C H A P T E R 1 4
WHAT CAUSES
DEMOCRATIZATION?
B A R B A R A G E D D E S
R E S E A R C H on democratization has become increasingly sophisticated during the last
decade. With the completion and sharing of new datasets and the ratcheting up of
training in statistics and modeling, approaches to studying democratization have
changed greatly since the mid-1990s. Economic models of democratization and
large-N statistical investigations of its causes play an ever larger role in its study.
What we think we know about democratization has changed much less, though we
have some intriguing new ideas to think about. Recent research has confirmed what
we thought we knew several decades ago: richer countries are more likely to be
democratic. Controversy continues about whether economic development increases
the likelihood of transitions to democracy. Przeworski and his co-authors (2000)
have argued emphatically that development does not cause democratization; rather,
development reduces the likelihood of democratic breakdown, thus increasing the
number of rich democratic countries even though it has no causal effect on transitions
to democracy. Other careful analyses of regime change, however, continue to
find a relationship between development and transitions to democracy (e.g., Boix
and Stokes 2003; Epstein et al. forthcoming).
Several other empirical regularities have achieved the status of stylized facts,
though all have also been challenged. Reliance on oil, and perhaps other mineral
exports, reduces the likelihood of democracy (Barro 1996; Ross 2001; Fish 2002).
Countries with large Muslim populations are less likely to be democratic (Fish 2002).
Weiner (1987) and Payne (1993) among others have suggested that British colonial
heritage contributes to better prospects for democracy later, and Barro (1996) finds
support for their claims.1
1
But Fish (2002) finds no relationship between British colonial heritage and democracy.
3l8 B A R B A R A G E D D E S
As with the relationship between development and democracy, controversy
continues about whether these are causal relationships or correlations explained by
something else. Among those who believe relationships are causal, there are disagreements
about the processes through which the causes produce the outcome. Middle
East experts explain the correlation between oil wealth and dictatorship as a consequence
of a rentier state that can use its rents from the sale of natural resources to
distribute subsidies to large parts of the population and thus to maintain popular
compliance with the regime (Anderson 1987; Crystal 1995). In a parallel argument,
Dunning (2006) argues that oil rents can in some circumstances be used to sustain
democracy. Herb (2005), however, shows that when a measure of development that
excludes the effect of oil on the economy is used in place of GDP per capita in
statistical analysis of the causes of democratization, oil-rich countries fit the same
patterns as other countries. The proxy measure of development has a strong positive
effect on changes in democracy scores, and rent dependence, measured separately,
has no effect. In short, he challenges the existence of a relationship between oil wealth
and regime type. Some observers have suggested an affinity between Muslim doctrine
or the attitudes of believers and authoritarianism, but Fish (2002) suggests that
Muslim countries tend to be authoritarian not for the reasons usually mentioned
but because of the suppression of women's rights in these countries.
In 1959, Seymour Martin Lipset argued that modernization caused democracy. He
supported his claim with what was then a state-of-the-art quantitative test, a table
showing a relationship between various measures of development and democracy in
a cross-section of countries. In succeeding decades, analytic techniques have become
much more sophisticated, more data have become available, and scholars have
developed more nuanced measures of democracy. In ever more sophisticated ways,
analysts have confirmed the existence of a correlation between democracy and
development (Bollen and Jackman 1985; Burkhart and Lewis-Beck 1994; Gasiorowrski
1995; Barro 1996; Przeworski et al. 2000).2
Without denigrating their contribution,
which has been very great, it is still possible to note that little beyond greater
certainty about that original claim has been added to the pile of knowledge we can be
reasonably sure we know.
In trying to understand democratization, we have traditionally relied on descriptions
of transitions in individual countries and small groups of countries or large-N
statistical studies. The case studies have been very useful in providing information
about particular transitions. Large-N studies typically include all countries for which
information about proposed causes is available. These studies have built the current
accumulation of knowledge about the relationship between development and democracy.
The authors of the large-N studies have suggested various processes through
which growth and the related spread of education, urbanization, and individual
mobility might lead to demands for democracy, and many of these arguments have
2
But see Acemoglu et al. (2005) for an empirical challenge. Acemoglu and Robinson's (2001)
deductive model, however, could easily lead to a correlation between democracy and development.
W H A T C A U S E S D E M O C R A T I Z A T I O N ? 319
been tested. A correlation between education, especially primary education, and
democracy is well established (Barro 1996). Some studies have found a relationship
between the income share of the middle class and democracy (Barro 1996). The
results on urbanization are mixed, with some showing a negative effect on democracy.
These studies have not actually modeled the process of democratization via
these avenues, however. They all seem to assume that if citizens want democracy and
have the required skills, they can achieve it.
Given the quality and amount of effort expended on understanding democratization,
it is frustrating to understand so little. Scholars have responded by pushing
the research frontier in two intriguing directions. Some have taken up Robert
Barro's (1996) challenge: "Given the strength of the Lipset/Aristotle hypothesis as
an empirical regularity, it is surprising that convincing theoretical models of the
relation do not exist. Thus development of such a theory is a priority for future
research (S182)." Modeling and testing interactions between elites, who may not
want to share power, and citizens, who may want to influence distribution and
therefore demand democracy as a means of gaining influence, have now moved to
the top of the research agenda. Several scholars have proposed plausible deductive
arguments that identify underlying causes of democratization, most of which are
correlated with development, and that therefore explain the correlation. The next
section discusses recent models of the process of democratization and the evidence
supporting them.
A different direction has been taken by other analysts, who claim that international
factors have played a much larger role in explaining democratization than
earlier observers had realized. If international forces have a major effect on democratization,
and especially if there is an interaction between international and
domestic factors, their exclusion from statistical tests may explain some of the
limited and contradictory results obtained in these tests. International influences
have barely figured in the historical literature on democratization, but studies
including them have produced interesting results in the last few years. The second
section below summarizes recent findings about international effects on transitions
to democracy.
In response to the mix of success and failure to which the study of democratization
has led, I suggest that the reason results have been somewhat limited so far is that the
phenomenon we label democratization actually includes several different causal
processes. If the large-N studies have lumped multiple causal processes into the
same statistical models, it is not surprising that only the most basic relationships
have emerged. Similarly, if the models that have been proposed fit democratization in
some contexts but not others, then it is also not surprising that empirical support for
the models has been modest. A different approach to understanding democratization
would begin by disaggregating into several distinct processes or subgroups and then
theorizing different transition processes separately. In the third section I discuss some
different ways to think about theoretically useful disaggregations of the process of
democratization.
320 B A R B A R A G E D D E S
i I N V E S T I G A T I N G T H E P R O C E S S : W H A T
C A U S E S T H E C O R R E L A T I O N B E T W E E N
D E V E L O P M E N T AND D E M O C R A C Y ?
In a very influential book and article, Przeworski and co-authors (1997, 2000) have
argued that there is no relationship between levels of economic development and
transitions to democracy They note that transitions can occur for many reasons,
not all of which are systematic. They claim that the apparent relationship results
from the political stability of rich democracies. Although poor democracies sometimes
collapse and return to dictatorship, rich democracies never do, which over
time leads to a high proportion of rich countries among democracies. Using a
different measure of democracy and a dataset covering a much longer period of
time, Gleditsch and Choun (2004) also find no relationship between development
and transitions to democracy after controlling for characteristics of countries'
neighbors.3
Other analysts, however, have been unpersuaded by Przeworski et al.'s argument.
In a very careful reanalysis that extends the time period back to 1850, Boix and Stokes
(2003) show that development does contribute to democratic transitions, though the
average effect for the whole period is small relative to the effect of development on
maintaining democracy. In fact, they note that a careful reading of Democracy and
Development shows that even Przeworski et al. (2000) find a small statistically significant
effect of development on the likelihood of transitions to democracy Boix and
Stokes (2003) show that when the dataset is divided by time periods, economic
development is an extremely important predictor of transition prior to 1950, but has
only a small (though statistically significant) effect in the post-1950 period. Epstein
et al. (forthcoming) also challenge the Przeworski et al. (2000) findings. They show
that results are changed by using a trichotomous measure of democracy rather than a
dichotomous one, as Przeworski et al. did. They find that development has strong
predictive power for transitions into and out of the category they call partial
democracy, but less effect on transitions from full autocracy to full democracy.
Epstein et al.'s (forthcoming) findings should probably be interpreted as meaning
that development is a good predictor of the softening or routinization of authoritarian
regimes, though not necessarily of regime change.
3
Pevehouse (2002) also finds no relationship between development and democratization after
controlling for the average level of democracy in the members of regional international organizations
that countries belong to. These findings are open to different interpretations. Since development tends to
vary by region, level of development is likely to be collinear with average democraticness in neighboring
countries or regional international organizations. It might be that development in a region causes
democraticness in a region, thus accounting for the correlation between neighbors' regime type and
the likelihood of democratization, even if neighbors have no direct influence. Alternatively, it might be
that neighbors' influence is the reason for the correlation between development and democracy.
W H A T C A U S E S D E M O C R A T I Z A T I O N ? 321
Economic development is correlated with many other trends, and one or more of
those may be the causal mechanism that accounts for the apparent relationship
between development and democracy. Lipset and other modernization theorists
suggested that increasing education, equality, urbanization, experience of working
in factories, and the weakening of traditional loyalties to tribe and village—all
correlates of economic development—would result in citizens with more tolerant
and participatory attitudes who would demand a say in government (Lipset 1959;
Inkeles and Smith 1974). These arguments stressed the experiences and values of
ordinary citizens as the bases for democracy without specifying the process through
which transitions might occur or giving much attention to the possible reluctance
of elites to give up power. Scholars influenced by Marx expect the middle class—
which tends to grow as the economy develops—to be the carrier of the demand for
democracy: "no bourgeoisie, no democracy."4
Zak and Feng (2003) have modeled a
process through which this relationship might unfold, but have not tested it.
Boix (2003), Acemoglu and Robinson (2001,2005), and Zak and Feng (2003) argue
that democratization is more likely when the income distribution—which tends to
even out as countries reach high levels of development—is more equal. Boix and
Acemoglu/Robinson argue that elites fear redistribution less when income distribution
is relatively equal because the median voter's preference with regard to taxes will
then be less confiscatory. Elites, according to Acemoglu and Robinson (2001), are
willing to cede some power rather than risk the costs of revolution when they expect
democracy not to lead to extremely redistributive taxation. Boix (2003) expects a
linear relationship between equality and the likelihood of democratization. Acemoglu
and Robinson's (2001) model suggests a non-monotonic relationship: at low
levels of inequality, an increase can promote democracy by increasing the threat of
revolution, but at higher levels of inequality, elites will repress rather than offering
concessions because of their fear of the redistributive consequences of democratization.
An empirical challenge to these arguments is that evidence of more equal
income distributions in democracies is at best mixed (Bollen and Jackman 1985).
There is little evidence that the current set of recalcitrant dictatorships is made up of
countries with especially unequal income distributions. In the post-Second World
War period, longer-lived dictatorships (excluding monarchies) have more equal
income distributions than brief ones.
Boix (2003) and Rogowski (1998) argue that capital mobility, which also tends to
rise with development, also contributes to democratization. When capital is mobile,
it can flee in response to high taxes. Knowing that, democratic governments are
expected to refrain from taxing heavily; so elites need not fear democracy. In the Boix
(2003) model, elites' interests can be protected either by a relatively equal income
distribution or by capital mobility. Where capital mobility is low, as in countries with
predominantly agricultural economies, and income unequal, however, elites should
be unwilling to negotiate democratization. The Boix and Acemoglu/Robinson arguments
are discussed in more detail below.
4
This is Barrington Moore's summary of Marx (1966, 416).
322 B A R B A R A G E D D E S
2 M O D E L S OF D E M O C R A T I Z A T I O N AS
S T R A T E G I C I N T E R A C T I O N S B E T W E E N
E L I T E S AND C I T I Z E N S
Models of the interactions between ruling elites and others that may lead to democratization
can be divided into two categories depending on their basic assumptions
about who the relevant actors are and what their goals are. The Boix (2003) and
Acemoglu/Robinson (2001) models described above assume that the most important
division within society is between rich and poor, and that the rich form and maintain
dictatorships in order to protect their assets. They also assume, as do many economic
models of authoritarian politics, that the key policy decision that determines the level
of redistribution is the level of taxation on domestic capital. It is assumed that the
median voter, who is poor, prefers high taxes in order to redistribute wealth. The
more unequal the income distribution, the poorer the median voter and thus
the more confiscatory the tax rate can be expected to be in a democracy.
In short, the median voter in these models has never met "Homer."5
Elites consider
changing the rules, however, because of the threat of violence or revolution. In these
models politicians are perfect agents of societal interests, and political leaders do not
maximize their own revenue distinct from the revenue of the elite group they
represent.
An alternative conception of autocracy assumes that the most important division
in society is between the rulers (sometimes simplified to a single dictator) and the
ruled. They assume that rulers maximize their own income from tax revenue at the
expense of both rich and poor ruled. Rulers thus set taxes at the highest rate that does
not deter economic effort by citizens. In these models, rulers offer increments of
democracy when doing so can increase the credibility of their promises to provide
public goods and other policies that will increase economic growth and thus benefit
both rulers and ruled (North and Weingast 1989; Weingast 1997; Escriba Folch 2003).
Alternatively, democratic institutions may be offered as a means of directly increasing
revenues (Levi 1988; Bates and Lien 1985; Rogowski 1998). Bueno de Mesquita et al.
(2003) propose a more complicated set of societal divisions and actors: a leader; a
ruling coalition; a "selectorate" that includes those citizens who can affect the
composition of the ruling coalition; and residents, those who are taxed but politically
marginal. In all these models, the ruled care about growth and the share of their own
production they are allowed to keep. Taxation is not seen as a means of redistribution
in favor of the poor, but rather as a means of enriching rulers. Rulers become rich by
ruling; they do not rule because they were rich before achieving power. They cling to
power in order to continue collecting revenue from the productive population under
their control, not to protect themselves from redistributive taxation. The main
5
Larry Bartels (2005) has christened the real-life low-income voter who favors more social spending
but who nevertheless opposes the estate tax Homer after the famous Homer Simpson.
W H A T C A U S E S D E M O C R A T I Z A T I O N ? 323
constraint on rulers' pursuit of wealth for themselves is the threat of declining
revenue caused by capital flight or reduction in economic effort.
Both of these approaches offer some insights into the process of democratization.
The Boix (2003) and Acemoglu/Robinson (2001) models are plausible simplifications
of early democratizations in Western Europe and of many transitions in Latin
America, but models emphasizing the conflict between rulers and ruled are more
plausible when applied to recent struggles over democratization in Africa, Eastern
Europe, the Middle East, and parts of Asia. These models, like large-N studies of
democratization to date, have implicitly assumed that a single model will explain
democratization in all times and all circumstances.
2.1 Rich Rulers versus Poor Ruled
As noted above, Boix (2003) argues that income equality and capital mobility reduce
elite fears of democracy, the first because it reduces expected redistribution by
popular governments and the second because it provides capital holders with an
exit option if taxes become confiscatory. This is a seminal contribution to the
literature on democratization because it provides plausible microfoundations for
the observed correlation between development and democracy. Other laudable
aspects of the research include a serious effort to test the argument and the inclusion
of nineteenth- and early twentieth-century democratizations in the analysis. Virtually
all other quantitative studies of democratization have looked only at the postSecond
World War period because of data limitations. Boix has made a huge effort to
overcome those limitations.
The Boix (2003) study has not resolved all debates, however, in part because the
empirical support for the argument is somewhat ambiguous. On the positive side,
income inequality has a substantial effect on the likelihood of democratization in a
dataset that covers 1950-90 and thus excludes most African democratizations. We do
not know if the result would change if a number of transitions in poor African
countries were added. The percentage of family farms, used as a proxy for inequality
in the historical tests, has a negative effect on the probability of transition, contrary
to expectations. One of the measures of capital mobility, average share of agriculture
as a percentage of GDP, fails to produce expected results. Other indicators used to
measure capital mobility have strong effects but ambiguous interpretations.
The ratio of fuel exports to total exports, for example, is a plausible indicator of
capital mobility. The correlation between reliance on oil and authoritarianism,
however, is usually attributed to the oil-producing government's ability to provide
transfers to large parts of the population without relying on taxation.6
So how can we
tell whether the reported relationship between oil dependence and democratization
is caused by reduced capital mobility or the strategic use of resources by dictators to
buy popular support?
6
For rentier state arguments, see Anderson 1987 and Crystal (1995).
324 B A R B A R A G E D D E S
Average years of schooling is used as a measure of human capital, which is more
mobile than physical capital, and Boix finds a positive relationship between education
and democratization. Many other analysts have found this relationship, however,
and attributed it to the propensity of more-educated citizens to demand
democracy. In short, although Boix's argument is plausible and attractively simple,
the empirical investigation is not definitive. The argument fits well with the stylized
facts of West European democratization, however, and redistributive changes followed
democratization in Western Europe as this argument would predict (Lindert
1994). Further tests of this argument deserve to be important items on the research
agenda of students of democratization.
Acemoglu and Robinson's (2001) argument begins with many of the same basic
assumptions about the way the world works as Boix's. It also gives a central role in
resistance to democratization to elites' fear of redistribution when the starting
income distribution is unequal. Its predictions are complicated, however, by limiting
the threat of revolution to periods of recession. In this argument, when the rich are
threatened by revolution (which only occurs during recession), they can grant
redistribution without changing the political system, grant democracy as a way of
making the commitment to redistribution credible, or repress. Redistribution without
regime change is not credible to the poor because they know that they cannot
maintain the threat of revolution after the recession is over. According to Acemoglu
and Robinson, democratization is a more credible commitment to maintaining
redistribution over a longer time period. (Why the poor should accept democratization
as credible when even the model allows the rich to stage coups if they are
dissatisfied by the later tax rate is not clear.)
The introduction of recessions, which vary in both intensity and frequency,
substantially complicates making predictions about the effects of inequality on elite
behavior. Equality makes democratization less threatening to elites, but how they
react to inequality depends on the seriousness of the threat of revolution and the cost
of repression. In this model, the likelihood of revolution depends on inequality
(which increases the threat of revolution) and the intensity of recession (which
decreases revolution's cost to the poor). Frequent recessions, however, increase the
likelihood that the elite can credibly offer redistribution without democratization
because frequent recessions allow the poor to threaten revolution often, thus enforcing
the bargain. So intense recessions destabilize dictatorships leading to democratization,
revolution, or repression, but frequent recessions lead paradoxically to stable
authoritarianism with redistribution. The bottom line, according to Acemoglu and
Robinson (2001, 957), is that "democracy is more likely to be consolidated if the level
of inequality is limited, whereas high inequality is likely to lead to political instability,
either in the form of frequent regime changes or repression of social unrest."
In contrast to the Boix argument, Acemoglu and Robinson expect income inequality
to lead to unstable regime changes, not continued authoritarianism. One of the
attractive features of the Acemoglu and Robinson model is that it explains repeated
transitions between democracy and dictatorship, a phenomenon that has characterized
some parts of the developing world since the middle of the twentieth century. The
W H A T C A U S E S D E M O C R A T I Z A T I O N ? 325
model seems to be a plausible simplification of events in much of Latin America and
in a few other developing countries. It does not fit most of the Middle East, Eastern
Europe, Africa, or Asia, where fear of redistributive taxation is not a plausible reason
for resistance to democratization since substantial portions of productive assets were
state or foreign owned for much of the late twentieth century. State elites who control
a large portion of productive assets may certainly fear loss of power since it will
dispossess them, but they will not suffer less dispossession because the income
distribution is more equal. Acemoglu and Robinson do not offer systematic empirical
tests of their arguments so we cannot assess their fit with the real world.
Models linking democratization to inequality seem highly plausible initially, but
the empirical investigation of the relationship between regime type and income
inequality does not offer strong support for their basic assumptions. Nor does
empirical investigation of the relationship between democracy and redistribution.
If these arguments were correct, we would expect to find the remaining dictatorships
in the world more unequal on average than democracies, but Bollen and Jackman
(1985) find no relationship between democracy and inequality. Przeworski et al.
(2000) find a positive relationship between only one of three measures of inequality
tried and transitions to democracy. They find a stronger relationship between
inequality (in democracies) and democratic breakdown, which might explain any
relationship that exists between democracy and equality (if one does exist), but does
not support the idea that equality makes democratization more likely.
The models also assume that the main reason elites fear democracy and ordinary
citizens want it is that they expect it to lead to redistribution. LIndert (1994) has
shown that the expected redistribution occurred in Western Europe after the first
steps toward democratization were taken, but Mulligan, Sala-i-Martin, and Gil
(2003) show that contemporary democracies do not on average distribute more
than dictatorships.7
We should not be surprised by this result. Income distribution
varied greatly among late twentieth-century dictatorships. Many, both communist
and non-communist, expropriated traditional elites and redistributed income and
opportunities through land reform, much increased public education, and industrialization
policies that led to the movement of large numbers of people out of
agriculture and into factories. It is hard to imagine that elites in these kinds of
authoritarian regimes would be motivated by a fear of greater redistribution. They
would fear loss of their own power and wealth, but not via redistributive taxation.
Income equality would not reassure them.
2.2 Revenue Maximizing Rulers versus Politically
Powerless Citizens
This approach to the study of democratization, which owes much to seminal articles
by North and Weingast (1989) and Olson (1993), sees rulers as maximizing their own
7
Boix (2003) challenges this result.
326 B A R B A R A G E D D E S
individual revenue via taxation and citizens as sharing a desire for productivityenhancing
policies and public goods, regardless of whether they are rich or poor. In
this image of politics, taxes redistribute wealth from citizens to rulers, not from rich
to poor. Rulers may want revenue in order to pursue wars, to buy support in order to
stay in power, or for personal consumption; their reason does not affect the logic of
the argument. Rulers are motivated by their desire for revenue to offer public goods
and a tax rate that does not reduce investment or effort.
In some versions of this approach, societal elites or holders of capital are most
affected by the ruler's policies and can do most to destabilize his rule if they are
dissatisfied. Consequently, they are the ones most likely to be accommodated when
the ruler offers an institutionalized form of participation in return for their cooperation.
Rulers may offer representative institutions as a means of offering a credible
commitment to supply desired public goods (Levi 1988; North and Weingast 1989;
Escriba Folch 2003) or simply in exchange for wealth holders' contingent consent to
the taxation of mobile capital (Bates and Lien 1985). As in the Boix (2003) argument,
democratization becomes more likely as capital becomes more mobile, but the
reason for the relationship changes. The more mobile capital, according to Bates
and Lien (1985), the harder it is to tax without contingent consent and thus the more
likely the ruler will offer representative institutions. Rogowski (1998) suggests a more
general form of this logic in which citizens' ability to move away increases the
likelihood that rulers will offer them representative institutions or good government
in order to induce them to remain, along with their productive capacity, within the
ruler's territory8
Thus these models often explain the first small steps toward
democratization from absolutist monarchy.
Bueno de Mesquita et al. (2003) suggest a more complicated general framework for
understanding politics in both democracy and autocracy. Their model, to reiterate,
includes: a ruler supported by a winning coalition; a "selectorate," meaning those
citizens who have some influence on who can join the winning coalition; and
residents who play no role in selecting rulers. In democracies, the selectorate is the
enfranchised population, and the winning coalition is made up of those who voted
for the winning party or coalition, that is, roughly 50 percent of the selectorate. In
single-party authoritarian regimes, the winning coalition is the small group of actual
rulers, and the selectorate is made up of all members of the ruling party. In military
regimes, the winning coalition is the junta and the selectorate is the officer corps.
They do not discuss reasons for different authoritarian institutional choices.
Rulers maximize personal revenue via taxation but are constrained by the need to
provide private and public goods in order to maintain the support of the winning
coalition. If enough members of the ruling coalition defect because they are dissatisfied
with their share, the ruler is overthrown. Citizens outside the winning
8
But see Bravo (2006) for evidence that the exit of those citizens most dissatisfied with a ruler's
policies may increase the probability that he survives in office—thus giving the ruler a reason to provide
policies that induce the exit of those citizens most likely to join the opposition.
W H A T C A U S E S D E M O C R A T I Z A T I O N ? 327
coalition benefit only from the public goods provided when the winning coalition is
too large to be maintained by private goods alone.
Residents and sometimes members of the selectorate may hold demonstrations or
join rebellions to challenge rulers who tax them too heavily or provide insufficient
public goods, but rulers in this model always respond with repression. If revolutionary
challengers win despite repression, the new rulers face the same incentives that
other rulers do to narrow the winning coalition and keep resources for themselves. In
other words, revolutions and popular uprisings in this model do not threaten
redistribution or lead to democracy. Instead they lead to a seizure of power by a
new leader and winning coalition who maximize their own wealth at the expense of
those they exclude. One of the most useful and empirically realistic points made by
Bueno de Mesquita et al. is that participation in a coup, uprising, or revolution does
not guarantee the participant an improved share of power or wealth after the fall of
the old regime because those who lead such movements have incentives after they
win to renege on earlier promises.
Thus democracy cannot arise as a response to popular uprising in this model.
Instead, it arises when the members of the winning coalition can benefit themselves
by expanding its size. Members of winning coalitions are cross-pressured when it
comes to the size of coalition they prefer to be part of. Their individual share of
private goods is larger when the coalition is smaller, but the ruler keeps less for
himself and provides more public and total private goods when the coalition is larger.
In the model, the winning coalition has a tipping point at the size at which it prefers
to increase further. Once that happens, democracy will eventually follow. This model,
like those described above, portrays democratization as elite led. In the Bueno de
Mesquita et al. (2003) model, however, winning coalition elites are motivated simply
by wanting to improve their own welfare relative to that of the ruler. They are not
responding either to a challenge from the excluded or to the threat of capital strike.
Models that emphasize conflict between revenue-maximizing rulers and politically
powerless citizens capture elements of reality in many recent transitions in developing
countries. Once the changes in the international economy provoked by the debt
crisis had rendered state interventionist development strategies unsustainable, many
authoritarian governments were forced to begin liberalizing their economies. In
order to attract private investment to replace state investment that could not be
sustained without foreign inflows, governments had to offer more predictable policies
and certain public goods conducive to private investment (Roberts 2006). Like
democrats, dictators' survival in office is threatened by poor economic performance.
As noted by North and Weingast (1989), Acemoglu and Robinson (2001), Escriba
Folch (2003), and others, policy promises made by dictators inherently lack credibility.
Dictators can increase the credibility of these promises by creating institutions
that give capital holders a say in policy making and that increase the constraints on
the dictator's arbitrary power. Democratic institutions such as legislatures and
multiparty electoral competition can create those constraints if the commitment to
the institutional change is itself considered credible. If the institutions benefit both
the ruler, by increasing revenues, and the ruled, by increasing productivity or welfare,
328 B A R B A R A G E D D E S
then the institutional bargain is self-enforcing and thus credible. These models, in
other words, provide a reason for expecting institutional bargains to be more credible
than offers to provide desired policies in the absence of institutional change, which
the Acemoglu and Robinson (2001) model does not.
These models thus suggest intuitions about why democratization and economic
liberalization tended to vary together in the late twentieth century (Hellman 1998).
Prior to the debt crisis of the 1980s, governments had a choice between relying
primarily on state investment or private investment. Those that chose state investment
did not have to offer credible commitments to provide public goods, predictable
economic policy, or policies favorable to private investors in order to secure
revenue flows, and thus the economic pressure to initiate institutional constraints on
rulers' arbitrary powers was low. Since the 1980s, the state investment strategy has
become unworkable except possibly in countries reliant on the export of oil or other
high-priced natural resources. Consequently, governments have sought to attract
private investment via capital-friendly policies, and political institutions that constrain
the dictator's discretion help to make those policies credible to investors
(Roberts 2006).
The emphasis on the interest differences between rulers and ruled and on redistribution
in favor of rulers as a central fact of dictatorship fits well with what we know
about many of the dictatorships referred to as personalistic, sultanistic, or patrimonial
by different authors. These models do not accommodate the role that popular
uprisings have played in many late twentieth-century democratizations, however.
Moreover, most of these models are very abstract, and most tests of them have been
narrowly focused or open to multiple interpretations.
Some features of late twentieth-century democratization have not found their way
into models, though they have been included in large-N statistical studies. The
correlation between reliance on oil exports and authoritarianism, for example, has
been found repeatedly. In developing countries, oil is usually state owned or owned
by foreign multinationals and taxed heavily. Whether it is state owned or not, the
government draws its revenues largely from natural resource production, not from
taxation on domestic wealth holders. A large mostly descriptive literature on the
effects of oil on politics exists (Karl 1997; Chaudhry 1997; Anderson 1987; Crystal
1995)- Yet, I know of no model that has grappled seriously with state ownership of
productive resources and its effect on the struggle over democratization. All models
assume a capitalist economy with private domestic investors as important actors.
During the third wave of democratization, however, most transitions affected authoritarian
regimes in which state investment was high. In many, foreign investment
also played a large role, and revenue from foreign aid was more important than
revenues from taxation in some.
International factors have also been largely absent from models of democratization.
Many observers have suggested that international forces, such as the diffusion of
democratic ideas and pressure from international financial institutions to democratize,
have affected transitions, especially since the 1980s. Earlier quantitative studies
found it hard to document these influences, but Gasiorowski (1995) and Gleditsch
W H A T C A U S E S D E M O C R A T I Z A T I O N ? 329
and Choun (2004) show that the proportion of democratic neighbors increases the
likelihood of transitions to democracy in neighboring countries, lending some
support to the diffusion argument. Jon Pevehouse (2002) shows that membership
in regional international organizations in which most other members are democratic
increases the likelihood of democratization. Since membership in democratic regional
international organizations is likely to be correlated with having democratic
neighbors, however, we cannot be sure whether organizations have an independent
effect beyond the effect of living in a "good" neighborhood. Bueno de Mesquita,
Siverson, and Woller (1992; Bueno de Mesquita and Siverson 1995) show that war
affects the survival of both political leaders and regimes. Gleditsch and Choun (2004)
show that wars increase the likelihood of transition from one authoritarian government
to another, but neither Gleditsch and Choun (2004) nor Pevehouse (2002)
shows strong evidence that wars in the neighborhood decrease the likelihood of
democratization, as some have suggested. Marinov (2005) shows that although
sanctions are effective at bringing down democratic leaders, they have little effect
on the survival of dictators and therefore we can infer little effect on authoritarian
regimes.9
Theoretical treatments of democratization, however, continue to focus on
domestic causes. It may be that the focus on domestic causes is appropriate when
explaining democratizations before the Second World War, but that international
influences—both economic and political—have become more pronounced over
time.
3 D I S A G G R E G A T I N G D E M O C R A T I Z A T I O N
Assuming that there is one explanation of democratization may be the reason that
scholars continue to disagree about its causes. Different analysts have deeper knowledge
about some sets of cases than others, and naturally their intuitions formalized
in models fit the cases they know best better than those they know less well. The
findings of large-N studies differ from each other depending on specification, time
period included, and cases used, leaving very basic ideas contested. Such varying
results should be expected if single statistical models are being imposed on a set of
disparate processes without efforts to specify how the process might differ over time
or in different kinds of transitions.1 0
I suggest that it would be useful to consider the
possibility that processes of democratization might be different in different contexts,
9
He does not test the effect of sanctions on economic performance, and growth is included as a
control variable in the test of the effect of sanctions, so it is quite possible that sanctions do affect
authoritarian survival through their effect on growth. In democracies, though, sanctions affect leadership
survival even with growth controlled for.
10
Besides the exceptions noted in the text above, a number of large-N studies have modeled factors
that have changed over time. Gleditsch and Choun (2004), for example, show that the predominance of
330 B A R B A R A G E D D E S
that these differences might be systematic, and that developing a theoretical understanding
of these differences would lead to useful empirical results and a better
understanding of how transitions really take place.
Two context differences that might influence the democratization process are the
historical period in which it takes place and the type of regime that democracy
replaces. Early democratizations took place in capitalist economies in which the rich
usually held political power. Later democratizations have also occurred in countries
with high levels of state ownership of productive assets, especially natural resources.
State ownership makes possible both the accumulation of wealth by political leaders
and also the distribution of benefits to supporters, and in some cases citizens,
without the need for high taxation of private wealth holders. Rulers who have
acquired wealth through access to state resources, in contrast to those who hold
political power because they own private wealth, have to fear losing most of their
assets if they are deposed, regardless of the income distribution or other factors that
might affect future taxation.
Most transitions before the Second World War were transitions from some form of
oligarchic government; many were gradual transitions from very limited suffrage to
nearly universal. Post-Second World War democratizations have occurred in several
quite different ways, but nearly all have involved a transition to immediate universal
suffrage democracy. These have included the transition from colonial rule to universal
suffrage democracy at independence; transitions from universal suffrage authoritarianism
to universal suffrage democracy; and redemocratizations in which
most of the parties and political institutions of a prior democracy are reinstated at
the conclusion of an authoritarian interlude. Gradual transitions from limited to
almost universal suffrage have been rare during the last fifty years (cf. Huntington
1991).
If elite opposition to democracy is motivated by fear of redistributive taxation,
gradual increases in suffrage should be easier than rapid ones because the median
voter after a limited enfranchisement would be richer and thus demand less redistribution.
Such institutional choices are often made during bargaining over the
conditions of transition. We might expect authoritarian rulers concerned about
redistributive taxation to negotiate incremental enfranchisement, but dictators
with different fears might not consider universal suffrage threatening.
Various international influences on democratization have arguably had greater
effects since the Second World War and perhaps greater still since the 1980s. The
differences in the sources of dictators' wealth before and after the Second World War
noted above are associated with a change in economic strategy that swept through
the developing world between about 1930 and 1970. Nearly all developing countries
initiated development strategies that increased state investment, ownership, and
regulation of their economies. These strategies reduced governments' dependence
Catholicism in countries has a negative effect on prospects for democratization before Vatican II and a
positive effect afterward. Gasiorowrski (1995) shows that economic crisis has different effects on the
likelihood of democratization during different time periods.
W H A T C A U S E S D E M O C R A T I Z A T I O N ? 331
on private investors and created non-tax sources of revenue, which could then be
distributed along with monopolies and subsidies of various kinds in exchange for
support. The ability to use state resources to expropriate traditional and foreign
wealth holders and create new elites beholden to the government may have reduced
pressures for democratization during the decades when this strategy remained viable.
A second change in the international economy, beginning around 1980 with the
debt crisis, brought that period to an end. When foreign lending was no longer
available to cover the trade and budget deficits characteristic of the state interventionist
development strategy, developing country governments faced intense pressure
to adopt policies conducive to attracting investment. Attracting investments depends
on credible policy commitments and secure property rights. If, as various analysts
have argued, dictators can use legislatures and other quasi-democratic institutions to
make their policy commitments credible, the economic strategy changes brought
about by the debt crisis of the 1980s should have created strong incentives toward
some degree of democratization. In the post-1980 period, we see an increase in both
democratizations and also the adoption of quasi-democratic institutions by authoritarian
regimes (Levitsky and Way 2006).
The end of the Cold War has also changed the process of democratization. Before
1990, authoritarian regimes were supported with extensive aid and other help from
both superpowers. Such aid both increased the regimes' repressive capacity (Boix
2003, 29-30) and also added to dictators' ability to buy support without redistributing
from domestic producers. Since 1990, Levitsky and Way (2006) show that those
authoritarian regimes with the closest linkages to the USA and Western Europe are
the most likely to have democratic-looking institutions such as multiparty elections
in which some real competition is allowed. Such regimes may be easier to dislodge
since opposition is usually less risky and costly in them. The reduction in foreign
support for dictatorships since the end of the Cold War also contributed to the
increase in democratizations in the late twentieth century.
Thus, for both domestic and international reasons, we might think that a model of
the early process of democratization would be different from a model of the later
process. The finding by Boix and Stokes (2003) that economic development and
income distribution have much stronger effects on the likelihood of democratization
before 1950 than after lends support to the idea that modeling separate processes for
the two time periods would be fruitful.
These cross-time differences in the causes of democratization may be caused in
part by differences in the kinds of regimes from which democracies emerge. PreSecond
World War democratizations, which occurred primarily in Europe and Latin
America, generally replaced governments controlled by the rich, whether these were
planter oligarchies or monarchies, through electoral systems with very limited
suffrage. In most of these non-democratic regimes, legislatures existed, elite parties
or proto-parties competed for office, and struggles by legislatures to limit the power
of monarchs or executives had played an important role in determining the shape of
political institutions. Democratization tended to occur through the extension of
suffrage to new groups without other large institutional changes. More citizens
332 B A R B A R A G E D D E S
voted, sometimes new parties formed to attract the votes of the newly enfranchised,
and elections became fairer, but parliamentary systems in Europe and separation of
powers systems in Latin America accommodated the inclusion of new voters and
parties.
We cannot make the same kinds of generalizations about late twentieth-century
transitions. The authoritarian regimes from which late twentieth-century democratizations
emerged differed from the stylized portrait in the paragraph above. Few of
their rulers were born to wealthy families. Most came to prominence via either a
military career or a rise to leadership in a revolutionary or nationalist party. Some
contemporary authoritarian regimes have repressed all political activity, but many
have held regular elections with universal suffrage. Competition for control of
government has been limited by restrictions on opposition parties or manipulation
of voters and playing field, not restrictions on suffrage. Some contemporary authoritarian
regimes have protected the interests of the rich, but others have redistributed
land, nationalized natural resources, and expropriated other wealth. Sometimes these
expropriations have led to more equal income distributions and other times to an
altered but equally uneven distribution with wealth concentrated in the hands of the
dictator's family and supporters. In the former situation, regime supporters fear the
loss of power entailed by more competitive politics, not redistribution. In the latter,
they fear confiscation, being brought to trial for corruption and human rights abuses,
prison, and execution (Kaminski, Nalepa, and O'Neill 2006), but these dangers are
not lessened by a relatively equal income distribution.
Because of these differences, late twentieth- and twenty-first-century democratizations
may not only be different from earlier ones but also different from each other. If
wealthy private sector elites rule countries, then they may indeed resist democratization
when they expect more redistributive taxation, and their fears may be allayed by
a relatively equal income distribution or capital mobility (Boix 2003). Incremental
suffrage extensions may be especially easy for them to endure. If, however, ruling
elites came to power either through election or revolution as the leaders of movements
determined to overthrow traditional elites, then regardless of whether they
actually carried out their promises or have simply stolen in their turn, their fears of
being deposed seem unlikely to be allayed by factors that reduce future taxation.
Instead, their fears might be allayed by enforceable bargains not to prosecute them
for corruption and human rights abuses (i.e. allowing them to go into friendly exile)
or institutional bargains that give them a good chance of returning to office in
competitive elections in the future.
These differences do not imply a return to case studies or within-region comparisons
as the main way of studying democratization, however.1 1
Rather, they suggest
the possibility that there are theoretically relevant differences among authoritarian
11
Mainwaring and Perez-Linan (2003) argue that democratization in Latin America differed from the
general path shown by Przeworski et al. (2000). Different models could be appropriate for different
regions, as they argue, but we would only know it by comparing regions with each other. Stokes (2004)
provides a thoughtful discussion of why regional differences in democratization processes might occur.
W H A T C A U S E S D E M O C R A T I Z A T I O N ? 333
governments themselves and in the ways that they interact with the ruled that may
require different explanations of how transitions from them occur. As Diamond
(2002, 33) notes, contemporary authoritarian regimes differ from each other "and if
we are to understand the contemporary dynamics, causes, limits, and possibilities of
regime change (including future democratization), we must understand the different,
and in some respects new, types of authoritarian rule." His attempt at classification,
however, simply relies on drawing lines between scores on the Freedom House
scale to create categories. If the relevant differences are in degrees of "not democraticness,"
then we do not need to theorize processes separately; we can simply include
a measure of democracy in statistical models, as a number of analysts have. Unsurprisingly,
they find that countries that are more democratic at time one are likely to
be even more democratic at time two. We cannot tell, however, whether the analysis
means that less repressive forms of authoritarianism are less stable or that democratization
is often incremental, and dictatorships that have liberalized somewhat in one
year often continue on that path in subsequent years. A more fruitful approach to
classification would begin by thinking about how the causes of democratization seem
to vary from one context to another. Then classification could be based on expectations
about how those differences would be likely to unfold.1 2
Linz and Stepan (1996) take a first step toward the kind of theoretically based
classification that might help explain differences in democratization processes with
their classification of some authoritarian regimes as "sultanistic," meaning one
individual has discretion over all important personnel and policy decisions, some
as "neo-totalitarian," meaning, in effect, post-Stalinist, and so on. They expect the
usual characteristics of these different kinds of authoritarian regime to have systematic
effects on different aspects of democratic consolidation. These arguments have
not been tested, but they do suggest plausible links between characteristics of
particular kinds of authoritarianism and expected outcomes.
If post-Second World War authoritarian regimes with different kinds of leadership
tend to have different institutional structures and different relationships with supporters
and ordinary citizens, then we would expect them to break down differently
because different institutions privilege and disadvantage different groups. A simple
and intuitive way to categorize these different kinds of leadership and institutions is
as professionalized military, hegemonic party, and personalistic. These regime types
emerge from struggles among elite contenders with different backgrounds, support
bases, and resources after seizures of power. They do not derive in an obvious way
from underlying social or economic structures, and all have been compatible with a
wide range of economic ideologies. All types were common in the late twentieth
century, so understanding something about how they break down might help to
explain why post-1950 democratizations have been different from those that came
12
Some early descriptions of democratization classified transition processes themselves using categories
such as from above, from below, by transaction, and so on. What I suggest here might build on
these earlier ideas but differs from them in that the classifications of differences would be rooted in basic
features of authoritarian regimes.
334 B A R B A R A G E D D E S
before. In the real world, there are of course lots of borderline cases, but we can use
the simple types to develop theories and empirical expectations.
On average, governments ruled by the professionalized military are more fragile
than other kinds of authoritarianism (Gasiorowrski 1995; Geddes 2003).1 3
They are
more easily destabilized by poor economic performance because factionalization
over how to respond to crisis causes many officers to want to return to the barracks
in order to reunify the armed forces. Because of officers' dread of factionalism, the
first moves toward liberalization often arise within the military elite, as noted by
O'Donnell, Schmitter, and Whitehead (1986). Since military rulers usually decide to
return to the barracks rather than being forced out, transitions from military rule
tend to be negotiated and orderly. Negotiation is more likely to lead to democracy
than is violent overthrow, and the successors to professionalized military regimes are
nearly always elected in competitive elections. Thus the fall of a military regime
usually results in a democracy, though it may not last.
In contrast to the military, several scholars have noted the robustness of hegemonic
party regimes. Geddes (2003) shows that regimes ruled by dominant parties
last substantially longer than other non-monarchic forms of authoritarianism.1 4
Gandhi and Przeworski (2006) argue that dictators supported by single parties
survive longer in office. When dominant or single-party regimes face severe challenges,
they try to hang on by changing institutions to allow some participation by
moderate opponents—thus isolating and rendering less threatening more extreme
opponents (Lust-Okar 2005; Magaloni 2006). When they see the writing on the wall,
they put great effort into negotiating electoral institutions that will benefit them
when they become ex-authoritarians competing in fair elections (Geddes 1995;
Magaloni 2006). If members of a dominant party regime cannot maintain their
monopoly on power, they prefer to be replaced by a democracy since they have a
good chance of being able to continue their political careers as democratic politicians.
Replacement by an opposing authoritarian regime is likely to exclude them
from the political game at best. Consequently dominant party governments negotiate
their extrications through elections. The elections that end the rule of hegemonic
parties most often initiate a democracy, but sometimes they result in a new hegemonic
party regime. This happens because the new ruling party can sometimes make
use of institutions originally devised to help the previous ruling party.
Regimes in which power has been personalized under one individual, however, are
more likely to be replaced by a new dictatorship than by a democracy (Hadenius and
Teorell 2005). Personalistic dictators are less willing to negotiate leaving office
because they face a greater likelihood of assassination, prosecution, confiscation, or
exile than do the leaders of other kinds of authoritarianism. Transitions from
personalist dictatorship are seldom initiated by regime insiders; instead, popular
13
For the logic underlying this argument and some of the evidence supporting it, see Geddes (2003).
14
Hadenius and Teorell (2005) find different survival rates than do most other scholars because their
coding rules do not allow them to distinguish between what most other analysts would identify as a
regime change and smaller institutional changes that occur while a regime, in the usual sense of the word,
remains in power.
W H A T C A U S E S D E M O C R A T I Z A T I O N ? 335
opposition, strikes, and demonstrations often force dictators to consider allowing
multiparty elections (Bratton and van de Walle 1997). Personalistic dictators are
more likely to be overthrown in revolutions, civil wars, popular uprisings, or
invasions (Skocpol and Goodwin 1994; Geddes 2003). Linz and Chehabi (1998)
have described the difficulties of democratization following what they call sultanistic
regimes. Several observers have suggested that transitions from personalist rule are
more affected by international factors, such as pressures from international financial
institutions and invasion by neighboring or ex-colonial countries, than are other
kinds of authoritarianism. International financial institutions pressured a number of
African dictators to agree to multiparty elections (Bratton and van de Walle 1997).
For these reasons, the process of transition from personalized dictatorship should
not be modeled as an elite-led bargain. Transitions from personalized dictatorship
are less likely to result in democracy, but sometimes they do. A model that focused on
such transitions would help us to understand the special circumstances that lead to
this outcome. Neeman and Wantchekon (2002) have proposed that democracy
occurs when neither of two contending forces can defeat the other. They address
situations in which opposition to dictatorship has developed into civil war, but the
model might be generalizable to non-violent forms of political conflict. Models that
explain transitions to democracy from personalized dictatorship should be on the
democratization research agenda, as should models that include foreign pressures.
There may be other fruitful ways of disaggregating the democratization process.
My point in this section has not been to argue that there is one true way to break the
process into theorizable parts, but rather that we have considerable evidence that not
all democratizations occur in the same way and that these differences are systematic
not random. The identification of democratization as one "thing" is an artifact of
our use of normal language to describe the process. If the current state of empirical
knowledge allows us to see that there are theoretically important differences in
democratization processes depending on when they happened, what kinds of dictatorship
were being replaced, or something else, we should not expect a single model
to capture all the processes well. Nor should we combine all democratizations in the
same statistical tests without making an effort to specify cross-time or other theoretically
relevant differences.
4 C O N C L U S I O N
Recent empirical research on democratization has confirmed the relationship between
economic development and democracy. Most research also agrees that countries with
oil and mostly Muslim populations are less likely to be democratic, though these
conclusions have been challenged by some analysts. It has also confirmed that
countries with highly educated populations are more likely to be democratic.
336 B A R B A R A G E D D E S
The explanations for these correlations remain contested. Przeworski et al. (2000) argue
that economic development causes democratic stability not democratization. Boix and
Stokes (2003), however, show that economic development had a substantial impact on
democratization before the Second World War and continues to have a smaller effect.
Middle East scholars have described a process through which oil rents are translated
into popular acquiescence to authoritarianism, but Herb (2005) argues that oil wealth
leads to a misspecification of statistical tests of the effect of economic development
on democratization in oil-rich countries, not to a special kind of rentier authoritarianism.
Most observers have attributed the apparent affinity between Islam and authoritarianism
to traditional values widely held by individual Muslims, but Fish (2002)
claims that the treatment of women in Muslim societies hinders democratization.
These empirical regularities with contested interpretations bring two tasks to the
forefront of the research agenda in the study of democratization: empirical studies
aimed explicitly at testing different causal mechanisms; and the creation of carefully
specified models to explain democratization. Some progress is being made on both
fronts. Fish (2002) tests his argument about the treatment of women. Herb (2005)
attempts to disaggregate the effects of rentierism from the effect of economic
development as a way of testing the rentier state argument. Boix (2003) tests his
argument that income equality and capital mobility increase the likelihood of
democratization. None of these tests is fully persuasive, but they are very useful
steps in the direction of identifying causal mechanisms. Boix (2003), Acemoglu and
Robinson (2001, 2005), Zak and Feng (2003), North and Weingast (1989), Weingast
(1997), Bates and Lien (1985), Neeman and Wantchekon (2002), and others have
proposed formal models of democratization that offer a number of useful insights.
Most of these models have been proposed as universal explanations of democratization,
but when examined carefully, most turn out to be useful simplifications of
democratization or elements of it in one specific context.
I suggest that we take seriously our own research showing systematic differences in
the process of democratization across time and type of authoritarianism. Other
differences in the process may also be theoretically important. We might make
progress faster, both empirically and theoretically, if we identified clear domains
for our arguments about the causes of democratization rather than assuming that
just because we cover many processes of democratization with one word we should
also uncritically model it as one process regardless of what we know about historical
and other differences.
REFERENCES
A C E M O G L U , D., and R O B I N S O N , J . 2001. A theory of political transitions. American Economic
Review, 91: 938-63.
2005. Economic Origins of Dictatorship and Democracy. New York: Cambridge
University Press.
W H A T C A U S E S D E M O C R A T I Z A T I O N ? 337
J O H N S O N , S., R O B I N S O N , J. and Y A R E D , P. 2005. Income and democracy. NBER Working
Paper 11205.
A N D E R S O N , L. 1987. The state in the Middle East and North Africa. Comparative Politics,
20:118.
B A R R O , R. 1996. Determinants of democracy. Journal of Political Economy, 107: S158-S183.
B A R T E L S , L. 2005. Homer gets a tax cut: inequality and public policy in the American mind.
Perspectives on Politics, 3:15-31.
B A T E S , R., and L I E N , D. 1985. A note on taxation, development, and representative government.
Politics and Society, 14: 53-70.
Boix, C. 2003. Democracy and Redistribution. Cambridge: Cambridge University Press.
and S T O K E S , S. 2003. Endogenous democratization. World Politics, 55: 517-49.
B O L L E N , K., and J A C K M A N , R. 1985. Economic and non-economic determinants of political
democracy in the 1960s. Pp. 27-48 in Research in Political Sociology, ed. R. G. Braungart and
M. M. Braungart. Greenwich, Conn.: JAI Press.
B R A T T O N , M., and V A N D E W A L L E , N. 1997. Democratic Experiments in Africa: Regime
Transitions in Comparative Perspective. Cambridge: Cambridge University Press.
B R A V O , J. 2006. The political economy of recent Mexico-U.S. migration: a view into Mexican
sub-national politics. Ph.D. dissertation. Duke University.
B U E N O D E M E S Q U I T A , B., and S I V E R S O N , R. M. 1995. War and the survival of political leaders:
a comparative study of regime types and political accountability. American Political Science
Review, 89: 841-55.
and W O L L E R , G. 1992. War and the fate of regimes: a comparative analysis.
American Political Science Review, 86 (3): 638-46.
S M I T H , A., S I V E R S O N , R. M. and M O R R O W , J. D. 2003. The Logic of Political Survival.
Cambridge, Mass.: MIT Press.
B U R K H A R T , R., and L E W I S - B E C K , M. 1994. Comparative democracy: the economic development
thesis. American Political Science Review, 88: 903-10.
C H A U D H R Y , K. 1997. The Price of Wealth: Economies and Institutions in the Middle East. Ithaca,
NY: Cornell University Press.
C R Y S T A L , J. 1995. Oil and Politics in the Gulf: Rulers and Merchants in Kuwait and Qatar.
Cambridge: Cambridge University Press.
D I A M O N D , L. 2002. Thinking about hybrid regimes. Journal of Democracy, 13: 21-35.
D U N N I N G , T. 2006. Does oil promote democracy? Regime change in rentier states. Presented
at Annual Conference of the International Society for New Institutional Economics,
Boulder, Colo.
E P S T E I N , D., B A T E S , R., G O L D S T O N E , I., D R I S T E N S E N , I., and O ' H A L L O R A N , S. Forthcoming.
Democratic transitions. American Journal of Political Science.
E S C R I B A F O L C H , A. 2003. Legislatures in authoritarian regimes. Working Paper 196, Instituto
Juan March de Estudios e Investigaciones.
F I S H , M. S. 2002. Islam and authoritarianism. World Politics, 55: 4-37.
G A N D H I , I., and P R Z E W O R S K I , A. 2006. Authoritarian institutions and the survival of autocrats.
Unpublished MS. Emory University.
G A S I O R O W S K I , M. 1995. Economic crisis and political regime change: an event history analysis.
American Political Science Review, 89: 882-97.
G E D D E S , B. 1996. The initiation of new democratic institutions in Eastern Europe and Latin
America. In Institutional Design in New Democracies, ed. A. Lijphart and C. Waisman.
Boulder, Colo.: Westview.
2003. Paradigms and Sand Castles: Theory Building and Research Design in Comparative
Politics. Ann Arbor: University of Michigan Press.
338 B A R B A R A G E D D E S
G L E D I T S C H , K. S., and C H O U N , J. L. 2004. Autocratic transitions and democratization.
Prepared for presentation at the International Studies Association, Montreal.
H A D E N I U S , A., and T E O R E L L , J. 2005. Learning more about democratization: persistence and
fall of authoritarian regimes 1972-2003. Prepared for presentation at the APSA.
H E L L M A N , J. 1998. Winner take all: the politics of partial reform in postcommunist transitions.
World Politics, 50: 203-34.
H E R B , M. 2005. No representation without taxation? Rents, development and democracy.
Comparative Politics, 37: 297-317.
H U N T I N G T O N , S. 1991. The Third Wave: Democratization in the Late Twentieth Century. Norman:
Oklahoma University Press.
I N K E L E S , A., and S M I T H , D. H. 1974. Becoming Modern: Individual Change in Six Developing
Countries. Cambridge, Mass.: Harvard University Press.
K A M I N S K I , M., N A L E P A , M., and O ' N E I L L , B. 2006. Normative and strategic aspects of
transitional justice. Journal of Conflict Resolution, 50: 292-302.
K A R L , T. 1997. The Paradox of Plenty: Oil Booms and Petro-States. Berkeley and Los Angeles:
University of California Press.
L E V I , M. 1988. Of Rule and Revenue. Berkeley and Los Angeles: University of California Press.
L E V I T S K Y , S., and W A Y , L. 2006. Competitive authoritarianism: origins and evolution of
hybrid regimes in the post-Cold War era. Presented at APSA, Philadelphia.
L I N D E R T , P. 1994. The rise of social spending, 1880-1930. Explorations in Economic History, 31:
i-37L
I N Z , J., and C H E H A B I , H. E. eds. 1998. Sultanistic Regimes. Baltimore: Johns Hopkins
University Press.
and S T E P A N , A. 1996. Problems of Democratic Transition and Consolidation: Southern
Europe, South America, and Post-Communist Europe. Baltimore: Johns Hopkins University
Press.
L I P S E T , S. M. 1959. Some social requisites of democracy: economic development and political
legitimacy. American Political Science Review, 53: 69-105.
L U S T - O K A R , E. 2005 Structuring Conflict in the Arab World: Incumbents, Opponents and
Institutions. Cambridge: Cambridge University Press.
M A G A L O N I , B. 2006. Voting for Autocracy: The Politics of Party Hegemony and its Decline.
Cambridge: Cambridge University Press.
M A I N W A R I N G , S., and P E R E Z - L I N A N , A. 2003. Levels of development and democracy: Latin
American exceptionalism, 1945-1996. Comparative Political Studies, 36:1031-67.
M A R I N O V , N. 2005. Do economic sanctions destabilize country leaders? American Journal of
Political Science, 49: 564-76.
M O O R E , B. 1966. Social Origins of Dictatorship and Democracy: Lord and Peasant in the Making
of the Modern World. Boston: Beacon Press.
M U L L I G A N , C, S A L A - I - M A R T I N , X., and G I L , R. 2003. Do democracies have different public
policies than non-democracies? National Bureau of Economic Research.
N E E M A N , Z., and W A N T C H E K O N , L. 2002. A theory of post Civil-War democratization. Journal
of Theoretical Politics, 14: 439-64.
N O R T H , D., and W E I N G A S T , B. 1989. Constitutions and commitment: evolution of the
institutions governing public choice in 17th century England. Journal of Economic History,
49: 803-32.
O ' D O N N E L L , G., S C H M I T T E R , P., and W H I T E H E A D , L. 1986. Transitions from Authoritarian Rule:
Tentative Conclusions about Uncertain Democracies. Baltimore: Johns Hopkins University
Press.
W H A T C A U S E S D E M O C R A T I Z A T I O N ? 339
O L S O N , M. 1993. Dictatorship, democracy, and development. American Political Science
Review, 83: 567-76.
P A Y N E , A. 1993. Westminster adapted: the political order of the Commonwealth Caribbean. In
Democracy in the Caribbean, ed. J. Dominguez, R. Pastor, and R. DeLisle Worrell. Baltimore:
Johns Hopkins University Press.
P E V E H O U S E , J. 2002. Democracy from the outside-in? International organizations and democratization.
International Organization, 56: 515-49.
P R Z E W O R S K L A., and L I M O N G I , F. 1997. Modernization: theories and facts. World Politics, 49:
155-83.
A L V A R E Z , M. E., C H E I B U B , J. A. and L I M O N G I , F. 2000. Democracy and Development:
Political Institutions and Weil-Being in the World, 1950-1990. Princeton: Princeton University
Press.
R O B E R T S , T. 2006. An international political economy theory of democratic transition.
Unpublished MS. UCLA.
R O G O W S K I , R. 1998. Democracy, capital, skill, and country size: effects of asset mobility and
regime monopoly on the odds of democratic rule. Ch. 4 in The Origins of Liberty: Political
and Economic Liberalization in the Modern World, ed. P. Drake and M. McCubbins.
Princeton: Princeton University Press.
Ross, M. 2001. Does oil hinder democracy? World Politics, 53: 325-61.
S K O C P O L , X , and G O O D W I N , J. 1994. Explaining revolutions in the contemporary Third World.
Pp. 301-44 in Social Revolutions in the Modern World, ed. T. Skocpol. Cambridge: Cambridge
University Press.
S M I T H , B. 2004. Oil wealth and regime survival in the developing world. American Journal of
Political Science, 48: 232-46.
S T O K E S , S. 2004. Region, contingency, and democratization. Presented at Conference on
Contingency in the Study of Politics, Yale University.
W E I N E R , M. 1987. Empirical democratic theory. Pp. 3-34 in Competitive Elections in Developing
Countries, ed. M. Weiner and E. Ozbudun. Durham: University of North Carolina Press.
W E I N G A S T , B. 1997. The political foundations of democracy and the rule of law. American
Political Science Review, 91: 245-63.
W H I T E H E A D , L. 1996. Three international dimensions of democratization. Pp. 3-25 in The
International Dimensions of Democratization: Europe and the Americas, ed. L. Whitehead.
Oxford: Oxford University Press.
Z A K , P., and F E N G , Y. 2003. A dynamic theory of the transition to democracy. Journal of
Economic Behavior and Organization, 52:1-25.