Shohov, 59–78.
Huntington, NY: Nova Science Publishers.
Brox, Ottar. 1990. “The Common Property Theory: Epistemological Status and
Analytical Utility.” Human Organization 49(3):227–35.
Bruner, Aaron, Raymond Gullison, Richard Rice, and Gustavo da Fonseca. 2001.
“Effectiveness of Parks in Protecting Tropical Biodiversity.” Science
291(5501):125–29.
Bruner, Jerome. 2004. “The Psychology of Learning: A Short History.” Daedalus
133(1):13–20.
Buchan, Nancy R., Rachel T. Croson, and Eric J. Johnson. 1999. “Getting to
Know You: An International Experiment of the Influence of Culture, Communication,
and Social Distance on Trust and Reciprocation.” Wharton School
Working Paper no. 98–03–05. Philadelphia: University of Pennsylvania.
Buchanan, James M., and Alberto di Pierro. 1980. “Cognition, Choice, and Entrepreneurship.”
Southern Economic Journal 46(3):693–701.
Buchanan, James M., and Gordon Tullock. 1962. The Calculus of Consent. Ann
Arbor: University of Michigan Press.
Burns, James MacGregor. 1963. The Deadlock of Democracy: Four-Party Politics
in America. Englewood Cliffs, NJ: Prentice Hall.
Burns, Tom R., and Tom Dietz. 1991. “Institutional Dynamics: An Evolutionary
Perspective.” Paper presented at the Fifteenth World Congress of the International
Political Science Association, Buenos Aires, Argentina, July 21–25.
Burns, Tom R., and Helena Flam. 1987. The Shaping of Social Organization:
Social Rule System Theory with Applications. Beverly Hills, CA: Sage.
Burns, Tom R., and Anna Gomolin´ska. 1998. “Modelling Social Game Systems
by Rule Complexes.” In Rough Sets and Current Trends in Computing:
Proceedings of the First International Conference, Warsaw, Poland, ed. Lech
Polkowski and Andrzej Skowron, 581–84. Berlin/Heidelberg: Springer-Verlag.
Busemeyer, Jerome R., and In Jae Myung. 1992. “An Adaptive Approach to
Human Decision Making: Learning Theory, Decision Theory, and Human Performance.”
Journal of Experimental Psychology: General 121(2):177–94.
  
R EF ER EN CE S314
Bushouse, Brenda K. 1999. “The Mixed Economy of Child Care: An Institutional
Analysis of Nonprofit, For-Profit, and Public Enterprises.” Ph.D. diss., Indiana
University.
Cain, Michael. 1998. “An Experimental Investigation of Motives and Information
in the Prisoner’s Dilemma Game.” Advances in Group Processes
15:133–60.
Calvert, Randall L. 1992. “Rational Actors, Equilibrium, and Social Institutions.”
Paper presented at the Public Choice Meetings, New Orleans, LA, March.
. 1995. “The Rational Choice Theory of Social Institutions: Cooperation,
Coordination, and Communication.” In Modern Political Economy: Old Topics,
New Directions, ed. Jeffrey Banks and Eric Hanushek, 216–18. New York:
Cambridge University Press.
Camerer, Colin F. 2003. Behavioral Game Theory: Experiments in Strategic Interaction.
Princeton: Princeton University Press.
Cameron, Lisa A. 1999. “Raising the Stakes in the Ultimatum Game: Experimental
Evidence from Indonesia.” Economic Inquiry 37(1):47–59.
Campbell, Donald T. 1969. “Reforms as Experiments.” American Psychologist
24(4):409–29.
. 1975. “On the Conflicts between Biological and Social Evolution and
between Psychology and Moral Tradition.” American Psychologist
30(11):1103–26.
Cardenas, Juan-Camilo. 2000. “How Do Groups Solve Local Commons Dilemmas?
Lessons from Experimental Economics in the Field.” Environment, Development
and Sustainability 2:305–22.
. 2003. “Real Wealth and Experimental Cooperation: Evidence from Field
Experiments.” Journal of Development Economics 70:263–89.
Cardenas, Juan-Camilo, John K. Stranlund, and Cleve E. Willis. 2000. “Local
Environmental Control and Institutional Crowding-Out.” World Development
28(10):1719–33.
Carlson, Jean M., and John Doyle. 2002. “Complexity and Robustness.”
Proceedings of the National Academy of Sciences 9 (suppl. 1) (February 19):
2499–545.
Carr, Jered, and Richard Feiock. 2004. City-County Consolidation and Its
Alternatives: Reshaping the Local Governing Landscape. Armonk, NY:
M. E. Sharpe.
Carson, Richard T., Nadja Marinova, and David Zilberman. 1999. “Smoothing
the Waters: The Jordan Rift.” Policy Brief 11–1. San Diego: University of
California, Institute on Global Conflict and Cooperation.
Casari, Marco, and Charles R. Plott. 2003. “Decentralized Management of
Common Property Resources: Experiments with a Centuries-Old Institution.”
Journal of Economic Behavior and Organization 51:217–47.
Cashdan, Elizabeth. 1990. Risk and Uncertainty in Tribal and Peasant Economies.
Boulder, CO: Westview Press.
Cernea, Michael M. 1989. User Groups as Producers in Participatory Afforestation
Strategies. World Bank Discussion Papers no. 70. Washington, DC:
World Bank.
  
R EF ER EN CE S 315
Chakraborty, Rabindra N. 2004. “Sharing Rules and the Commons: Evidence
from Ha’apai, Tonga.” Environment and Development Economics 9:455–72.
Chamberlin, John. 1974. “Provision of Collective Goods as a Function of Group
Size.” American Political Science Review 68(2) (June): 707–16.
Chan, Kenneth S., Rob Godby, Stuart Mestelman, and R. Andrew Muller. 1997.
“Equity Theory and the Voluntary Provision of Public Goods.” Journal of Economic
Behavior and Organization 32:349–64.
Charness, Gary, and Matthew Rabin. 2003. “Understanding Social Preferences
with Simple Tests.” Quarterly Journal of Economics 117:817–69.
Choe, Jaesong. 1992. “The Organization of Urban Common-Property Institutions:
The Case of Apartment Communities in Seoul.” Ph.D. diss., Indiana
University.
Choe, Olivia S. 2004. “Appurtenancy Reconceptualized: Managing Water in an
Era of Scarcity.” Yale Law Journal 113 (June): 1909–53.
Ciriacy-Wantrup, Siegfried V., and Richard C. Bishop. 1975. “ ‘Common Property’
as a Concept in Natural Resource Policy.” Natural Resources Journal
15(4):713–27.
Clark, Andy, and Annette Karmiloff-Smith. 1991. “The Cognizer’s Innards: A
Psychological and Philosophical Perspective on the Development of Thought.”
Mind and Language 8(4) (Winter): 487–519.
Coase, Ronald H. 1937. “The Nature of the Firm.” Economica 4:386–405.
Coglianese, Cary, Jennifer Nash, and Todd Olmstead. 2003. “Performance-Based
Regulation: Prospects and Limitations in Health, Safety, and Environmental
Protection.” Administrative Law Review 55(4):705–29.
Coleman, James S. 1964. Introduction to Mathematical Sociology. New York:
Free Press.
. 1973. The Mathematics of Collective Action. Chicago: Aldine.
. 1987. “Norms as Social Capital.” In Economic Imperialism, ed. Gerard
Radnitzky and Peter Bernholz, 133–55. New York: Paragon.
. 1988. “Free Riders and Zealots: The Role of Social Networks.” Sociological
Theory 6 (Spring): 52–7.
Commons, John R. [1924] 1968. Legal Foundations of Capitalism. Madison: University
of Wisconsin Press.
Conning, Jonathan, and Michael Kevane. 2002. “Community Based Targeting
Mechanisms for Social Safety Nets: A Critical Review.” World Development
30(3):375–94.
Cook, Karen S., and Robin M. Cooper. 2003. “Experimental Studies of Cooperation,
Trust, and Social Exchange.” In Trust and Reciprocity: Interdisciplinary
Lessons from Experimental Research, ed. Elinor Ostrom and James Walker,
209–44. New York: Russell Sage Foundation.
Cooter, Robert D. 1994. “Structural Adjudication and the New Law Merchant:
A Model of Decentralized Law.” International Review of Law and Economics
14:215–31.
Cooter, Robert D., and Thomas S. Ulen. 1996. Law and Economics. 2d ed. Reading,
MA: Addison-Wesley.
Copes, Parzival. 1986. “A Critical Review of the Individual Quota as a Device in
Fisheries Management.” Land Economics 62(3):278–91.
  
R EF ER EN CE S316
Cordell, John C., and Margaret McKean. 1992. “Sea Tenure in Bahia, Brazil.” In
Making the Commons Work: Theory, Practice, and Policy, ed. Daniel W. Bromley
et al., 183–205. San Francisco: ICS Press.
Corson, Trevor. 2002. “Stalking the American Lobster.” Atlantic Monthly (April):
62–81.
Cox, James C. 2004. “How to Identify Trust and Reciprocity.” Games and
Economic Behavior 46:260–81.
Cox, James C., and Cary Deck. In press. “On the Nature of Reciprocal
Motives.” Economic Inquiry 43(3).
Cox, James C., Daniel Friedman, and Steven Gjerstad. 2004. “A Tractable Model
of Reciprocity and Fairness.” Working Paper. Tucson: University of Arizona,
Department of Economics.
Cox, James C., and Vjollca Sadiraj. 2004. “Direct Tests of Models of Social Preferences
and a New Model.” Economics Lab Working Paper. Tucson: University
of Arizona.
Crawford, Sue E. S., and Elinor Ostrom. 2000. “A Grammar of Institutions.”
Reprinted in Polycentric Games and Institutions: Readings from the Workshop
in Political Theory and Policy Analysis, ed. Michael McGinnis, 114–55. Ann
Arbor: University of Michigan Press. Originally published in American Political
Science Review 89(3) (September 1995): 582–600.
Crawford, Vincent P., and Joel Sobel. 1982. “Strategic Information Transmission.”
Econometrica 50:1431–51.
Crook, Darren S., and Anne M. Jones. 1999. “Design Principles from Traditional
Mountain Irrigation Systems (Bisses) in the Valais, Switzerland.” Mountain
Research and Development 19(2):79–122.
Crutchfield, James P., and Peter Schuster. 2003. Evolutionary Dynamics: Exploring
the Interplay of Selection, Accident, Neutrality, and Function. Oxford:
Oxford University Press.
Cruz, Wilfrido D. 1986. “Overfishing and Conflict in a Traditional Fishery: San
Miguel Bay, Philippines.” In Proceedings of the Conference on Common Property
Resource Management, National Research Council, 115–35. Washington,
DC: National Academy Press.
Cummins, Denise D. 1996. “Evidence of Deontic Reasoning in Three- and FourYear-Old
Children.” Memory and Cognition 24(6):823–29.
Dahlman, Carl J. 1980. The Open Field System and Beyond: A Property Rights
Analysis of an Economic Institution. New York: Cambridge University Press.
Dana, David. 2001. “Rethinking the Puzzle of Escalating Penalties for Repeat
Offenders.” Yale Law Journal 110 (March): 733–83.
Dannin, Ellen, and Gangaram Singh. Forthcoming. “Creating a Law Reform Laboratory:
Empirical Research and Labor Law Reform.” Wayne Law Review 50.
Dasgupta, Partha. 1982. The Control of Resources. Cambridge: Harvard University
Press.
Dawes, Robyn M. 1988. Rational Choice in an Uncertain World. Chicago:
Harcourt Brace Jovanovich.
Deadman, Peter. 1997. “Modeling Individual Behavior in Common Pool Resource
Management Experiments with Autonomous Agents.” Ph.D. diss.,
University of Arizona, Tucson.
  
R EF ER EN CE S 317
de Castro, Fabio. 2000. “Fishing Accords: The Political Ecology of Fishing Intensification
in the Amazon.” Ph.D. diss., Indiana University.
Deci, Edward L. 1975. Intrinsic Motivation. New York: Plenum Press.
Deci, Edward L., Richard Koestner, and Richard M. Ryan. 1999. “A MetaAnalytic
Review of Experiments Examining the Effects of Extrinsic Rewards
on Intrinsic Motivation.” Psychological Bulletin 125(6):627–68.
Deci, Edward L., and Richard M. Ryan. 1985. Intrinsic Motivation and SelfDetermination
in Human Behavior. New York: Plenum Press.
Decker, Christopher S., and Rafael Reuveny. 2005. “Endogenous Technological
Progress and the Malthusian Trap: Could Simon and Boserup Have Saved Easter
Island?” Human Ecology 33(1):119–40.
Delhey, Jan, and Ken Newton. 2003. “Who Trusts? The Origins of Social Trust
in Seven Societies.” European Societies 5:93–137.
de los Reyes, Romana P. 1980. Managing Communal Gravity Systems: Farmers’
Approaches and Implications for Program Planning. Quezon City, Philippines:
Ateneo de Manila University, Institute of Philippine Culture.
de Moor, Martina, Leigh Shaw-Taylor, and Paul Warde. 2002. The Management
of Common Land in North West Europe, c. 1500–1850. Turnhout, Belgium:
Brepols Publishers.
Demsetz, Harold. 1967. “Toward a Theory of Property Rights.” American
Economic Review 57 (May): 347–59.
Denzau, Arthur T., and Douglass C. North. 2000. “Shared Mental Models:
Ideologies and Institutions.” In Elements of Reason—Cognition, Choice, and
the Bounds of Rationality, ed. Arthur Lupia, Mathew D. McCubbins, and
Samuel L. Popkin, 23–46. Cambridge: Cambridge University Press.
de Soto, Hernando. 2000. The Mystery of Capital: Why Capitalism Triumphs in
the West and Fails Everywhere Else. New York: Basic Books.
Diamond, Jared. 2005. Collapse: How Societies Choose to Fail or Succeed. New
York: Viking.
Dickhaut, John, John Hubbard, Kevin McCabe, and Vernon Smith. 1997. “Trust,
Reciprocity, and Interpersonal History: Fool Me Once, Shame on You, Fool Me
Twice, Shame on Me.” Working Paper. Tucson: University of Arizona.
Diermeier, Daniel, and Keith Krehbiel. 2003. “Institutionalism as a Methodology.”
Journal of Theoretical Politics 15(2):123–44.
Dietz, Thomas, Elinor Ostrom, and Paul Stern. 2003. “The Struggle to Govern
the Commons.” Science 302 (December 12): 1907–12.
DiMaggio, Paul J., and Walter W. Powell. 1991. Introduction to The New Institutionalism
in Organizational Analysis, ed. Walter W. Powell and Paul J. DiMaggio,
1–34. Chicago: University of Chicago Press.
Dixit, Avinash K., and Susan Skeath. 1999. Games of Strategy. New York:
W. W. Norton.
Dolsˇak, Nives. 2000. “Marketable Permits: Managing Local, Regional, and
Global Commons.” Ph.D. diss., Indiana University.
Dougherty, Keith. 2001. Collective Action under the Articles of Confederation.
Cambridge: Cambridge University Press.
Downs, Anthony. 1957. An Economic Theory of Democracy. New York: Harper
and Row.
  
R EF ER EN CE S318
Dryzek, John S. 2000. Deliberative Democracy and Beyond: Liberals, Critics,
Contestations. Oxford: Oxford University Press.
Dye, Thomas R. 1981. Understanding Public Policy. 4th ed. Englewood Cliffs,
NJ: Prentice Hall.
Ebbin, Syma A. 2002. “Enhanced Fit through Institutional Interplay in the Pacific
Northwest Salmon Co-management Regime.” Marine Policy 26:253–59.
. 2004. “The Anatomy of Conflict and the Politics of Identity in Two Cooperative
Salmon Management Regimes.” Policy Sciences 37(1):71–87.
Eggertsson, Thra´inn. 2005. Imperfect Institutions: Possibilities and Limits of
Reform. Ann Arbor: University of Michigan Press.
Ellickson, Robert C. 1991. Order without Law: How Neighbors Settle Disputes.
Cambridge: Harvard University Press.
Elster, Jon. 1979. Ulysses and the Sirens: Studies in Rationality and Irrationality.
Cambridge: Cambridge University Press.
. 1989a. The Cement of Society: A Study of Social Order. Cambridge:
Cambridge University Press.
. 1989b. Solomonic Judgements: Studies in the Limitations of Rationality.
Cambridge: Cambridge University Press.
, ed. 1998. Deliberative Democracy. New York: Cambridge University
Press.
Ensminger, Jean. 1990. “Co-opting the Elders: The Political Economy of State
Incorporation in Africa.” American Anthropologist 92:662–75.
. 2000. “Experimental Economics in the Bush: Why Institutions Matter.”
In Institutions, Contracts, and Organizations, ed. Claude Menard, 158–71.
London: Edward Elgar.
. 2001. “Reputations, Trust, and the Principal Agent Problem.” In Trust
in Society, ed. Karen S. Cook, 185–201. New York: Russell Sage Foundation.
Ensminger, Jean, and Jack Knight. 1997. “Changing Social Norms: Common
Property, Bridewealth, and Clan Exogamy.” Current Anthropology 38(1):
1–24.
Epstein, Joshua M. 2001. “Learning to Be Thoughtless: Social Norms and Individual
Computation.” Computational Economics 18(1):9–24.
Esarey, Justin, and T. K. Ahn. 2004. “Strategic and Norm-Based Behavior in a
Laboratory Experiment.” Paper presented at the American Political Science Annual
Meeting, Chicago.
Eshel, Ilan, Larry Samuelson, and Avner Shaked. 1998. “Altruists, Egoists, and
Hooligans in a Local Interaction Model.” American Economic Review 88(1)
(March): 157–79.
Etzioni, Amitai. 1988. The Moral Dimension: Toward a New Economics. New
York: Free Press.
Falk, Armin, Ernst Fehr, and Urs Fischbacher. 2002. “Appropriating the Commons:
A Theoretical Explanation.” In The Drama of the Commons, National
Research Council, Committee on the Human Dimensions of Global Change,
ed. Elinor Ostrom, Thomas Dietz, Nives Dolsˇak, Paul C. Stern, Susan Stonich,
and Elke Weber, 157–92. Washington, DC: National Academy Press.
Farr, James. 1985. “Situational Analysis: Explanation in Political Science.” Journal
of Politics 47(4) (November): 1085–107.
  
R EF ER EN CE S 319
Farrell, James. 1993. “Meaning and Credibility in Cheap-Talk Games.” Games
and Economic Behavior 5:514–31.
Faysse, Nicolas. 2003. “Allocating Irrigation Water: The Impact of Strategic Interactions
on the Efficiency of Rules.” European Review of Agricultural Economics
30(3):305–32.
Feeny, David. 1988. “Agricultural Expansion and Forest Depletion in Thailand,
1900–1975.” In World Deforestation in the Twentieth Century, ed. John F.
Richards and Richard P. Tucker, 112–43. Durham, NC: Duke University Press.
Feeny, David, Susan Hanna, and Arthur F. McEvoy. 1996. “Questioning the Assumptions
of the ‘Tragedy of the Commons’ Model of Fisheries.” Land Economics
72(2) (May): 187–205.
Fehr, Ernst, and Armin Falk. 1999. “Wage Rigidity in a Competitive Incomplete
Contract Market.” Journal of Political Economy 107:106–34.
Fehr, Ernst, and Simon Ga¨chter. 1998. “Reciprocity and Economics: The Economic
Implications of Homo Reciprocans.” European Economic Review
42(3–5):845–59.
. 2000a. “Cooperation and Punishment in Public Goods Experiments.”
American Economic Review 90(4):980–94.
. 2000b. “Fairness and Retaliation: The Economics of Reciprocity.” Journal
of Economic Perspectives 14(3):159–81.
. Forthcoming. “Do Incentive Contracts Undermine Voluntary Cooperation?”
Review of Economic Studies.
Fehr, Ernst, Simon Ga¨chter, and Georg Kirchsteiger. 1996. “Reciprocal Fairness
and Non-compensating Wage Differentials.” Journal of Institutional and
Theoretical Economics 152(4):608–40.
Fehr, Ernst, Georg Kirchsteiger, and Arno Riedl. 1993. “Does Fairness Prevent
Market Clearing? An Experimental Investigation.” Quarterly Journal of Economics
108(2):437–60.
Fehr, Ernst, and Bettina Rockenbach. 2003. “Detrimental Effects of Sanctions on
Human Altruism.” Nature 422 (March 13): 137–40.
Fehr, Ernst, and Klaus Schmidt. 1999. “A Theory of Fairness, Competition, and
Cooperation.” Quarterly Journal of Economics 114(3):817–68.
Finlayson, A. Christopher, and Bonnie J. McCay. 1998. “Crossing the Threshold
of Ecosystem Resilience: The Commercial Extinction of Northern Cod.” In
Linking Social and Ecological Systems: Management Practices and Social
Mechanisms for Building Resilience, ed. Fikret Berkes and Carl Folke, 311–38.
New York: Cambridge University Press.
Firmin-Sellers, Kathryn. 1996. The Transformation of Property Rights in the
Gold Coast: An Empirical Study Applying Rational Choice Theory. New York:
Cambridge University Press.
Fishburn, Peter C. 1974. “Lexicographic Orders, Utilities, and Decision Rules: A
Survey.” Management Science 20:1442–71.
Fisher, Frank. 2003. Reframing Public Policy: Discursive Politics and Deliberative
Practices. Oxford: Oxford University Press.
Foliente, Greg C. 2000. “Developments in Performance-Based Building Codes
and Standards.” Forest Products Journal 50(7–8):12–21.
  
R EF ER EN CE S320
Folke, Carl, Fikret Berkes, and Johan Colding. 1998. “Ecological Practices and
Social Mechanisms for Building Resilience and Sustainability.” In Linking Social
and Ecological Systems: Management Practices and Social Mechanisms for
Building Resilience, ed. Fikres Berkes and Carl Folke, 414–36. New York: Cambridge
University Press.
Fortmann, Louise, and John W. Bruce. 1988. Whose Trees? Proprietary Dimensions
of Forestry. Boulder, CO: Westview Press.
Frank, Robert. 1988. Passions within Reason: The Strategic Role of the Emotions.
New York: W. W. Norton.
Frey, Bruno S. 1994. “How Intrinsic Motivation Is Crowded Out and In.” Rationality
and Society 6:334–52.
. 1997a. Not Just for the Money: An Economic Theory of Personal Motivation.
Cheltenham, England: Edward Elgar.
. 1997b. “A Constitution for Knaves Crowds out Civic Virtue.” Economic
Journal 107:1043–53.
Frey, Bruno S., Matthias Benz, and Alois Stutzer. 2004. “Introducing Procedural
Utility: Not Only What, but also How Matters.” Journal of Institutional and
Theoretical Economics 160(3):377–401.
Frey, Bruno S., and Reto Jegen. 2001. “Motivation Crowding Theory: A Survey
of Empirical Evidence.” Journal of Economic Surveys 15(5):589–611.
Frohlich, Norman, and Joe A. Oppenheimer. 1996. “Experiencing Impartiality
to Invoke Fairness in the n-PD: Some Experimental Results.” Public Choice
86(1–2):117–35.
. 2001. “Choosing: A Cognitive Model of Economic and Political Choice.”
Working Paper. Winnipeg: University of Manitoba, Faculty of Management.
Frohlich, Norman, Joe A. Oppenheimer, and Cheryl Eavey. 1987. “Choices of
Principles of Distributive Justice in Experimental Groups.” American Journal
of Political Science 31(3):606–36.
Frohlich, Norman, Joe Oppenheimer, and Anja Kurki. 2004. “Modeling OtherRegarding
Preferences and an Experimental Test.” Public Choice 119(1–2):
91–117.
Fudenberg, Drew, Bengt Holmstro¨m, and Paul Milgrom. 1990. “Short-Term Contracts
and Long-Term Agency Relationship.” Journal of Economic Theory
51:1–31.
Futemma, Celia. 2000. “Collective Action and Assurance of Property Rights to
Natural Resources: A Case Study from the Lower Amazon Region, Santarem,
Brazil.” Ph.D. diss., Indiana University.
Gadgil, Madhav, Natabar Shyam Hemam, and B. Mohan Reddy. 1998. “People,
Refugia, and Resilience.” In Linking Social and Ecological Systems: Management
Practices and Social Mechanisms for Building Resilience, ed. Fikret Berkes
and Carl Folke, 30–47. New York: Cambridge University Press.
Gadgil, Madhav, and Prema Iyer. 1989. “On the Diversification of CommonProperty
Resource Use by Indian Society.” In Common Property Resources:
Ecology and Community-Based Sustainable Development, ed. Fikret Berkes,
240–72. London: Belhaven Press.
Gambetta, Diego, ed. 1988. Trust: Making and Breaking Cooperative Relations.
Oxford: Oxford University Press.
  
R EF ER EN CE S 321
Ganz, Joan S. 1971. Rules: A Systematic Study. Paris: Mouton.
Gardner, Roy. 2003. Games for Business and Economics. 2d ed. New York: Wiley.
Gardner, Roy, and Elinor Ostrom. 1991. “Rules and Games.” Public Choice 70(2)
(May): 121–49.
Gavrilets, Sergey. 2003. “Evolution and Speciation in a Hyperspace: The Roles of
Neutrality, Selection, Mutation, and Random Drift.” In Evolutionary Dynamics:
Exploring the Interplay of Selection, Accident, Neutrality, and Function,
ed. James P. Crutchfield and Peter Schuster, 135–62. Oxford: Oxford University
Press.
Gellar, Sheldon. 2005. Democracy in Senegal: Tocquevillian Analytics in Africa.
New York: Palgrave Macmillan.
Ghate, Rucha. 2000. “The Role of Autonomy in Self-Organizing Processes: A
Case Study of Local Forest Management in India.” Paper presented at the
Workshop in Political Theory and Policy Analysis miniconference, Indiana University,
Bloomington.
. 2003. “Global Gains at Local Costs: Imposing Protected Areas; Evidence
from Central India.” International Journal of Sustainable Development and
World Ecology 10(4):377–89.
Gibbard, Allan. 1990. “Norms, Discussion, and Ritual: Evolutionary Puzzles.”
Ethics 100 (July): 787–802.
Gibson, Clark. 1999. Politicians and Poachers: The Political Economy of Wildlife
Policy in Africa. New York: Cambridge University Press.
. 2001. “Forest Resources: Institutions for Local Governance in Guatemala.”
In Protecting the Commons: A Framework for Resource Management
in the Americas, ed. Joanna Burger, Elinor Ostrom, Richard B. Norgaard, David
Policansky, and Bernard D. Goldstein, 71–89. Washington, DC: Island Press.
Gibson, Clark, Krister Andersson, Elinor Ostrom, and Sujai Shivakumar. 2005.
The Samaritan’s Dilemma: The Political Economy of Development Aid. Oxford:
Oxford University Press.
Gibson, Clark, and C. Dustin Becker. 2000. “A Lack of Institutional Demand:
Why a Strong Local Community in Western Ecuador Fails to Protect Its Forest.”
In People and Forests: Communities, Institutions, and Governance, ed.
Clark Gibson, Margaret McKean, and Elinor Ostrom, 135–61. Cambridge:
MIT Press.
Gibson, Clark, Margaret McKean, and Elinor Ostrom, eds. 2000. People and
Forests: Communities, Institutions, and Governance. Cambridge: MIT Press.
Gibson, Clark, John Williams, and Elinor Ostrom. 2005. “Local Enforcement
and Better Forests.” World Development 33(2):273–84.
Giddens, Anthony. 1979. Central Problems in Social Theory: Action, Structure,
and Contradiction in Social Analysis. Berkeley and Los Angeles: University of
California Press.
Gigerenzer, Gerd, and Daniel G. Goldstein. 1996. “Reasoning the Fast and Frugal
Way: Models of Bounded Rationality.” Psychological Review 103(4):650–69.
Gigerenzer, Gerd, and Reinhard Selten, eds. 2001. Bounded Rationality: The
Adaptive Toolbox. Cambridge: MIT Press.
Gigerenzer, Gerd, Peter M. Todd, and the ABC Research Group. 1999. Simple
Heuristics that Make Us Smart. Oxford: Oxford University Press.
  
R EF ER EN CE S322
Gilles, Jere L., and Keith Jamtgaard. 1981. “Overgrazing in Pastoral Areas: The
Commons Reconsidered.” Sociologia Ruralis 21:129–41.
Gintis, Herbert. 2000a. “Beyond Homo Economicus: Evidence from Experimental
Economics.” Ecological Economics 35(3):311–22.
. 2000b. Game Theory Evolving: A Problem-Centered Introduction to
Modeling Strategic Interaction. Princeton: Princeton University Press.
. 2004. “The Genetic Side of Gene-Culture Coevolution: Internalization of
Norms and Prosocial Emotions.” Journal of Economic Behavior and Organization
53(1):57–67.
Glaeser, Edward, David Labson, Jose Scheinkman, and Christine Soutter. 2000.
“Measuring Trust.” Quarterly Journal of Economics 115(3):811–46.
Gode, Dhananjay K., and Shyam Sunder. 1997. “What Makes Markets Allocationally
Efficient?” Quarterly Journal of Economics 112(2) (May): 603–30.
Goffman, Irving. 1974. Frame Analysis: An Essay on the Organization of Experience.
Cambridge: Harvard University Press.
Gombya-Ssembajjwe, William. 1995. “Sacred Forests in Modern Ganda Society.”
Uganda Journal 42:32–44.
Goodland, Robert, George Ledec, and Maryla Webb. 1989. “Meeting Environmental
Concerns Caused by Common-Property Mismanagement in Economic
Development Projects.” In Common Property Resources: Ecology and Community-Based
Sustainable Development, ed. Fikret Berkes, 148–63. London:
Belhaven Press.
Gopnik, Alison. 2004. “Finding Our Inner Scientist.” Daedalus 133(1):21–28.
Gordillo, Gustavo, and Krister Andersson. 2004. “From Policy Lessons to Policy
Actions: Motivation to Take Evaluation Seriously.” Public Administration and
Development 24:305–20.
Gordon, H. Scott. 1954. “The Economic Theory of a Common Property Resource:
The Fishery.” Journal of Political Economy 62:124–42.
Grafton, R. Quentin. 2000. “Governance of the Commons: A Role for the State.”
Land Economics 76(4):504–17.
Grainger, Alan. 1993. Controlling Tropical Deforestation. London: Earthscan
Publications.
Granovetter, Mark. 1978. “Threshold Models of Collective Behavior.” American
Journal of Sociology 83:1420–43.
Greif, Avner, Paul Milgrom, and Barry R. Weingast. 1994. “Coordination,
Commitment, and Enforcement: The Case of the Merchant Guild.” Journal of
Political Economy 102:745–77.
Grether, David M., R. Mark Isaac, and Charles R. Plott. 1979. “Alternative Methods
of Allocating Airport Slots: Performance and Evaluation.” Report prepared
for the Civil Aeronautics Board and Federal Aviation Administration, Polinomics
Research Laboratories, Inc., Pasadena, California.
. 1981. “The Allocation of Landing Rights by Unanimity among Competitors.”
American Economic Review 71(2):166–71.
Grether, David M., and Charles R. Plott. 1979. “Economic Theory of Choice
and the Preference Phenomenon.” American Economic Review 69 (September):
623–38.
  
R EF ER EN CE S 323
Greve, Michael S. 2001. “Affirmative Action Is on the Rocks, Thanks to College
Leaders.” Chronicle of Higher Education, April 20, 2001.
Grimm, Volker, and Christian Wissel. 1997. “Babel, or the Ecological Stability
Discussions: An Inventory and Analysis of Terminology and a Guide for
Avoiding Confusion.” Oecologia 109:323–34.
Guillet, David W. 1992a. “Comparative Irrigation Studies: The O´ rbigo Valley of
Spain and the Colca Valley of Peru´.” Polı´gonos 2:141–50.
. 1992b. Covering Ground: Communal Water Management and the State
in the Peruvian Highlands. Ann Arbor: University of Michigan Press.
Gunderson, Lance H., and Crawford S. Holling, eds. 2001. Panarchy: Understanding
Transformations in Human and Natural Systems. Washington, DC:
Island Press.
Gunderson, Lance H., and Lowell Pritchard, Jr., eds. 2002. Resilience and the
Behavior of Large-Scale Systems. Washington, DC: Island Press.
Gupta, Radhika, and Sunandan Tiwari. 2002. “At the Crossroads: Continuity
and Change in the Traditional Irrigation Practices of Ladakh.” Paper presented
at “The Commons in an Age of Globalization,” the Ninth Conference of the
International Association for the Study of Common Property, Victoria Falls,
Zimbabwe.
Gu¨rerk, O¨ zgu¨r, Bernd Irlenbusch, and Bettina Rockenbach. 2004. “On the
Evolvement of Institutions in Social Dilemmas.” Working Paper. Erfurt,
Germany: University of Erfurt, Laboratory for Experimental Economics.
Gu¨th, Werner. 1988. “On the Behavioral Approach to Distributive Justice—A
Theoretical and Experimental Investigation.” In Applied Behavioral Economics,
vol. 2, ed. Shlomo Maital, 703–17. New York: New York University Press.
. 1995. “An Evolutionary Approach to Explaining Cooperative Behavior
by Reciprocal Incentives.” International Journal of Game Theory 24:323–44.
Gu¨th, Werner, and Hartmut Kliemt. 1998. “The Indirect Evolutionary Approach:
Bridging the Gap between Rationality and Adaptation.” Rationality and Society
10(3) (August): 377–99.
Gu¨th, Werner, and Wilhelm Neuefeind. 2001. “Heuristics as Decision Rules—
Part 1: The Single Consumer.” Discussion paper. Berlin: Humboldt University.
Gu¨th, Werner, Peter Ockenfels, and Markus Wendel. 1997. “Cooperation Based
on Trust: An Experimental Investigation.” Journal of Economic Psychology
18(1):15–43.
Gu¨th, Werner, and Menahem Yaari. 1992. “An Evolutionary Approach to
Explaining Reciprocal Behavior in a Simple Strategic Game.” In Explaining
Process and Change: Approaches to Evolutionary Economics, ed. Ulrich Witt,
23–34. Ann Arbor: University of Michigan Press.
Habermas, Ju¨rgen. 1996. Between Facts and Norms: Contributions to a Discourse
Theory of Law and Democracy. Cambridge: MIT Press.
Hackett, Steven. 1992. “Heterogeneity and the Provision of Governance for
Common-Pool Resources.” Journal of Theoretical Politics 4 (July): 325–42.
Hackett, Steven, Edella Schlager, and James Walker. 1994. “The Role of Communication
in Resolving Commons Dilemmas: Experimental Evidence with
Heterogeneous Appropriators.” Journal of Environmental Economics and
Management 27:99–126.
  
R EF ER EN CE S324
Haley, Sharman. 2002. “Sustainable Development and Oil: Lessons from the Ecuadorian
Amazon.” Cultural Survival Quarterly 26:65–68.
Haller, Tobias. 2001. “Rules Which Pay Are Going to Stay: Indigenous Institutions,
Sustainable Resouerce Use and Land Tenure among the Ouldem and
Paltha, Mandara Mountains, Northern Cameroon.” In Everyday Governance
of Land in Africa, ed. Pierre-Yyes Le Meur and Christian Lund, 117–32. New
Brunswick, NJ: Transaction Publishers.
. 2002. “Common Property Resource Management, Institutional
Change, and Conflict in African Floodplain Wetlands.” African Anthropologist
9(1):25–35.
Hamburger, Henry. 1979. Games as Models of Social Phenomena. San Francisco:
W. H. Freeman.
Hammerstein, Peter, ed. 2003. Genetic and Cultural Evolution of Cooperation.
Cambridge: MIT Press.
Hammond, Thomas H., and Christopher K. Butler. 2003. “Some Complex
Answers to the Simple Question ‘Do Institutions Matter?’: Policy Choice and
Policy Change in Presidential and Parliamentary Systems.” Journal of Theoretical
Politics 15(2):145–200.
Hardin, Garrett. 1968. “The Tragedy of the Commons.” Science 162:1243–48.
Hardin, Russell. 1982. Collective Action. Baltimore: Johns Hopkins University
Press.
Harre´, Rom. 1974. “Some Remarks on ‘Rule’ as a Scientific Concept.” In
Understanding Other Persons, ed. Theodore Mischel, 143–84. Oxford: Basil
Blackwell.
Harsanyi, John C. 1955. “Cardinal Welfare, Individual Ethics, and Interpersonal
Comparison of Utility.” Journal of Political Economy 63(4):309–21.
. 1967–68. “Games with Incomplete Information Played by ‘Bayesian’
Players.” Management Science 14:159–82, 320–34, 486–502.
Hart, Herbert L. A. 1961. The Concept of Law. Oxford: Clarendon Press.
Hasson, Uri, Yuval Nir, Ifat Levy, Galit Fuhrmann, and Rafael Malach. 2004.
“Intersubject Synchronization of Cortical Activity during Natural Vision.”
Science 303 (March 12): 1634–40.
Hawkes, Kristen, James F. O’Connell, and Nicholas G. Blurton Jones. 2001.
“Hadza Meat Sharing.” Evolution and Human Behavior 22:113–42.
Hawley, Amos, and Basil G. Zimmer. 1970. The Metropolitan Community: Its
People and Government. Beverly Hills, CA: Sage.
Hayes, Tanya. 2004. “Parks, People, and Forest Protection: An Institutional Assessment
of the Effectiveness of Protected Areas.” Working Paper. Bloomington:
Indiana University, Center for the Study of Institutions, Population, and Environmental
Change (CIPEC).
Heal, Geoffrey M. 1998. Valuing the Future: Economic Theory and Sustainability.
New York: Colombia University Press.
Hechter, Michael, and Karl-Dieter Opp, eds. 2001. Social Norms. New York:
Russell Sage Foundation.
Heikkila, Tanya. 2004. “Institutional Boundaries and Common-Pool Resource
Management: A Comparative Analysis of Water Management Programs in California.”
Journal of Policy Analysis and Management 23(1) (Winter): 97–117.
  
R EF ER EN CE S 325
Heiner, Ronald A. 1983. “The Origin of Predictable Behavior.” American Economic
Review 73(4):560–95.
. 1990. “Rule-Governed Behavior in Evolution and Human Society.”
Constitutional Political Economy 1:19–46.
Heise, David R. 1979. Understanding Events: Affect and the Construction of
Social Action. New York: Cambridge University Press.
Henrich, Joseph. 2004. “Cultural Group Selection, Coevolutionary Processes and
Large-Scale Cooperation.” Journal of Economic Behavior and Organization
53(1):3–35.
Henrich, Joseph, Robert Boyd, Samuel Bowles, Colin Camerer, Ernst Fehr, and
Herbert Gintis, eds. 2004. Foundations of Human Sociality: Economic Experiments
and Ethnographic Evidence from Fifteen Small-Scale Societies. Oxford:
Oxford University Press.
Herzberg, Roberta, and Vincent Ostrom. 2000. “Votes and Vetoes.” Reprinted
in Polycentric Games and Institutions: Readings from the Workshop in Political
Theory and Policy Analysis, ed. Michael McGinnis, 168–83. Ann Arbor: University
of Michigan Press. Originally published in Guidance, Control, and Evaluation
in the Public Sector, ed. Franz-Xaver Kaufmann, Giandomenico Majone,
and Vincent Ostrom, 431–43 (Berlin and New York: Walter de Gruyter, 1986).
Hess, Charlotte. 1999. A Comprehensive Bibliography of Common Pool Resources.
(CD-ROM) Bloomington: Indiana University, Workshop in Political
Theory and Policy Analysis.
Hess, Charlotte, and Elinor Ostrom. 2003. “Ideas, Artifacts, and Facilities: Information
as a Common-Pool Research.” Law and Contemporary Problems
66(1 & 2):111–45.
Higgs, Robert. 1996. “Legally Induced Technical Regress in the Washington
Salmon Fishery.” In Empirical Studies in Institutional Change, ed. Lee J. Alston,
Thra´inn Eggertsson, and Douglass C. North, 247–79. New York: Cambridge
University Press.
Hilpinen, Risto, ed. 1971. Deontic Logic: Introductory and Systematic Readings.
Dordrecht, Holland: D. Reidel.
, ed. 1981. New Studies in Deontic Logic: Norms, Actions, and the Foundations
of Ethics. Dordrecht, Holland: D. Reidel.
Hilton, Rita. 1992. “Institutional Incentives for Resource Mobilization: An
Analysis of Irrigation Schemes in Nepal.” Journal of Theoretical Politics
4(3):283–308.
Hirschman, Albert O. 1985. “Against Parsimony: Three Easy Ways of Complicating
Some Categories of Economic Discourse.” Economics and Philosophy 1
(April): 16–19.
Hockings, Marc, and Adrian Phillips. 1999. “How Well Are We Doing? Some
Thoughts on the Effectiveness of Protected Areas.” Parks 9(2):5–16.
Hodgson, Geoffrey M. 2004a. The Evolution of Institutional Economics: Agency,
Structure, and Darwinism in American Institutionalism. London: Routledge.
. 2004b. “Opportunism Is Not the Only Reason Why Firms Exist: Why an
Explanatory Emphasis on Opportunism May Mislead Management Strategy.”
Industrial and Corporate Change 13(2):401–18.
  
R EF ER EN CE S326
Hodgson, Geoffrey M., and Thorbjφrn Knudsen. 2004. “The Complex Evolution
of a Simple Traffic Convention: The Functions and Implications of Habit.”
Journal of Economic Behavior and Organization 54:19–47.
Hofstadter, Douglas R. 1979. “. . . Ant Fugue.” In Go¨del, Escher, Bach: An Eternal
Golden Braid, ed. Douglas R. Hofstadter, 310–36. New York: Basic Books.
Hohfeld, Wesley N. 1964. Fundamental Legal Conceptions. New Haven: Yale
University Press.
Holland, John H. 1995. Hidden Order: How Adaptation Builds Complexity.
Reading, MA: Addison-Wesley.
Holland, John H., Keith J. Holyoak, Richard E. Nisbett, and Paul R. Thagard.
1986. Induction: Processes of Inference, Learning, and Discovery. Cambridge:
MIT Press.
Holling, Crawford S. 1973. “Resilience and Stability of Ecological Systems.” Annual
Review of Ecology and Systematics 4:2–23.
Holling, Crawford S., Lance H. Gunderson, and Donald Ludwig. 2001. “In Quest
of a Theory of Adaptive Change.” In Panarchy: Understanding Transformations
in Human and Natural Systems, ed. Lance Gunderson and Crawford S.
Holling, 3–24. Washington, DC: Island Press.
Hong, Lu, and Scott E. Page. 2004. “Groups of Diverse Problem Solvers Can
Outperform Groups of High-Ability Problem Solvers.” Proceedings of the
National Academy of Science 101(46):16385–89.
Huber, Oswald. 1980. “The Influence of Some Task Variables on Cognitive Operations
in an Information-Processing Decision Model.” Acta Psychologica
45:187–96.
Hudson, William E. 1980. “The New Federalism Paradox.” Policy Studies Journal
8:900–6.
Hupe, Peter, and Michael Hill. 2004. “The Three Action Levels of Governance:
Reframing the Policy Process beyond the Stages Model.” Working Paper.
Rotterdam, the Netherlands: Erasmus University, Department of Public
Administration.
Hurwicz, Leonid. 1994. “Institutional Change and the Theory of Mechanism
Design.” Academia Economic Papers 22:1–26.
Igoe, Jim. 2004. Conservation and Globalization: A Study of National Parks and
Indigenous Communities from East Africa to South Dakota. Belmont, CA:
Wadsworth Thompson Learning.
Isaac, R. Mark, and James Walker. 1988a. “Communication and Free-Riding
Behavior: The Voluntary Contribution Mechanism.” Economic Inquiry
26(4):585–608.
. 1988b. “Group Size Effects in Public Goods Provision: The Voluntary
Contributions Mechanism.” Quarterly Journal of Economics 103:179–99.
Isaac, R. Mark, James Walker, and Arlington W. Williams. 1994. “Group Size
and the Voluntary Provision of Public Goods: Experimental Evidence Utilizing
Large Groups.” Journal of Public Economics 54(1) (May): 1–36.
Ito, Masaru, Tatsuyoshi Saijo, and Masashi Une. 1995. “The Tragedy of the Commons
Revisited: Identifying Behavioral Principles.” Journal of Economic Behavior
and Organization 28(3):311–35.
  
R EF ER EN CE S 327
Ives, Anthony R., Jennifer L. Klug, and Kevin Goss. 2000. “Stability and Species
Richness in Complex Communities.” Ecology Letters 3:399–411.
Jacob, Francois. 1977. “Evolution and Tinkering.” Science 196 (4295) (June 10):
1161–66.
Jager, Wander, and Marco A. Janssen. 2003. “Using Artificial Agents to Understand
Laboratory Experiments of Common-Pool Resources with Real Agents.”
In Complexity and Ecosystem Management: The Theory and Practice of MultiAgent
Systems, ed. Marco A. Janssen, 75–102. Cheltenham, UK/Northampton,
MA: Edward Elgar.
Jager, Wander, Marco A. Janssen, and Charles A. J. Viek. 2001. “Experimentation
with Household Dynamics: The Consumat Approach.” International Journal
of Sustainable Development 4(1):90–100.
Janssen, Marco A., ed. 2003. Complexity and Ecosystem Management: The Theory
and Practice of Multi-agent Systems. Northampton, MA: Edward Elgar.
. 2004. “Evolution of Cooperation When Feedback to Reputation
Scores Is Voluntary.” Working Paper. Bloomington: Indiana University,
Center for the Study of Institutions, Population, and Environmental Change
(CIPEC).
Janssen, Marco A., and T. K. Ahn. 2003. “Adaptation vs. Anticipation in Linear
Public Good Games.” Paper presented at the first conference of the European
Social Simulation Association, Groningen University, Groningen, the Netherlands,
September 18–21.
Janssen, Marco, and Elinor Ostrom. Forthcoming a. “Governing Social-Ecological
Systems.” In Handbook of Computational Economics II: Agent-Based
Computational Economics, ed. Kenneth L. Judd and Leigh Tesfatsion. Amsterdam:
Elsevier.
Janssen, Marco, and Elinor Ostrom. Forthcoming b. “Adoption of a New Regulation
for the Governance of Common-Pool Resources by a Heterogeneous
Population.” In Inequality, Cooperation, and Environmental Sustainability, ed.
Jean-Marie Baland, Pranab Bardhan, and Samuel Bowles. Princeton: Princeton
University Press.
Jodha, Narpat S. 1990. “Depletion of Common Property Resources in India:
Micro-Level Evidence.” In Rural Development and Population: Institutions
and Policy, ed. Geoffrey McNicoll and Mead Cain, 261–83. Oxford: Oxford
University Press.
. 1996. “Property Rights and Development.” In Rights to Nature, ed.
Susan Hanna, Carl Folke, and Karl-Goran Ma¨ler, 205–22. Washington, DC:
Island Press.
Johnson, Craig, and Timothy Forsyth. 2002. “In the Eyes of the State: Negotiating
a ‘Rights-Based Approach’ to Forest Conservation in Thailand.” World Development
30(9):1591–605.
Johnson, Ronald N., and Gary D. Libecap. 1982. “Contracting Problems and
Regulation: The Case of the Fishery.” American Economic Review 72(5):
1005–23.
Jones, Bryan D. 2001. Politics and the Architecture of Choice: Bounded Rationality
and Governance. Chicago: University of Chicago Press.
  
R EF ER EN CE S328
Jones, Bryan D., Frank R. Baumgartner, and James L. True. 1996. “The Shape
of Change: Punctuations and Stability in U.S. Budgeting, 1947–1994.” Paper
presented at the Annual Meeting of the Midwest Political Science Association,
Chicago.
Jones, Eric C. 2003. “Building on Ostrom’s ‘The Rudiments of a Theory of the
Origins, Survival and Performance of Common-Property Institutions.’ ” Journal
of Ecological Anthropology 7:65–72.
Joshi, Neeraj N., Elinor Ostrom, Ganesh Shivakoti, and Wai Fung Lam. 2000.
“Institutional Opportunities and Constraints in the Performance of FarmerManaged
Irrigation Systems in Nepal.” Asia-Pacific Journal of Rural Development
10(2) (December): 67–92.
Kagel, John, and Alvin Roth, eds. 1995. The Handbook of Experimental Economics.
Princeton: Princeton University Press.
Kahneman, Daniel, and Amos Tversky, eds. 2000. Choice, Values, and Frames.
Cambridge: Cambridge University Press.
Kahneman, Daniel, Jack Knetsch, and Richard Thaler. 1990. “Experimental Tests
of the Endowment Effect and the Coase Theorem.” Journal of Political Economy
98(6):1325–48.
Kameda, Tatsuya, Masanori Takezawa, and Reid Hastie. 2003. “The Logic of
Social Sharing: An Evolutionary Game Analysis of Adaptive Norm Development.”
Personality and Social Psychology Review 7(1):2–19.
Kaminski, Antoni. 1992. An Institutional Theory of Communist Regimes: Design,
Function, and Breakdown. Oakland, CA: ICS Press.
Kaminski, Juliane, Josep Call, and Julia Fischer. 2004. “Word Learning in a Domestic
Dog: Evidence for ‘Fast Mapping.’ ” Science 304 (June 11): 1682–83.
Kaplan, Hillard, Kim Hill, and A. Magdalena Hurtado. 1990. “Risk, Foraging,
and Food Sharing among the Ache.” In Risk and Uncertainty in Tribal and
Peasant Economies, ed. Elizabeth Cashdan, 107–44. Boulder, CO: Westview
Press.
Karkkainen, Bradley. 2001/2. “Collaborative Ecosystem Governance: Scale,
Complexity, and Dynamism.” Virginia Environmental Law Journal 21:
189–243.
Karotkin, Drora, and Jacob Paroush. 1994. “Variability of Decisional Ability and
the Essential Order of Decision Rules.” Journal of Economic Behavior and Organization
23:343–54.
Kikuchi, Masao, Masako Fujita, Esther Marciano, and Yujiro Hayami. 1998.
“State and Community in the Deterioration of a National Irrigation System.”
Paper presented at the World Bank-EDI Conference “Norms and Evolution in
the Grassroots of Asia,” Stanford University, Stanford, CA, February 6–7.
Kikuchi, Masao, Yoriko Watanabe, and Toshio Yamagishi. 1996. “Accuracy in
the Prediction of Others’ Trustworthiness and General Trust: An Experimental
Study.” Japanese Journal of Experimental Social Psychology 37(1):23–36.
Kirchler, Erich, Ernst Fehr, and Robert Evans. 1996. “Social Exchange in the
Labor Market: Reciprocity and Trust versus Egoistic Money Maximization.”
Journal of Economic Psychology 17:313–41.
  
R EF ER EN CE S 329
Kiser, Larry L., and Elinor Ostrom. 1982. “The Three Worlds of Action: A Metatheoretical
Synthesis of Institutional Approaches.” In Strategies of Political
Inquiry, ed. Elinor Ostrom, 179–222. Beverly Hills, CA: Sage.
Knack, Stephen. 1992. “Civic Norms, Social Sanctions, and Voter Turnout.”
Rationality and Society 4:133–56.
Knight, Frank H. 1921. Risk, Uncertainty, and Profit. Boston: Houghton Mifflin.
. 1965. Freedom and Reform: Essays in Economics and Social Philosophy.
New York: Harper and Row.
Knight, Jack. 1992. Institutions and Social Conflict. New York: Cambridge University
Press.
Knox, Anna, and Ruth Meinzen-Dick. 2001. “Collective Action, Property Rights,
and Devolution of Forest and Protected Area Management.” In Collective Action,
Property Rights, and Devolution of Natural Resource Management, ed.
Ruth Meinzen-Dick, Anna Knox, and Monica DiGregorio, 41–74. Feldafing,
Germany: Deutsche Stiftung fu¨r internationale Entwicklung.
Koestler, Arthur. 1973. “The Tree and the Candle.” In Unity through Diversity,
pt I, ed. William Gray and Nicholas D. Rizzo, 287–314. New York: Gordon
and Breach Science Publishers.
Koford, Kenneth. 2003. “Experiments on Trust and Bargaining in Bulgaria: The
Effects of Institutions and Culture.” Prepared for presentation at the International
Society of New Institutional Economics 7th Annual Meeting, Budapest,
September 11–13.
Koford, Kenneth, and Jeffrey B. Miller. 1995. “Contracts in Bulgaria: How Firms
Cope When Property Rights Are Incomplete.” Working Paper. Newark: University
of Delaware.
Koremenos, Barbara, Charles Lipson, and Duncan Snidal. 2001. “The Rational
Design of International Institutions.” International Organization 55(4) (Autumn):
761–99.
Korten, David C. 1980. “Community Organization and Rural Development: A
Learning Process Approach.” Public Administration Review 40(5):480–511.
Kreps, David M. 1990. “Corporate Culture and Economic Theory.” In Perspectives
on Positive Political Economy, ed. James Alt and Kenneth Shepsle, 90–
143. New York: Cambridge University Press.
Kreps, David M., Paul Milgrom, John Roberts, and Robert Wilson. 1982.
“Rational Cooperation in the Finitely Repeated Prisoners’ Dilemma.” Journal
of Economic Theory 27:245–52.
Kuran, Timur. 1989. “Sparks and Prairie Fires: A Theory of Unanticipated Political
Revolution.” Public Choice 61:41–74.
Kurzban, Robert. 2003. “Biological Foundations of Reciprocity.” In Trust and
Reciprocity: Interdisciplinary Lessons from Experimental Research, ed. Elinor
Ostrom and James Walker, 105–27. New York: Russell Sage Foundation.
Laffont, Jean-Jacques, and David Martimort. 2002. The Theory of Incentives:
The Principal-Agent Model. Princeton: Princeton University Press.
Laland, Kevin N., John Odling-Smee, and Marcus W. Feldman. 2000. “Niche
Construction, Biological Evolution, and Cultural Change.” Behavioral and
Brain Sciences 23:131–75.
  
R EF ER EN CE S330
Lam, Wai Fung. 1998. Governing Irrigation Systems in Nepal: Institutions, Infrastructure,
and Collective Action. Oakland, CA: ICS Press.
Landau, Martin. 1969. “Redundancy, Rationality, and the Problem of Duplication
and Overlap.” Public Administration Review 29(4):346–58.
. 1973. “Federalism, Redundancy, and System Reliability.” Publius
3(2):173–96.
Lansing, J. Stephen. 1991. Priests and Programmers: Technologies of Power in
the Engineered Landscape of Bali. Princeton: Princeton University Press.
Lasswell, Harold, and Abraham Kaplan. 1950. Power and Society: A Framework
for Political Inquiry. New Haven: Yale University Press.
Lawry, Steven W. 1990. “Tenure Policy toward Common Property Natural Resources
in Sub-Saharan Africa.” Natural Resources Journal 30:403–22.
Lederman, Leandra. 2003. “The Interplay between Norms and Enforcement in
Tax Compliance.” Ohio State Law Journal 64(6):1453–514.
Ledyard, John. 1995. “Public Goods: A Survey of Experimental Research.” In
The Handbook of Experimental Economics, ed. John Kagel and Alvin Roth,
111–94. Princeton: Princeton University Press.
Less, Hangwoo. 2002. “‘No Artificial Death, Only Natural Death’: The Dynamics
of Centralization and Decentralization of Usenet Newsgroups.” Information
Society 18(5):361–70.
Levi, Margaret. 1988. Of Rule and Revenue. Berkeley and Los Angeles: University
of California Press.
. 1990. “A Logic of Institutional Change.” In The Limits of Rationality,
ed. Karen S. Cook and Margaret Levi, 403–18. Chicago: Chicago University
Press.
. 1997a. Consent, Dissent, and Patriotism. New York: Cambridge University
Press.
. 1997b. “A Model, a Method, and a Map: Rational Choice in Comparative
and Historical Analysis.” In Comparative Politics: Rationality, Culture,
and Structure, ed. Mark Irving Lichbach and Alan S. Zuckerman, 19–42. Cambridge:
Cambridge University Press.
Levin, Jonathan, and Barry Nalebuff. 1995. “An Introduction to Vote-Counting
Schemes.” Journal of Economic Perspectives 9(1):3–26.
Levin, Simon A. 1992. “The Problem of Pattern and Scale in Ecology.” Ecology
73(6) (December): 1943–67.
. 1998. “Ecosystems and the Biosphere as Complex Adaptive Systems.”
Ecosystems 1:431–36.
Levine, Gilbert. 1980. “The Relationship of Design, Operation, and Management.”
In Irrigation and Agricultural Development in Asia, ed. E. Walter Coward,
51–64. Ithaca: Cornell University Press.
Lewis, David K. 1969. Convention: A Philosophical Study. Cambridge: Harvard
University Press.
Lian, Peng, and Charles R. Plott. 1998. “General Equilibrium, Markets, Macroeconomics,
and Money in a Laboratory Experimental Environment.” Economic
Theory 12(1):21–75.
Libecap, Gary D. 1989. Contracting for Property Rights. New York: Cambridge
University Press.
  
R EF ER EN CE S 331
. 1995. “The Conditions for Successful Collective Action.” In Local Commons
and Global Interdependence: Heterogeneity and Cooperation in Two Domains,
ed. Robert O. Keohane and Elinor Ostrom, 161–90. London: Sage.
. 1996. “Economic Variables and the Development of the Law: The Case
of Western Mineral Rights.” In Empirical Studies in Institutional Change, ed.
Lee Alston, Thra´inn Eggertsson, and Douglass North, 34–58. New York: Cambridge
University Press.
Libecap, Gary D., and Steven N. Wiggins. 1984. “Contractual Responses to the
Common Pool: Prorationing of Crude Oil Production.” American Economic
Review 74:87–98.
. 1985. “The Influence of Private Contractual Failure on Regulation: The
Case of Oil Field Unitization.” Journal of Political Economy 93:690–714.
Lobe, Kenton, and Fikret Berkes. 2004. “The Padu System of Community-Based
Fisheries Management: Change and Local Institutional Innovation in South
India.” Marine Policy 28:271–81.
Loveman, Brian. 1993. The Constitution of Tyranny: Regimes of Exception in
Spanish America. Pittsburgh: University of Pittsburgh Press.
Low, Bobbi S. 1989. “Cross-Cultural Patterns in the Training of Children: An
Evolutionary Perspective.” Journal of Comparative Psychology 103:311–19.
. 1990. “Human Responses to Environmental Extremeness and Uncertainty:
A Cross-Cultural Perspective.” In Risk and Uncertainty in Tribal and
Peasant Economies, ed. Elizabeth Cashdan, 229–56. Boulder, CO: Westview
Press.
Low, Bobbi S., Elinor Ostrom, Carl Simon, and James Wilson. 2003. “Redundancy
and Diversity: Do They Influence Optimal Management?” In Navigating
Social-Ecological Systems: Building Resilience for Complexity and Change, ed.
Fikret Berkes, Johan Colding, and Carl Folke, 83–114. New York: Cambridge
University Press.
Luce, R. Duncan. 1956. “Semiorders and a Theory of Utility Discrimination.”
Econometrica 24:178–91.
Maass, Arthur, and Raymond L. Anderson. 1986. . . . and the Desert Shall
Rejoice: Conflict, Growth, and Justice in Arid Environments. Malabar, FL:
R. E. Krieger.
Maier-Rigaud, Frank P., and Jose´ Apesteguia. 2003. “The Role of Choice in Social
Dilemma Experiments.” Working Paper no. 2003/7. Bonn, Germany: Max
Planck Institute for Human Development.
Manktelow, Ken I., and David E. Over. 1991. “Social Roles and Utilities in
Reasoning with Deontic Conditionals.” Cognition 39:85–105.
Mansbridge, Jane. 1990. “On the Relation of Altruism and Self-Interest.” In
Beyond Self-Interest, ed. Jane Mansbridge, 133–43. Chicago: University of
Chicago Press.
. 1994. “Public Spirit in Political Systems.” In Values and Public Policy,
ed. Henry J. Aaron, Thomas E. Mann, and Timothy Taylor, 146–72. Washington,
DC: Brookings Institution.
Mansuri, Ghazala, and Vijayendra Rao. 2003. “Evaluating Community Driven
Development: A Review of the Evidence.” Washington, DC: World Bank,
Development Research Group.
  
R EF ER EN CE S332
Mantzavinos, Chris, Douglass C. North, and Syed Shariq. 2004. “Learning, Institutions,
and Economic Performance.” Perspectives on Politics 2(1):75–84.
March, James G., and Johan P. Olsen. 1989. Rediscovering Institutions: The
Organizational Basis of Politics. New York: Free Press.
Marcus, Gary. 2004. The Birth of the Mind: How a Tiny Number of Genes
Creates the Complexities of Human Thought. New York: Basic Books.
Margolis, Howard. 1991. “Self-Interest and Social Motivation.” In Individuality
and Cooperative Action, ed. Joseph E. Earley, 129–36. Washington, DC:
Georgetown University Press.
Marwell, Gerald, and Pamela Oliver. 1993. The Critical Mass in Collective Action.
Cambridge: Cambridge University Press.
Maynard Smith, John, and David Harper. 2003. Animal Signals. Oxford: Oxford
University Press.
Maynard Smith, John, and Eo¨rs Szathma´ry. 1997. The Major Transitions in
Evolution. Oxford: Oxford University Press.
Mayntz, Renate, and Fritz W. Scharpf. 1995. Steuerung and Selbstorganisation in
Staatsnahen Sektoren. Frankfurt am Main: Campus-Verlag.
McAdams, Richard H. 2001. “Conventions and Norms: Philosophical Aspects.”
In International Encyclopedia of the Social and Behavioral Sciences, ed. Neil
Smelser and Paul Baltes, 2735–41. Oxford: Pergamon.
McAdams, Richard H., and Eric Rasmusen. Forthcoming. “Norms in Law and
Economics.” In Handbook of Law and Economics, ed. A. Mitchell Polinsky
and Steven Shavell. Amsterdam: Elsevier.
McCabe, Kevin A. 2003. “A Cognitive Theory of Reciprocal Exchange.” In
Trust and Reciprocity: Interdisciplinary Lessons from Experimental Research,
ed. Elinor Ostrom and James Walker, 147–69. New York: Russell Sage
Foundation.
McCabe, Kevin A., Stephen J. Rassenti, and Vernon L. Smith. 1996. “Game Theory
and Reciprocity in Some Extensive Form Experimental Games.” Proceedings
of the National Academy of Sciences 93 (November): 13421–428.
McCabe, Kevin A., and Vernon Smith. 2003. “Strategic Analysis in Games: What
Information Do Players Use?” In Trust and Reciprocity: Interdisciplinary
Lessons from Experimental Research, ed. Elinor Ostrom and James Walker,
275–301. New York: Russell Sage Foundation.
McCay, Bonnie J., and James M. Acheson. 1987. The Question of the Commons:
The Culture and Ecology of Communal Resources. Tucson: University of
Arizona Press.
McDermott, Rose. 2001. “The Psychological Ideas of Amos Tversky and Their
Relevance for Political Science.” Journal of Theoretical Politics 13(1):5–33.
McGinnis, Michael, ed. 1999a. Polycentric Governance and Development: Readings
from the Workshop in Political Theory and Policy Analysis. Ann Arbor:
University of Michigan Press.
, ed. 1999b. Polycentricity and Local Public Economies: Readings from
the Workshop in Political Theory and Policy Analysis. Ann Arbor: University
of Michigan Press.
  
R EF ER EN CE S 333
, ed. 2000. Polycentric Games and Institutions: Readings from the
Workshop in Political Theory and Policy Analysis. Ann Arbor: University of
Michigan Press.
. Forthcoming. Dilemmas of Global Response to Conflict: Lessons from
the Horn of Africa. Bloomington: Indiana University, Workshop in Political
Theory and Policy Analysis.
McGinnis, Michael, and John T. Williams. 1989. “Change and Stability in
Superpower Rivalry.” American Political Science Review 83(4) (December):
1101–23.
. 2001. Compound Dilemmas: Democracy, Collective Action, and Superpower
Rivalry. Ann Arbor: University of Michigan Press.
McKean, Margaret A. 1992. “Management of Traditional Common Lands
(Iriaichi) in Japan.” In Making the Commons Work: Theory, Practice, and Policy,
ed. Daniel W. Bromley et al., 63–98. San Francisco: ICS Press.
. 2000. “Common Property: What Is It, What Is It Good For, and What
Makes It Work?” In People and Forests: Communities, Institutions, and
Governance, ed. Clark Gibson, Margaret McKean, and Elinor Ostrom, 27–55.
Cambridge: MIT Press.
Medard, Modesta, and Kim Geheb. 2001. Fisheries Co-management in the Tanzanian
Sector of Lake Victoria. Mwanza, Tanzania: Tanzania Fisheries Research
Institute.
Mehra, Shakuntla. 1981. Instability in Indian Agriculture in the Context of the
New Technology. Research Report no. 25. Washington, DC: International Food
Policy Research Institute.
Meinzen-Dick, Ruth, K. Vengamma Raju, and Ashok Gulati. 2002. “What
Affects Organization and Collective Action for Managing Resources? Evidence
from Canal Irrigation Systems in India.” World Development 30(4):
649–66.
Merrey, Douglas J. 1996. “Institutional Design Principles for Accountability on
Large Irrigation Systems.” Research Report no. 8. Colombo, Sri Lanka: International
Irrigation Management Institute.
Messner, Matthias, and Mattias K. Polborn. 2004. “Voting on Majority Rules.”
Review of Economic Studies 71(1):115–32.
Meyer, John W., and Brian Rowan. 1991. “Institutionalized Organizations: Formal
Structure as Myth and Ceremony.” In The New Institutionalism in Organizational
Analysis, ed. Walter W. Powell and Paul J. DiMaggio, 41–62. Chicago:
University of Chicago Press.
Milgrom, Paul R., Douglass C. North, and Barry R. Weingast. 1990. “The Role
of Institutions in the Revival of Trade: The Law Merchant, Private Judges, and
the Champagne Fairs.” Economics and Politics 2(1) (March): 1–23.
Milgrom, Paul R., and John Roberts. 1992. Economics, Organization, and
Management. Englewood Cliffs, NJ: Prentice Hall.
Millar, Karen U., and Abraham Tesser. 1988. “Deceptive Behavior in Social Relationships:
A Consequence of Violated Expectations.” Journal of Psychology
122:263–73.
  
R EF ER EN CE S334
Miller, Gary J. 2001. “Why Is Trust Necessary in Organizations? The Moral
Hazard of Profit Maximization.” In Trust in Society, ed. Karen S. Cook, 307–
31. New York: Russell Sage Foundation.
Monroe, Burt. Forthcoming. Electoral Systems in Theory and Practice: Toward
an Empirically Relevant Theory of Social Choice. Ann Arbor: University of
Michigan Press.
Moore, Mark H. 1995. Creating Public Value: Strategic Management in Government.
Cambridge: Harvard University Press.
Moran, Emilio, and Elinor Ostrom, eds. 2005. Seeing the Forest and the Trees:
Human-Environment Interactions in Forest Ecosystems. Cambridge: MIT
Press.
Morrow, Christopher E., and Rebecca Watts Hull. 1996. “Donor-Initiated
Common Pool Resource Institutions: The Case of the Yanesha Forestry Cooperative.”
World Development 24(10):1641–57.
Moxnes, Erling. 1996. “Mismanagement of Fish Resources.” Working Paper
no. 1. Bergen, Norway: SNF.
. 1998. “Not Only the Tragedy of the Commons: Misperceptions of
Bioeconomics.” Management Science 44(9):1234–48.
Mueller, Milton L. 2002. Ruling the Roots. Cambridge: MIT Press.
Mueller, Ulrich O., Benzion Chanowitz, and Ellen Langer. 1983. “The Rationality
of the Free Ride.” Paper presented at the Annual Meetings of the Public Choice
Society, Savannah, Georgia, March 24–26.
Mukand, Sharun, and Dani Rodrik. 2002. “In Search of the Holy Grail: Policy
Convergence, Experimentation, and Economic Performance.” Mimeo. Cambridge,
MA: National Bureau of Economic Research.
Mwangi, Esther. 2003. “Institutional Change and Politics: The Transformation
of Property Rights in Kenya’s Maasailand.” Ph.D. diss., Indiana University.
Myers, Ransom A., and Boris Worm. 2003. “Rapid Worldwide Depletion of
Predatory Fish Communities.” Nature 423 (May 15): 280–83.
Myerson, Roger B. 1991. Game Theory: Analysis of Conflict. Cambridge:
Harvard University Press.
Nardulli, Peter. 1995. “The Concept of a Critical Realignment, Electoral Behaviors,
and Political Change.” American Political Science Review 89:10–22.
National Research Council. 2002. Drama of the Commons. Committee on the
Human Dimensions of Global Change, ed. Elinor Ostrom, Thomas Dietz, Nives
Dolsˇak, Paul C. Stern, Susan Stonich, and Elke Weber. Washington, DC: National
Academy Press.
Nemarundwe, Nontokozo, and Witness Kozanayi. 2003. “Institutional Arrangements
for Water Resource Use: A Case Study from Southern Zimbabwe.” Journal
of Southern African Studies 29(1):193–206.
Netting, Robert McC. 1972. “Of Men and Meadows: Strategies of Alpine Land
Use.” Anthropological Quarterly 45:132–44.
. 1974. “The System Nobody Knows: Village Irrigation in the Swiss Alps.”
In Irrigation’s Impact on Society, ed. Theodore E. Downing and McGuire
Gibson, 67–75. Tucson: University of Arizona Press.
. 1993. Smallholders, Householders: Farm Families and the Ecology of
Intensive, Sustainable Agriculture. Stanford, CA: Stanford University Press.
  
R EF ER EN CE S 335
Niskanen, William A. 1971. Bureaucracy and Representative Government. Chicago:
Aldine Atherton.
North, Douglass C. 1981. Structure and Change in Economic History. New York:
W. W. Norton.
. 1986. “The New Institutional Economics.” Journal of Institutional and
Theoretical Economics 142:230–37.
. 1990. Institutions, Institutional Change, and Economic Performance.
New York: Cambridge University Press.
. 2005. Understanding the Process of Institutional Change. Princeton:
Princeton University Press.
Noussair, Charles N., Charles R. Plott, and Raymond G. Reizman. 1995. “An
Experimental Investigation of the Patterns of International Trade.” American
Economic Review 85(3) (June): 462–91.
Nowak, Martin A., and Karl Sigmund. 1998. “Evolution of Indirect Reciprocity
by Image Scoring.” Nature 393(6685):573–77.
Oakerson, Ronald J. 1981. “Erosion of Public Goods: The Case of Coal-Haul
Roads in Eastern Kentucky.” Research in Public Policy Analysis and Management
2:73–102.
. 1992. “Analyzing the Commons: A Framework.” In Making the Commons
Work: Theory, Practice, and Policy, ed. Daniel W. Bromley et al., 41–59.
Oakland, CA: ICS Press.
. 1994. “The Logic of Multi-level Institutional Analysis.” Paper presented
at the “Workshop on the Workshop” conference, Workshop in Political Theory
and Policy Analysis, Indiana University, Bloomington, June 15–19.
Oaksford, Mike, and Nick Chater. 1994. “A Rational Analysis of the Selection
Task as Optimal Data Selection.” Psychological Review 101(4):608–31.
Offe, Claus, and Helmut Wiesenthal. 1980. “Two Logics of Collective Action:
Theoretical Notes on Social Class and Organizational Form.” Political Power
and Social Theory 1:67–115.
O’Leary, Brendan, Ian S. Lustick, and Thomas Callaghy. 2001. Right-Sizing the
State: The Politics of Moving Borders. Oxford: Oxford University Press.
Olson, Mancur. 1965. The Logic of Collective Action: Public Goods and the
Theory of Groups. Cambridge: Harvard University Press.
. 1991. “The Role of Morals and Incentives in Society.” In Individuality
and Cooperative Action, ed. Joseph E. Earley, 117–28. Washington, DC:
Georgetown University Press.
Opp, Karl-Dieter. 1982. “The Evolutionary Emergence of Norms.” British Journal
of Social Psychology 21:139–49.
Orbell, John M., and Robyn M. Dawes. 1991. “A ‘Cognitive Miser’ Theory of
Cooperators’ Advantage.” American Political Science Review 85(2):515–28.
. 1993. “Social Welfare, Cooperators’ Advantage, and the Option of
Not Playing the Game.” American Sociological Review 58(6) (December):
787–800.
Orbell, John M., Tomonori Morikawa, Jason Hartwig, James Hanley, and Nicholas
Allen. 2004. “ ‘Machiavellian’ Intelligence as a Basis for the Evolution of
Cooperative Dispositions.” American Political Science Review 98(1):1–15.
  
R EF ER EN CE S336
Orbell, John M., Perry Schwarz-Shea, and Randy T. Simmons. 1984. “Do Cooperators
Exit More Readily Than Defectors?” American Political Science Review
78:147–62.
Orbell, John M., Alphons J. van de Kragt, and Robyn M. Dawes. 1991. “Covenants
without the Sword: The Role of Promises in Social Dilemma Circumstances.”
In Social Norms and Economic Institutions, ed. Kenneth J. Koford
and Jeffrey B. Miller, 117–34. Ann Arbor: University of Michigan Press.
Osterloh, Margit, and Bruno S. Frey. 2000. “Motivation, Knowledge Transfer,
and Organizational Forms.” Organization Science 11(5) (September–October):
538–50.
Ostrom, Elinor. 1986. “An Agenda for the Study of Institutions.” Public Choice
48:3–25.
. 1990. Governing the Commons: The Evolution of Institutions for Collective
Action. New York: Cambridge University Press.
. 1992a. Crafting Institutions for Self-Governing Irrigation Systems. San
Francisco, CA: ICS Press.
. 1992b. “The Rudiments of a Theory of the Origins, Survival, and Performance
of Common-Property Institutions.” In Making the Commons Work:
Theory, Practice, and Policy, ed. Daniel W. Bromley et al., 41–59. Oakland,
CA: ICS Press.
. 1996. “Incentives, Rules of the Game, and Development.” In Proceedings
of the Annual World Bank Conference on Development Economics 1995, 207–
34. Washington, DC: World Bank.
. 1998. “A Behavioral Approach to the Rational Choice Theory of Collective
Action.” American Political Science Review 92(1) (March): 1–22.
. 1999. “Institutional Rational Choice: An Assessment of the Institutional
Analysis and Development Framework.” In Theories of the Policy Process, ed.
Paul A. Sabatier, 35–71. Boulder, CO: Westview Press.
. 2001. “Reformulating the Commons.” In Protecting the Commons: A
Framework for Resource Management in the Americas, ed. Joanna Burger,
Elinor Ostrom, Richard B. Norgaard, David Policansky, and Bernard D.
Goldstein, 17–41. Washington, DC: Island Press.
Ostrom, Elinor, Arun Agrawal, William Blomquist, Edella Schlager, Shui Yan
Tang, et al. 1989. CPR Coding Manual. Bloomington: Indiana University,
Workshop in Political Theory and Policy Analysis.
Ostrom, Elinor, and T. K. Ahn, eds. 2003. Foundations of Social Capital. Cheltenham,
UK: Edward Elgar.
Ostrom, Elinor, Joanna Burger, Christopher Field, Richard B. Norgaard, and
David Policansky. 1999. “Revisiting the Commons: Local Lessons, Global
Challenges.” Science 284(5412) (April 9): 278–82.
Ostrom, Elinor, Roy Gardner, and James Walker. 1994. Rules, Games, and Common-Pool
Resources. Ann Arbor: University of Michigan Press.
Ostrom, Elinor, Roger B. Parks, and Gordon P. Whitaker. 1974. “Defining and
Measuring Structural Variations in Interorganizational Arrangements.” Publius
4(4) (Fall): 87–108.
  
R EF ER EN CE S 337
Ostrom, Elinor, Larry Schroeder, and Susan Wynne. 1993. Institutional Incentives
and Sustainable Development: Infrastructure Policies in Perspective. Boulder,
CO: Westview Press.
Ostrom, Elinor, and James Walker. 1991. “Communication in a Commons: Cooperation
without External Enforcement.” In Laboratory Research in
Political Economy, ed. Thomas R. Palfrey, 287–322. Ann Arbor: University of
Michigan Press.
, eds. 2003. Trust and Reciprocity: Interdisciplinary Lessons from Experimental
Research. New York: Russell Sage Foundation.
Ostrom, Elinor, James Walker, and Roy Gardner. 1992. “Covenants with and
without a Sword: Self-Governance Is Possible.” American Political Science Review
86(2) (June): 404–17.
Ostrom, Vincent. 1980. “Artisanship and Artifact.” Public Administration Review
40(4) (July/August): 309–17.
. 1986. “A Fallabilist’s Approach to Norms and Criteria of Choice.” In
Guidance, Control, and Evaluation in the Public Sector, ed. Franz K. Kaufmann,
Giandomenico Majone, and Vincent Ostrom, 229–44. Berlin: Walter de
Gruyter.
. 1987. The Political Theory of a Compound Republic: Designing the
American Experiment. 2d rev. ed. San Francisco, CA: ICS Press.
. 1991. The Meaning of American Federalism: Constituting a Self-Governing
Society. Oakland, CA: ICS Press.
. 1993. “Epistemic Choice and Public Choice.” Public Choice 77(1)
(September): 163–76.
. 1996. “Faustian Bargains.” Constitutional Political Economy 7:303–8.
. 1997. The Meaning of Democracy and the Vulnerability of Democracies:
A Response to Tocqueville’s Challenge. Ann Arbor: University of Michigan
Press.
. 1999. “Polycentricity (Parts 1 and 2).” In Polycentricity and Local
Public Economies: Readings from the Workshop in Political Theory and Policy
Analysis, ed. Michael McGinnis, 52–74, 119–38. Ann Arbor: University of
Michigan Press.
Ostrom, Vincent, David Feeny, and Hartmut Picht, eds. 1993. Rethinking Institutional
Analysis and Development: Issues, Alternatives, and Choices. 2d ed.
Oakland, CA: ICS Press.
Ostrom, Vincent, and Elinor Ostrom. 1970. “Conditions of Legal and Political
Feasibility.” In Natural Resource System Models in Decision Making, Proceedings
of a 1969 Water Resources Seminar, ed. Gerrit H. Toebes. Lafayette, IN:
Purdue University, Water Resources Research Center.
. 1977. “Public Goods and Public Choices.” In Alternatives for Delivering
Public Services: Toward Improved Performance, ed. E. S. Savas, 7–49. Boulder,
CO: Westview Press.
. 1999. “Legal and Political Conditions of Water Resource Development.”
Reprinted in Polycentric Governance and Development: Readings from the
Workshop in Political Theory and Policy Analysis, ed. Michael McGinnis, 42–
  
R EF ER EN CE S338
59. Ann Arbor: University of Michigan Press. Originally published in Land
Economics 48(1) (February 1972): 1–14.
Ostrom, Vincent, Charles M. Tiebout, and Robert Warren. 1961. “The Organization
of Government in Metropolitan Areas: A Theoretical Inquiry.” American
Political Science Review 55(4) (December): 831–42.
Palanisami, K. 1982. Managing Tank Irrigation Systems: Basic Issues and Implications
for Improvement. Presented at the workshop “Tank Irrigation: Problems
and Prospects.” Bogor, Indonesia: CIFOR.
Pant, Takur Nath, and Jyoti Lohani. 1983. “Some Observations on Irrigation
Development of Nepal.” In Water Management in Nepal: Proceedings of the
Seminar on Water Management Issues, held in Kathmandu, 31 July–2 August,
ed. Edward Martin and Robert Yoder, 3–46. Kathmandu: Ministry of Food and
Agriculture, Agricultural Projects Services Centre, Agricultural Development
Council.
Parker, Dawn C., Steven M. Manson, Marco A. Janssen, Mathew J. Hoffmann,
and Peter Deadman. 2003. “Multi-agent Systems for the Simulation of LandUse
and Land-Cover Change: A Review.” Annals of the Association of American
Geographers 93(2):314–37.
Pessoa, Luiz. 2004. “Seeing the World in the Same Way.” Science 3003 (March
12): 1617–18.
Pickett, Steward T. A., and Mary L. Cadenasso. 1995. “Landscape Ecology:
Spatial Heterogeneity in Ecological Systems.” Science 269 (July 21): 331–34.
Pinker, Steven. 1994. The Language Instinct. New York: W. Morrow.
Pinkerton, Evelyn, ed. 1989. Co-operative Management of Local Fisheries: New
Directions for Improved Management and Community Development. Vancouver:
University of British Columbia Press.
Platteau, Jean-Philippe. 2003. “Community Based Development in the Context
of Within Group Heterogeneity. ” Paper presented at the Annual Bank Conference
on Development Economics, Bangalore, India.
. 2004. “Monitoring Elite Capture in Community-Driven Development.”
Development and Change 35(2):223–46.
Platteau, Jean-Philippe, and Frederic Gaspart. 2003. “The Risk of Resource Misappropriation
in Community-Driven Development.” World Development
31(10):1687–1703.
Plott, Charles R. 1996. “Rational Individual Behaviour in Markets and Social
Choice Processes: The Discovered Preference Hypothesis.” In The Rational
Foundations of Economic Behaviour, ed. Kenneth Arrow, Enrico Colombatto,
Mark Perlman, and Christian Schmidt, 225–50. New York: St. Martin’s Press.
Plott, Charles R., and Michael E. Levine. 1978. “A Model of Agenda Influence
on Committee Decisions.” American Economic Review 68:146–60.
Plott, Charles R., and Robert A. Meyer. 1975. “The Technology of Public Goods,
Externalities, and the Exclusion Principle.” In Economic Analysis of Environmental
Problems, ed. Edwin S. Mills, 65–94. New York: Columbia University
Press.
Plott, Charles R., and David P. Porter. 1996. “Market Architectures and Institutional
Testbedding: An Experiment with Space Station Pricing Policies.” Journal
of Economic Behavior and Organization 31(2):237–72.
  
R EF ER EN CE S 339
Poffenberger, Mark, and Betsy McGean. 1996. Village Voices, Forest Choices:
Joint Forest Management in India. New Delhi: Oxford University Press.
Polski, Margaret M. 2003. The Invisible Hands of U.S. Commercial Banking Reform:
Private Action and Public Guarantees. Boston: Kluwer.
Popper, Karl. 1961. The Poverty of Historicism. New York: Harper and Row.
. 1976. “The Logic of the Social Sciences.” In The Positivist Dispute in
German Sociology, ed. Theodore W. Adorno, 87–104. New York: Harper and
Row.
Posner, Richard A., and Eric Rasmusen. 1999. “Creating and Enforcing Norms,
with Special Reference to Sanctions.” International Review of Law and Economics
19(3) (September): 369–82.
Poteete, Amy, and Elinor Ostrom. 2004. “Heterogeneity, Group Size, and Collective
Action: The Role of Institutions in Forest Management.” Development and
Change 35(3):435–61.
Poudel, Rabi, Kala N. Pandit, Keshav Adhikari, S. M. Shakya, D. N. Yadav, and
Neeraj R. Joshi. 1994. Inventory and Need Assessment of Irrigation Systems
in North-Eastern Tanahu. (Vol. I). Chitwan, Rampur, Nepal: Institute of Agriculture
and Animal Science, Irrigation Management Systems Study Group.
Pradhan, Naresh C., and Robert Yoder. 1989. “Improving Irrigation System Management
through Farmer-to-Farmer Training: Examples from Nepal.” IIMI
Working Paper no. 12. Colombo, Sri Lanka: International Irrigation Management
Institute.
Pradhan, Prachanda. 1989. “Patterns of Irrigation Organization in Nepal: A
Comparative Study of 21 Farmer-Managed Irrigation Systems.” Kathmandu,
Nepal: International Irrigation Management Institute.
Price, Michael E. In press. “Monitoring, Reputation and ‘Greenbeard’ Reciprocity
in a Shuar Work Team.” Journal of Organizational Behavior.
Pritchett, Lant, and Michael Woolcock. 2003. “Solutions When the Solution Is
the Problem: Arraying the Disarray in Development.” World Development
35(3):435–61.
Rabin, Matthew. 1993. “Incorporating Fairness in Game Theory and Economics.”
American Economic Review 83:1281–302.
. 1998. “Psychology and Economics.” Journal of Economic Literature 36
(March): 11–46.
Raiffa, Howard. 1982. The Art and Science of Negotiation. Cambridge: Harvard
University Press.
Rapoport, Anatol. 1966. Two-Person Game Theory: The Essential Ideas. Ann
Arbor: University of Michigan Press.
Rappaport, Roy A. 1979. Ecology, Meaning, and Religion. Richmond, CA: North
Atlantic Books.
Rasmusen, Eric. 1989. Games and Information: An Introduction to Game Theory.
Oxford: Basil Blackwell.
Raub, Werner, and Thomas Voss. 1986. “Conditions for Cooperation in Problematic
Social Situations.” In Paradoxical Effects of Social Behavior: Essays in
Honor of Anatol Rapoport, ed. Andreas Diekmann and Peter Mitter, 85–104.
Heidelberg: Physica-Verlag.
Rawls, John. 1971. A Theory of Justice. Cambridge: Harvard University Press.
  
R EF ER EN CE S340
Ray, Isha, and Jeffrey Williams. 1999. “Evaluation of Price Policy in the
Presence of Water Theft.” American Journal of Agricultural Economics
81(4):928–41.
Repetto, Fabia´n. 2002. “Plan Solidaridad: Notas Sobre la Incapacidad del Estado
Nacional para Gerenciar un Programa Complejo.” Working Paper no. 64. Buenos
Aires, Argentina: Fundacio´n Gobierno y Sociedad, Centro de Estudios para
el Desarrollo Institucional (CEDI).
Repetto, Robert. 1986. Skimming the Water: Rent-Seeking and the Performance
of Public Irrigation Systems. Research report no. 4. Washington, DC: World
Resources Institute.
Reuveny, Rafael, and John Maxwell. 2001. “Conflict and Renewable Resources.”
Journal of Conflict Resolution 45:719–42.
Reynolds, Andrew. 2002. The Architecture of Democracy: Constitutional Design,
Conflict Management, and Democracy. Oxford: Oxford University Press.
Richerson, Peter J., and Robert Boyd. 2002. “Culture Is Part of Human Biology.”
In Probing Human Origins, ed. Morris Goodman and Anne Simon Moffat, 59–
86. Cambridge, MA: American Academy of Arts and Sciences.
Richerson, Peter J., Robert Boyd, and Brian Paciotti. 2002. “An Evolutionary
Theory of Commons Management.” In The Drama of the Commons, National
Research Council, Committee on the Human Dimensions of Global Change,
ed. Elinor Ostrom, Thomas Dietz, Nives Dolsˇak, Paul C. Stern, Susan Stonich,
and Elke Weber, 403–42. Washington, DC: National Academy Press.
Rieskamp, Jo¨rg, and Gerd Gigerenzer. 2003. “Heuristics for Social Interactions:
How to Generate Trust and Fairness.” Working Paper. Berlin: Max Planck Institute
for Human Development.
Rieskamp, Jo¨rg, and Ulrich Hoffrage. 2003. “The Use of Simple Heuristics: Inferences
and Preferences under Time Pressure.” Working Paper. Berlin: Max
Planck Institute for Human Development.
Riker, William H. 1962. The Theory of Political Coalitions. New Haven: Yale
University Press.
Rilling, James K., David A. Gutman, Thorsten R. Zeh, Giuseppe Pagnoni, Gregory
S. Berns, and Clinton D. Kilts. 2002. “A Neural Basis for Social Cooperation.”
Neuron 35(2) (July): 395–405.
Riolo, Rick, Michael Cohen, and Robert Axelrod. 2001. “Evolution of Cooperation
without Reciprocity.” Nature 414 (November): 441–43.
Rocco, Elena, and Massimo Warglien. 1995. “Computer Mediated Communication
and the Emergence of ‘Electronic Opportunism.’ ” Venice, Italy: University
of Venice, Department of Economics, Laboratory of Experimental Economics.
Roconi, Lucas. 2002. “El Programa TRABAJAR.” Working Paper no. 63. Buenos
Aires, Argentina: Fundacio´n Gobierno y Sociedad, Centro de Estudios para el
Desarrollo Institucional (CEDI).
Rolett, Barry, and Jared Diamond. 2004. “Environmental Predictors of Pre-European
Deforestation on Pacific Islands.” Nature 431 (September 23): 443–46.
Romer, Thomas, and Howard Rosenthal. 1978. “Political Resource Allocation,
Controlled Agendas, and the Status Quo.” Public Choice 33(4):27–43.
Rose, Carol. 2002. “Common Property, Regulatory Property, and Environmental
Protection: Comparing Community-Based Management to Tradable Environ-
  
R EF ER EN CE S 341
mental Allowances.” In The Drama of the Commons, National Research Council,
Committee on the Human Dimensions of Global Change, ed. Elinor
Ostrom, Thomas Dietz, Nives Dolsˇak, Paul C. Stern, Susan Stonich, and Elke
Weber, 233–57. Washington, DC: National Academy Press.
Rothstein, Bo. 1998. Just Institutions Matter: The Moral and Political Logic of
the Universal Welfare State. Cambridge: Cambridge University Press.
. 2005. “Is Political Science Producing Technically Competent Barbarians?”
European Political Science 4:3–13.
. 2005. Social Traps and the Problem of Trust. Cambridge: Cambridge
University Press.
Rowe, Nicholas. 1989. Rules and Institutions. Ann Arbor: University of Michigan
Press.
Sally, David. 1995. “Conversation and Cooperation in Social Dilemmas: A Metaanalysis
of Experiments from 1958 to 1992.” Rationality and Society 7:58–92.
Sawyer, Amos. 1992. The Emergence of Autocracy in Liberia: Tragedy and Challenge.
Oakland, CA: ICS Press.
. 2005. Beyond Plunder: Toward Democratic Governance in Liberia. Boulder,
CO: Lynne Rienner Publishers.
Schaaf, Jeanne. 1989. “Governing a Monopoly Market under Siege: Using Institutional
Analysis to Understand Competitive Entry into Telecommunications
Markets, 1944–1982.” Ph.D. diss., Indiana University.
Schank, Roger C., and Robert P. Abelson. 1977. Scripts, Plans, Goals, and Understanding:
An Inquiry in Human Knowledge Structures. Hillsdale, NJ: Lawrence
Erlbaum Associates.
Scharpf, Fritz W. 1997. Games Real Actors Play: Actor-Centered Institutionalism
in Policy Research. Boulder, CO: Westview Press.
Schelling, Thomas C. 1978. Micromotives and Macrobehavior. New York:
W. W. Norton.
Schiemann, John W. 2002. “History and Emotions, Beliefs and Mental Models:
Toward a Hermeneutics of Rational Choice.” Paper presented at the American
Political Science Association meeting, Boston, 29 August–1 September.
Schlager, Edella. 1990. “Model Specification and Policy Analysis: The Governance
of Coastal Fisheries.” Ph.D. diss., Indiana University.
. 1994. “Fishers’ Institutional Responses to Common-Pool Resource Dilemmas.”
In Rules, Games, and Common-Pool Resources, ed. Elinor Ostrom,
Roy Gardner, and James Walker, 247–66. Ann Arbor: University of Michigan
Press.
. 1999. “A Comparison of Frameworks, Theories, and Models of Policy
Processes.” In Theories of the Policy Process, ed. Paul A. Sabatier, 233–60.
Boulder, CO: Westview Press.
. 2004. “Common-Pool Resource Theory.” In Environmental Governance
Reconsidered: Challenges, Choices, and Opportunities, ed. Robert Durant,
Daniel Fiorino, and Rosemary O’Leary, 145–75. Cambridge: MIT Press.
Schlager, Edella, William Blomquist, and Shui Yan Tang. 1994. “Mobile Flows,
Storage, and Self-Organized Institutions for Governing Common-Pool Resources.”
Land Economics 70(3) (August): 294–317.
  
R EF ER EN CE S342
Schlager, Edella, and Elinor Ostrom. 1992. “Property-Rights Regimes and
Natural Resources: A Conceptual Analysis.” Land Economics 68(3) (August):
249–62.
. 1993. “Property-Rights Regimes and Coastal Fisheries: An Empirical
Analysis.” In The Political Economy of Customs and Culture: Informal Solutions
to the Commons Problem, ed. Terry L. Anderson and Randy T. Simmons,
13–41. Lanham, MD: Rowman and Littlefield.
Schmid, Allan. 1999. “Government, Property, Markets . . . In That Order . . . Not
Government versus Markets.” In The Fundamental Interrelationships between
Government and Property, ed. Nicholas Mercuro and Warren Samuels, 233–
37. Stamford, CT: JAI Press.
. 2004. Conflict and Cooperation: Institutional and Behavioral Economics.
Malden, MA: Blackwell.
Schmitt, Pamela, Kurtis Swope, and James Walker. 2000. “Collective Action with
Incomplete Commitment: Experimental Evidence.” Southern Economic Journal
66(4):829–54.
Schotter, Andrew. 1981. The Economic Theory of Social Institutions. Cambridge:
Cambridge University Press.
Schuessler, Rudolf. 1989. “Exit Threats and Cooperation under Anonymity.”
Journal of Conflict Resolution 33(4):728–49.
Schweik, Charles M. 2000. “Optimal Foraging, Institutions, and Forest Change:
A Case from Nepal.” In People and Forests: Communities, Institutions, and
Governance, ed. Clark Gibson, Margaret McKean, and Elinor Ostrom, 57–85.
Cambridge: MIT Press.
Schweik, Charles M., Tom Evans, and J. Morgan Grove. Forthcoming. “Open
Source and Open Content: A Framework for the Development of Socio-Ecological
Research.” Ecology and Society.
Schweik, Charles M., and Andrei Semenov. 2003. “The Institutional Design of
‘Open Source’ Programming: Implications for Addressing Complex Public
Policy and Management Problems.” First Monday 8(1). <http://www.first
monday.org/issues/issue8_1/schweik/>
Scott, Anthony D. 1955. “The Fishery: The Objectives of Sole Ownership.”
Journal of Political Economy 63:116–24.
. 1993. “Obstacles to Fishery Self-Government.” Marine Resource
Economics 8:187–99.
. Forthcoming. The Evolution of Resource Rights. Oxford: Oxford University
Press.
Seabright, Paul. 1993. “Managing Local Commons: Theoretical Issues in Incentive
Design.” Journal of Economic Perspectives 7(4):113–34.
Searle, John. 1969. Speech Acts: An Essay in the Philosophy of Language. New
York: Cambridge University Press.
Seixas, Cristiana, and Fikret Berkes. 2003. “Dynamics of Social-Ecological
Changes in a Lagoon Fishery in Southern Brazil.” In Navigating Social-Ecological
Systems, ed. Fikret Berkes, Johan Colding, and Carl Folke, 271–98. Cambridge:
Cambridge University Press.
  
R EF ER EN CE S 343
Sekher, Madhushree. 2000. “Local Organisations and Participatory CPR Management:
Some Reflections.” Working Paper no. 61. Bangalore, India: Institute
for Social and Economic Change.
Selten, Reinhard. 1998. “Features of Experimentally Observed Bounded Rationality.”
European Economic Review 42:413–36.
Selten, Reinhard, Klaus Abbink, and Ricarda Cox. 2001. “Learning Direction
Theory and the Winners Curse.” Discussion paper no. 10/2001. Bonn, Germany:
University of Bonn, Department of Economics.
Sen, Amartya K. 1977. “Rational Fools: A Critique of the Behavioral Foundations
of Economic Theory.” Philosophy and Public Affairs 6(4):317–44.
Sengupta, Nirmal. 1991. Managing Common Property: Irrigation in India and
the Philippines. New Delhi: Sage.
Sethi, Rajiv. 1996. “Evolutionary Stability and Social Norms.” Journal of Economic
Behavior and Organization 29(1):113–40.
Shepherd, Gill. 1992. Managing Africa’s Tropical Dry Forests: A Review of Indigenous
Methods. London: Overseas Development Institute.
Shepsle, Kenneth A. 1975. “Congressional Committee Assignments: An Optimization
Model with Institutional Constraints.” Public Choice 22:55–78.
. 1979a. “Institutional Arrangements and Equilibrium in Multidimensional
Voting Models.” American Journal of Political Science 23:27–59.
. 1979b. “The Role of Institutional Structure in the Creation of Policy Equilibrium.”
In Public Policy and Public Choice, ed. Douglas W. Rae and Theodore
J. Eismeier, 249–83. Sage Yearbooks in Politics and Public Policy, vol. 6. Beverly
Hills, CA: Sage.
. 1989. “Studying Institutions: Some Lessons from the Rational Choice
Approach.” Journal of Theoretical Politics 1(2) (April): 131–47.
Shepsle, Kenneth A., and Barry R. Weingast. 1984. “When Do Rules of Procedure
Matter?” Journal of Politics 46:206–21.
. 1987. “The Institutional Foundations of Committee Power.” American
Political Science Review 81 (March): 85–104.
Sherman, K., and T. Laughlin, eds. 1992. NOAA Technical Memorandum NMFSF/NEC
91. The Large Marine Ecosystem (LME) Concept and Its Application
to Regional Marine Resource Management. Woods Hole, MA: Northeast Fisheries
Science Center.
Shimanoff, Susan B. 1980. Communication Rules: Theory and Research. Beverly
Hills, CA: Sage.
Shivakoti, Ganesh, and Elinor Ostrom, eds. 2002. Improving Irrigation Governance
and Management in Nepal. Oakland, CA: ICS Press.
Shivakumar, Sujai J. 2005. The Constitution of Development: Crafting Capabilities
for Self-Governance. New York: Palgrave Macmillan.
Shubik, Martin. 1986. “The Games within the Game: Modeling PoliticoEconomic
Structures.” In Guidance, Control, and Evaluation in the Public Sector,
ed. Franz-Xaver Kaufmann, Giandomenico Majone, and Vincent Ostrom,
571–91. Berlin and New York: Walter de Gruyter.
Shukla, Ashutosh K. 2002. “Policies, Processes, and Performance of Management:
Turnover and Agency-Initiated Interventions.” In Improving Irrigation
  
R EF ER EN CE S344
Governance and Management in Nepal, ed. Ganesh Shivakoti and Elinor Ostrom,
71–102. Oakland, CA: ICS Press.
Shukla, Ashutosh K., Kishor Gajurel, Ganesh Shivakoti, Rabi Poudel, Kala N.
Pandit, Keshav R. Adhikari, Tej Thapa, S. M. Shakya, D. N. Yadav, Neeraj R.
Joshi, and A. P. Shrestha. 1993. Irrigation Resource Inventory of East Chitwan.
Rampur, Chitwan, Nepal: Institute of Agriculture and Animal Science, Irrigation
Management Systems Study Group.
Sieber, Sam D. 1981. Fatal Remedies: The Ironies of Social Intervention. New
York: Plenum Press.
Simon, Adam, and Heather Gorgura. 2003. “Say the Magic Word: Effective
Communication in Social Dilemmas.” Working Paper. Seattle: University of
Washington, Department of Political Science.
Simon, Herbert A. 1955. “A Behavioural Model of Rational Choice.” Quarterly
Journal of Economics 69:99–188.
. 1957. Models of Man. New York: Wiley.
. 1972. “Theories of Bounded Rationality.” In Decision and Organization:
A Volume in Honor of Jacob Marschak, ed. C. B. McGuire and Roy Radner,
161–76. Amsterdam: North Holland.
. 1981. The Sciences of the Artificial. 2d ed. Cambridge: MIT Press.
. 1995. “Near Decomposability and Complexity: How a Mind Resides
in a Brain.” In The Mind, the Brain, and Complex Adaptive Systems, ed.
Harold J. Morowitz and Jerome L. Singer, 25–44. Reading, MA:
Addison-Wesley.
. 1999. “The Potlatch between Political Science and Economics.” In Competition
and Cooperation: Conversations with Nobelists about Economics and
Political Science, ed. James Alt, Margaret Levi, and Elinor Ostrom, 112–9. New
York: Russell Sage Foundation.
Singleton, Sara. 1998. Constructing Cooperation: The Evolution of Institutions
of Comanagement in Pacific Northwest Salmon Fisheries. Ann Arbor: University
of Michigan Press.
Singleton, Sara, and Michael Taylor. 1992. “Common Property Economics: A
General Theory and Land Use Applications.” Journal of Theoretical Politics
4(3):309–24.
Skyrms, Brian. 1997. “Chaos and the Explanatory Significance of Equilibrium:
Strange Attractors in Evolutionary Game Dynamics.” In The Dynamics of
Norms, ed. Cristina Bicchieri, Richard Jeffrey, and Brian Skyrms, 199–222.
Cambridge: Cambridge University Press.
Smith, Robert J. 1981. “Resolving the Tragedy of the Commons by Creating
Private Property Rights in Wildlife.” CATO Journal 1:439–68.
Smith, Vernon L. 1982. “Microeconomic Systems as an Experimental Science.”
American Economic Review 72 (December): 923–55.
. 1991. Papers in Experimental Economics. New York: Cambridge University
Press.
. 2000. Essays in Experimental Economics. New York: Cambridge University
Press.
. 2001. “Mind, Reciprocity, and Markets in the Laboratory.” Wirtschaft
(August 10): 21 (English version of German translation).
  
R EF ER EN CE S 345
Spooner, Brian. 1974. “Irrigation and Society: The Iranian Plateau.” In Irrigation’s
Impact on Society, ed. Theodore E. Downing and McGuire Gibson,
43–57. Tucson: University of Arizona Press.
Sproule-Jones, Mark. 1993. Governments at Work: Canadian Parliamentary Federalism
and Its Public Policy Effects. Toronto: University of Toronto Press.
Stevens, Stan. 1997. “New Alliances for Conservation.” In Conservation through
Cultural Survival: Indigenous Peoples and Protected Areas, ed. Stan Stevens,
33–62. Washington, DC: Island Press.
Straffin, Philip D. 1977. “Majority Rule and General Decision Rules.” Theory
and Decision 8:351–60. Reprinted in Rational Man and Irrational Society?: An
Introduction and Sourcebook, ed. Brian Barry and Russell Hardin, 316–24.
Beverly Hills, CA: Sage, 1982.
Sugden, Robert. 1986. The Economics of Rights, Co-operation and Welfare.
Oxford: Basil Blackwell.
Sugrue, Leo P., Greg S. Corrado, and William T. Newsome. 2004. “Matching
Behavior and the Representation of Value in the Parietal Cortex.” Science 304
(June 18): 1782–87.
Sundquist, James L. 1968. Politics and Policy: The Eisenhower, Kennedy, and
Johnson Years. Washington, DC: Brookings Institution.
Sussman, Robert W., and Audrey R. Chapman, eds. 2004. The Origins and
Nature of Sociality. New York: Walter de Gruyter.
Tang, Shui Yan. 1992. Institutions and Collective Action: Self-Governance in
Irrigation. Oakland, CA: ICS Press.
. 1994. “Institutions and Performance in Irrigation Systems.” In Rules,
Games, and Common-Pool Resources, ed. Elinor Ostrom, Roy Gardner, and
James Walker, 225–45. Ann Arbor: University of Michigan Press.
Taylor, Michael. 1982. Community, Anarchy, and Liberty. New York: Cambridge
University Press.
. 1987. The Possibility of Cooperation. New York: Cambridge University
Press.
Tendler, Judith. 2000. “Why Are Social Funds So Popular?” In Local Dynamics
in an Era of Globalization: 21st-Century Catalysts for Development, ed. Shahid
Yusuf, Weiping Wu, and Simon Evenett, 114–29. Oxford: Oxford University
Press.
Tesfatsion, Leigh. 2002. “Agent-Based Computational Economics: Growing
Economies from the Bottom Up.” ISU Economics Working Paper no. 1. Ames:
Iowa State University.
Theesfeld, Insa. 2004. “Constraints on Collective Action in a Transitional
Economy: The Case of Bulgaria’s Irrigation Sector.” World Development
32(2):251–71.
Thompson, Leigh L., Elizabeth A. Mannix, and Max H. Bazerman. 1988. “Group
Negotiation: Effects of Decision Rule, Agenda, and Aspiration.” Journal of
Personality and Social Psychology 54(1):86–95.
Thomson, James T. 1977. “Ecological Deterioration: Local-Level Rule Making
and Enforcement Problems in Niger.” In Desertification: Environmental Degradation
in and around Arid Lands, ed. Michael H. Glantz, 57–79. Boulder, CO:
Westview Press.
  
R EF ER EN CE S346
Thomson, James T., David Feeny, and Ronald J. Oakerson. 1992. “Institutional
Dynamics: The Evolution and Dissolution of Common-Property Resource
Management.” In Making the Commons Work: Theory, Practice, and Policy,
ed. Daniel W. Bromley et al., 129–60. Oakland, CA: ICS Press.
Tietenberg, Tom. 2002. “The Tradable Permits Approach to Protecting the Commons:
What Have We Learned?” In The Drama of the Commons, National
Research Council, Committee on the Human Dimensions of Global Change,
ed. Elinor Ostrom, Thomas Dietz, Nives Dolsˇak, Paul C. Stern, Susan Stonich,
and Elke Weber, 197–232. Washington, DC: National Academy Press.
Tilman, David. 1999. “The Ecological Consequences of Changes in Biodiversity:
A Search for General Principles.” Ecology 80:1455–74.
Tilman, David, Clarence L. Lehman, and Charles E. Bristow. 1998. “DiversityStability
Relationships: Statistical Inevitability or Ecological Consequences?”
American Naturalist 151:277–82.
Tomasello, Michael. 2004. “Learning through Others.” Daedalus 133(1):51–58.
Toulmin, Stephen. 1974. “Rules and Their Relevance for Understanding Human
Behavior.” In Understanding Other Persons, ed. Theodore Mischel, 185–215.
Oxford: Basil Blackwell.
Trawick, Paul B. 2001. “Successfully Governing the Commons: Principles of
Social Organization in an Andean Irrigation System.” Human Ecology
29(1):1–25.
Trosper, Ronald L. 2002. “Northwest Coast Indigenous Institutions That Supported
Resilience and Sustainability.” Ecological Economics 41(2):329–44.
Tsebelis, George. 1989. “The Abuse of Probability in Political Analysis: The
Robinson Crusoe Fallacy.” American Political Science Review 83:77–91.
. 1991. “The Effect of Fines on Regulated Industries: Game Theory versus
Decision Theory.” Journal of Theoretical Politics 3(1):81–101.
. 2002. Veto Players: How Political Institutions Work. Princeton:
Princeton University Press.
Tucker, Catherine M. 1999. “Common Property Design Principles and Development
in a Honduran Community.” Praxis: The Fletcher Journal of Development
Studies 15:47–76.
Tversky, Amos. 1972. “Elimination by Aspects: A Theory of Choice.” Psychological
Review 79:281–99.
Tyler, Tom R. 1990. Why People Obey the Law. New Haven: Yale University
Press.
Ude´hn, Lars. 1993. “Twenty-five Years with The Logic of Collective Action.”
Acta Sociologica 36:239–61.
Ullmann-Margalit, Edna. 1977. The Emergence of Norms. Oxford: Oxford University
Press.
Uphoff, Norman T. 1986. Getting the Process Right: Improving Irrigation Water
Management with Farmer Participation. Boulder, CO: Westview Press.
Valenzuela, Arturo. 1993. “Latin America: Presidentialism in Crisis.” Journal of
Democracy 4:3–16.
Vanberg, Viktor J. 2002. “Rational Choice versus Program-Based Behavior: Alternative
Theoretical Approaches and Their Relevance for the Study of Institutions.”
Rationality and Society 14(1):7–54.
  
R EF ER EN CE S 347
Vanberg, Viktor J., and Roger D. Congleton. 1992. “Rationality, Morality, and
Exit.” American Political Science Review 82(2):418–31.
Van Vugt, Mark, Mark Snyder, Tom R. Tyler, and Anders Biel, eds. 2000. Cooperation
in Modern Society: Promoting the Welfare of Communities, Organizations,
and States. London: Routledge.
Van Wendel de Joode, Ruben. 2004. “Continuity of the Commons in Open Source
Communities.” Paper presented at the Symposium of Santa Caterina on Challenges
in the Internet and Interdisciplinary Research, Amalfi, Italy.
Varughese, George. 2000. “Population and Forest Dynamics in the Hills of Nepal:
Institutional Remedies by Rural Communities.” In People and Forests: Communities,
Institutions, and Governance, ed. Clark Gibson, Margaret McKean,
and Elinor Ostrom, 193–226. Cambridge: MIT Press.
Varughese, George, and Elinor Ostrom. 2001. “The Contested Role of Heterogeneity
in Collective Action: Some Evidence from Community Forestry in Nepal.”
World Development 29(5) (May): 747–65.
Vermillion, Douglas. 2001. “Property Rights and Collective Action in the Devolution
of Irrigation System Management.” In Collective Action, Property Rights,
and Devolution of Natural Resource Management, ed. Ruth Meinzen-Dick,
Anna Knox, and Monica DiGregorio, 183–220. Feldafing, Germany: Deutsche
Stiftung fu¨r internationale Entwicklung.
Vogt, Nathan, Abwoli Banana, William Gombya-Ssembajjwe, and Joseph Bahati.
2005. “Understanding the Long Term Stability of the West Mengo Forest Reserve
Boundaries.” Working paper. Bloomington: Indiana University, Center for
the Study of Institutions, Population, and Environmental Change (CIPEC).
von Wright, Georg Henrik. 1951. “Deontic Logic.” Mind 60:1–15.
. 1966. “The Logic of Action: A Sketch.” In The Logic of Decision and
Action, ed. Nicholas Rescher, 121–36. Pittsburgh, PA: University of Pittsburgh
Press.
. 1968. “The Logic of Practical Discourse.” In Contemporary Philosophy,
ed. Raymond Klikensky, 141–67. Italy: La Nuava Italia Editrice.
Wade, Robert. 1994. Village Republics: Economic Conditions for Collective Action
in South India. San Francisco: ICS Press.
Wagner, Richard E. In press. “Self-Governance, Polycentrism, and Federalism:
Recurring Themes in Vincent Ostrom’s Scholarly Oeuvre.” Journal of Economic
Behavior and Organization 57(2).
Walker, S. Tjip. 1998. “Both Pretense and Promise: The Political Economy of
Privatization in Africa.” Ph.D. diss., Indiana University.
Water and Energy Commission Secretariat, Nepal, and International Irrigation
Management Institute. 1990. Assistance to Farmer-Managed Irrigation Systems:
Results, Lessons, and Recommendations from an Action-Research Project.
Colombo, Sri Lanka: IIMI.
Weber, Max. 1947. The Theory of Social and Economic Organization. Translated
by A. M. Henderson and Talcott Parsons. New York: Free Press.
Weimer, David L., and Aidan R. Vining. 1992. Policy Analysis: Concepts and
Practice. 2d ed. Englewood Cliffs, NJ: Prentice Hall.
Weingast, Barry R. 1989. “Floor Behavior in the U.S. Congress: Committee Power
under the Open Rule.” American Political Science Review 78:417–34.
  
R EF ER EN CE S348
Weinstein, Martin S. 2000. “Pieces of the Puzzle: Solutions for Community-Based
Fisheries Management from Native Canadians, Japanese Cooperatives, and
Common Property Researchers.” Georgetown International Environmental
Law Review 12(2):375–412.
Weissing, Franz J., and Elinor Ostrom. 1991a. “Irrigation Institutions and the
Games Irrigators Play: Rule Enforcement without Guards.” In Game Equilibrium
Models II: Methods, Morals, and Markets, ed. Reinhard Selten, 188–262.
Berlin: Springer-Verlag.
. 1991b. “Crime and Punishment.” Journal of Theoretical Politics
3(3):343–9.
. 1993. “Irrigation Institutions and the Games Irrigators Play: Rule Enforcement
on Government- and Farmer-Managed Systems.” In Games in Hierarchies
and Networks: Analytical and Empirical Approaches to the Study of
Governance Institutions, ed. Fritz W. Scharpf, 387–428. Frankfurt am Main:
Campus Verlag; Boulder, CO: Westview Press.
Wells, Michael, and Katrina Brandon. 1992. People and Parks: Linking Protected
Area Management with Local Communities. Washington, DC: World Bank.
Wiggins, Steven N., and Gary D. Libecap. 1987. “Firm Heterogeneities and Cartelization
Efforts in Domestic Crude Oil.” Journal of Law, Economics, and
Organization 3:1–25.
Williamson, Oliver E. 1975. Markets and Hierarchies: Analysis and Antitrust Implications.
New York: Free Press.
. 1985. The Economic Institutions of Capitalism. New York: Free Press.
. 2000. “The New Institutional Economics: Take Stock, Looking Ahead.”
Journal of Economic Literature 38 (September): 593–613.
Wilson, Douglas, Jesper Nielsen, and Poul Degnbol, eds. 2003. The Fisheries
Co-management Experience: Accomplishments, Challenges, and Prospects.
Dordrecht, the Netherlands: Kluwer.
Wilson, James A. 1990. “Fishing for Knowledge.” Land Economics 66:12–29.
. 1997. “Maine Fisheries Management Initiative.” In The Social Impacts
of Individual Transferable Quotas, ed. Gisli Palsson, 335–53. Copenhagen:
TemaNord.
. 2002. “Scientific Uncertainty, Complex Systems, and the Design of Common-Pool
Institutions.” In The Drama of the Commons, National Research
Council, Committee on the Human Dimensions of Global Change, ed. Elinor
Ostrom, Thomas Dietz, Nives Dolsˇak, Paul C. Stern, Susan Stonich, and Elke
Weber, 327–59. Washington, DC: National Academy Press.
Wilson, James A., James M. Acheson, Mark Metcalfe, and Peter Kleban. 1994.
“Chaos, Complexity, and Community Management of Fisheries.” Marine Policy
18:291–305.
Wilson, James A., Bobbi Low, Robert Costanza, and Elinor Ostrom. 1999. “Scale
Misperceptions and the Spatial Dynamics of a Social-Ecological System.” Ecological
Economics 31(2) (November): 243–57.
Wilson, Paul N., and Gary D. Thompson. 1993. “Common Property and Uncertainty:
Compensating Coalitions by Mexico’s Pastoral Ejidatorios.” Economic
Development and Cultural Change 41(2):299–318.
  
R EF ER EN CE S 349
World Bank. 2002. Social Funds—Assessing Effectiveness. Washington, DC:
World Bank Operations Evaluation Department (OED).
Wynne, Susan G. 1989. “The Land Boards of Botswana: A Problem in Institutional
Design.” Ph.D. diss., Indiana University.
Yandle, Tracy. 2001. “Market-Based Natural Resource Management: An Institutional
Analysis of Individual Tradable Quotas in New Zealand’s Commercial
Fisheries.” Ph.D. diss., Indiana University.
Yandle, Tracy, and Christopher Dewees. 2003. “Privatizing the Commons . . .
Twelve Years Later: Fishers’ Experiences with New Zealand’s Market-Based
Fisheries Management.” In The Commons in the New Millennium: Challenges
and Adaptations, ed. Nives Dolsˇak and Elinor Ostrom, 101–27. Cambridge:
MIT Press.
Yoder, Robert D. 1991. “Peer Training as a Way to Motivate Institutional Change
in Farmer-Managed Irrigation Systems.” In Proceedings of the Workshop on
Democracy and Governance. Decentralization: Finance and Management Project
Report, 53–67. Burlington, VT: Associates in Rural Development.
. 1994. Locally Managed Irrigation Systems. Colombo, Sri Lanka: International
Irrigation Management Institute.
Young, H. Peyton. 1998. Individual Strategy and Social Structure: An Evolutionary
Theory of Institutions. Princeton: Princeton University Press.
Young, Oran R. 1997. Global Governance: Drawing Insights from the Environmental
Experience. Cambridge: MIT Press.
. 2002. The Institutional Dimensions of Environmental Change: Fit, Interplay,
and Scale. Cambridge: MIT Press.
  
  
Index
Abernathy, Charles L., 260 239; residency-based, 225; technologybased,
225Acheson, James, 230, 236, 284–85
action arena, 13–16, 55 Brander, James A., 249
Brox, Ottar, 273action situation, 32; potential outcomes of,
42; participants in, 37–40; opportunity Buchan, Nancy R., 74–75
Butler, Christopher K., 10in, 43; repeated interaction in, 53; organizational
linkage of, 55
actions, 45 Cadenasso, Mary L., 12
Cain, Michael, 129Actor-Centered Institutionalism framework,
29 Camerer, Colin, 124–25
Cardenas, Juan-Camilo, 93–94ADICO, 140–41; in repeated PD game,
158; and field research, 171–74 Carson, Richard T., 237
Casari, Marco, 95agenda control, 201
aggregation rules, 202–6; nonsymmetric, case studies, 35
centralized governance, 237–40, 278–79,202–3; symmetric, 203–4; lack of
agreement, 205 283–84
cheap talk, 85. See also communication,Agrawal, Arun, 227, 234, 251
AIM, 140, 148–49, 186–87 game theory
Choe, Jaesong, 269–70Alchian, Armen, 101
alcoholic beverages, 182–83 choice rules, 200–2, 228–33
Coglianese, Cary, 209–11Alcorn, Janis B., 284
altruism, 110 Coleman, James, 9, 146
collective choice, 33, 57, 63; rules of, 58–ATTRIBUTES, 139, 141; of goods, 23–26;
of community, 15, 26; of biophysical 60; examples of, 60. See also levels of
analysisworld, 26; of participants, 40; of resources,
247–48; of appropriators, common-pool resource (CPR), 79–80,
219–54, 280; attributes of, 23–25; pub-248–54; rapid change of, 272–73
lic provision of, 24; dilemma, 78–97;
open access, 80; and generic rules, 183;babbling equilibrium. See language
Banana, Abwoli Y., 265 and field research, 221–22; threats to,
271–79Bardhan, Pranab, 264
Bates, Robert, 277–78 Commons, John R., 20, 144–45, 187
communication, 64–65, 98; in Trust Game,Becker, C. Dustin, 268–69
Berg, Joyce, 72, 77 74–75; in Commons Dilemma, 85; faceto-face,
85–87, 93; among heteroge-Berkes, Fikret, 248, 263, 268
biophysical world, 15, 22–24, 26, 42 neous participants, 87–89; lack of, 89;
among Columbian villagers, 107; andBlack, Max, 16–17
Blomquist, William, 235, 237 mental models, 107; in PD game, 130;
intergenerational, 273blueprint thinking, 274–77
Blurton Jones, Nicholas G., 49 community, 15, 26, 274; governance, 220–
1, 243–45, 279–83boundary rules, 194–200; first order, 195;
second order, 195–96; and invitational complexity, 9–10, 13, 242–43, 256–57,
286entry, 196; and competitive entry, 196;
and compulsory entry, 196; and succes- CONDITIONS, 140, 144, 149, 171
constitutional choice, 33, 58, 61. See alsosion rules, 198; and exit rules, 198–99;
and common-pool resources, 223–26, levels of analysis
  
I ND EX352
contingent action situation, 44 124; and ADICO, 153–65; strategy, 156;
grim trigger strategy, 157contracting, 215
control, 49–50 games: bridge, 40; chess, 34–35; Snatch,
35–36; 5-person county council, 46;control variable, 46
corruption, 277 Trust Game, 71–78; commons dilemma,
78–97; Snatch with norms, 122–23;Cooter, Robert, 166
Crawford, Sue, 121, 137–74, 186–215 Prisoner’s Dilemma (PD), 129–30,
147–48, 154–55; repeated PD, 56–58;Croson, Rachel T., 74–75
culture, 27, 126–27 PD with monitor, monitoring norms,
and sanctions, 162–65; “urban taxi,”
191; Snatch with private property,Deadman, Peter, 84
default conditions, 182, 205, 210–14 211–14
Gardner, Roy, 82, 86, 92–93, 207de los Reyes, Romana, 228
delta parameter, 121–22, 146, 153. See Gibbard, Allan, 108
Gibson, Clark, 266, 268–69also norms
Denzau, Arthur T., 105, 107, 109 Gigerenzer, Gerd, 114
Gode, Dhananjay K., 101DEONTIC, 126, 140, 141–42, 172; operators,
142–44; in formal games, 146 Gombya-Ssembajjwe, William, 265
goods, excludability of, 24–25; sub-design principles, 259; and well-defined
boundaries, 260–62; and equivalence of tractability of, 25–26
Grainger, Alan, 237costs/benefits, 262–63; and collectivechoice
arrangements, 264–65; and moni- Grether, David M., 205, 240
grim trigger strategy, 113–14, 157toring, 265–66; and graduated sanctions,
266–67; and conflict-resolution Gu¨th, Werner, 75, 128
mechanism, 267–68; and recognition of
rights, 268–69; and nested enterprises, Hackett, Steven, 87
Haley, Sharman, 260269–70; in research, 270–71
Diamond, Jared, 248–49 Hammond, Thomas H., 10
Hardin, Garrett, 64Dickhaut, John, 72, 75, 77
duties, 145 Hawkes, Kristen, 49
Hechter, Michael, 179
Heiner, Ronald, 209Ebbin, Syma A., 286
evolutionary theory, 129. See also game heuristics, 98, 113–16; in Commons
Dilemma, 83; and grim trigger, 113–theory
exposure, 145 14; models of, 115
Hoffrage, Ulrich, 114–16
holons, 11–12, 103. See also levels ofFehr, Ernst, 72–73, 76
fisheries, 225–26, 227, 230, 238, 248, 273. analysis
Hull, Rebecca Watts, 262See also Maine lobster fishery, Pacific
salmon fishery
fMRI, 111 information, 50–51, 102, 104–9
information rules, 206–7foreign donors, 276–78
Forsyth, Timothy, 222 Ingram, Helen M., 237
Institutional Analysis and DevelopmentFrey, Bruno, 77, 95, 112–13
Frohlich, Norman, 100, 102, 103, 107–8, (IAD) framework, 6–7, 9, 27–29
institutional statement, 137, 178. See also130
norms, rules, strategies
institutions, 3, 179; diversity of, 4, 181–game theory, 6–7; decision point, 45; folk
theorem, 54–55; Tit for Tat, 54; inconsis- 82; commonalities of, 5–6, 30, 185; and
human behavior, 118. See also rulestencies of, 98; evolutionary, 116, 125–
31; and norms, 120–25; noncooperative, irrigation, 226, 227–28, 230–33, 251, 261
  
I ND EX 353
Irrigation Management Systems Study models, 28; of human behavior, 124; of
rules, 240–41Group (IMSSG), 231–33
Isaac, R. Mark, 205, 240–41 monitor, 158, 162, 227
monitoring, 265–66; of rules, 21; and
voter turnout, 168Jager, Wander, 84, 116
Janssen, Marco A., 84, 116 Morrow, Christopher E., 262
Moxnes, Erling, 238Jegen, Reto, 112–13
Johnson, Craig, 222
Johnson, Eric J., 74 Nalebuff, Barry, 202
Nash, Jennifer, 209–10Jones, Bryan, 109
niches, 124
norms, 112, 119, 123–24, 147; in socialKarotkin, Drora, 202–3
Kliemt, Hartmut, 128 dilemmas, 120; as payoffs, 121–22, 146;
evolution of, 123–27; no agreed defini-Knight, Frank, 170
Knox, Anna, 264 tion of, 178–79
North, Douglass C., 105, 107, 109, 179Koestler, Arthur, 11–12
Koford, Kenneth, 73–74, 102
Korten, David, 274–75 Ockenfels, Peter, 75
O’Connell, James F., 49Kreps, David, 147
Kurki, Anja, 102 Offe, Claus, 168
Oliver, Pamela, 10, 252
Olmstead, Todd, 209–10Lam, Wai Fung, 222, 251
language, 27, 126; and babbling equilib- Olson, Mancur, 253
operational choice, 34, 55; rules of, 58;rium, 176–80; in rules, 180–81; and
slogans, 181, 256 examples of, 60. See also levels
of analysislearning, 106, 128
levels of analysis, 56, 214–15; and IAD, operator. See DEONTIC
Opp, Karl-Dieter, 1798; and holons, 11–12; choosing levels in
research, 12, 61, 172; and nested rules, Oppenheimer, Joe A., 100, 102, 103,
107–8, 13058–60, 187; shifting among levels, 62–
64. See also operational choice, collec- OR ELSE, 140, 149–52. See also sanctions
Orbell, John, 41–42tive choice, constitutional choice
Levi, Margaret, 167, 267 organization, 57, 179
Ostrom, Elinor, 28, 29, 30, 82–84, 86, 92–Levin, Jonathan, 200
Lewis, David, 176–77 93, 121, 175, 207, 224–25, 241–42,
253, 257, 266Libecap, Gary, 248, 250, 251
liberty, 145, 170 Ostrom, Vincent, 20, 181
outcomes, 43–44, 64–65, 66, 233–36Lobe, Kenton, 268
local governance, 220–21, 243–45, 279–83
Pacific salmon fishery, 285–86
Paroush, Jacob, 202–3Maine lobster fisheries, 6, 248, 284–85
Margolis, Howard, 167–68 participants, 38–40; standing of, 41; assumptions
of 103–4, 117–18; informa-Marinova, Nadja, 237
Marwell, Gerald, 10, 252 tion process of, 104–9; valuation process
of, 109–13; selection process of, 113–16Maynard Smith, John, 126
Mayntz, Renate, 29 payoff rules, 42–43, 52, 233
permission rules, 144–46McAdams, Richard, 179
McCabe, Kevin, 72, 77 Pickett, Steward T. A., 12
Plott, Charles, 95, 182, 205, 240–41McGinnis, Michael, 45, 270
Meinzen-Dick, Ruth, 264 policy design, 240–43, 255–58, 274–77
polycentricity, 269–70, 281–86migration, 161
  
I ND EX354
positions, 40, 193–94, 197, 227–28 salience, 107
Sally, Hilmy, 260Poteete, Amy, 253
power, 50 sanctions, 92–93, 95–96; external, 77, 94–
95, 130–31; in Commons Dilemma, 79,principal-agent problem, 51
Pritchett, Lant, 275 91–92; in Italian Alps, 95–96; graduated,
152, 266–67; severity of, 152private property rights, 21–22, 29, 189–
90, 211–14, 237 Scharpf, Fritz, 10, 29, 39
Schiemann, John, 107public goods, 56, 23–25
Schlager, Edella, 87, 225–26, 230, 233
Schmitt, Pamela, 88quotas, 191–92, 230–31
Schwartz-Shea, Perry, 41–42
scientific management, 236–37; andRapoport, Anatol, 17–18
Rasmusen, Eric, 179 rational egoists, 238; and information
problems, 238; and central manage-rational egoist, 101, 127–28. See also
rationality ment, 240
Sekher, Madhushree, 264rationality, 101; and Trust Game, 78; in
competitive markets, 99–101; alternative self-organization, 220–21, 243–45; and
group size, 251–52; and group hetero-assumptions and, 103–4; in nonmarket
settings, 120 geneity, 252–53
Sen, Amartya, 111–12Ray, Isha, 264
reciprocity, 76–77, 112 Sengupta, Nirmal, 234
Shepsle, Kenneth, 258redundancy, 284
Repetto, Fabia´n, 275 Shukla, Ashutosh, 264
signaling, 128, 176–77reputation, 27
reversion level, 182. See also default Simmons, Randy T., 41–42
Simon, Herbert, 270–71conditions
Rieskamp, Jo¨rg, 114–16 Smith, Vernon, 69
social dilemma, 37, 79; and collectiverights, 144–46
Rilling, James K., 111 action, 119–21; and evolutionary cooperation,
127risk, 48
ritual, 108 spillovers, 249, 279
Steneck, Robert S., 230, 236, 284–85robustness, 243, 258–60, 271–79
Rocco, Elena, 93 Stranlund, John K., 94
strategies, 156, 178Rockenbach, Bettina, 76
Roconi, Lucas, 275 suboptimal solutions, 220, 236, 243, 254,
256, 270Rolett, Barry, 248
rules, 16–18; origin of, 19; in use, 19, 138, Sunder, Shyam, 101
Swope, Kurtis, 88166; predictability of, 20; enforcement
of, 21, 130–31; external, 130; changing, Szathma´ry, Eo¨rs, 126
138, 212–13, 215, 243–45, 245–47; generative
or regulatory, 138; legitimacy of, Tang, Shui Yan, 226, 228, 230–32, 233,
251167; lack of agreed definition of, 178; semantics
of, 180–81; variety of, 181–82; Taylor, M. Scott, 249
Theesfeld, Insa, 280and information transformation, 184–
85; types of, 190; interdependency of, theory, 28
Toledo, Victor, 284191; permission, 144–46; position, 193–
94, 227–28; boundary, 194–200, 223– transmission failures, 273–75
trust, 98, 74. See also games26; choice, 200–2, 228–33; aggregation,
202–6, 235–36; information, 206–7; pay- Tucker, Catherine M., 270
off, 207–8, 233–34; scope, 208–10; suboptimal
selection of, 220, 236, 243, 254, unanimity rule, 182, 203–4
uncertainty, 48–49256, 270; and CPR, 222–23
  
I ND EX 355
U.S. Geological Survey, 278–79 Wendel, Markus, 75
Wiesenthal, Helmut, 168utility, 53
Wiggins, Steven N., 248
Williams, Jeffrey, 264valuation, 43, 52; intrinsic, 53, 113, 119,
123–24, 146–47; process of, 109; hetero- Williams, John, 45, 266
Willis, Cleve E., 94geneity of, 123–24, 147
Viek, Charles A. J., 116 Wilson, James A., 230, 236, 284–85
Woolcock, Michael, 275vividness, 107–8
Vogt, Nathan, 262 Workshop in Political Theory and Policy
Analysis, 1, 34, 222von Wright, Georg Henrik, 142
World Bank, 275. See also foreign
donorsWade, Robert, 234
Walker, James, 82, 86, 87, 88, 92–93, 207
Warglien, Massimo, 93 Yadama, Gautam N., 227
Weinstein, Martin S., 260
Weissing, Franz J., 28 Zilberman, David, 237