PART FOUR
The analysis of implementation, evaluation, change and impact
Introduction
The 'problem' focus of the policy approach has meant that, until the 1970s, policy analysis was primarily concerned with the 'front end' of the policy process. This is to say, the analysis of the policy process tended to be preoccupied with issues such as how rational, open or fair decision-making was or could be; and, from the point of view of analysis in and for the policy process, how knowledge could improve decision-making (Van Meter and Van Horn, 1975: 450-1). However, in the aftermath of policy-making in the 1960s there emerged a growing interest in what might be termed the 'post-decisional' phases of public policy. It became apparent in the 1970s that many policies and programmes had not performed as well as their advocates had hoped. Problems had not been 'solved'; indeed some had been made worse. The best-laid plans had all too often gone awry, and in place of a 'problem/solution' discourse, analysts and actors began to direct their attention to what had gone wrong, why and how. This theme of good decisions and poor or unintended outcomes had, as we noted earlier, been touched upon in Selznick's study of the great hope of inter-war liberalism, the TVA (see 3.6.2). As it became evident that policy-making in so many areas had not achieved its stated goals, or that those goals had not been well defined, so students of public policy began to shift their attention from inputs and processes towards the withinputs, outputs and outcomes.
At first, this concern with the 'delivery' end of the policy process was framed by attempts to model a rational set of sequences involved in successful implementation, thus extending the logic of the policy stage approach to a more detailed analysis of the closing phases of the decision-making cycle (4.3.2). However, this so-called 'top-down' model soon came under attack from those who argued that implementation problems were far more complex a phenomenon than the hierarchical framework suggested. In the 1970s and 1980s the study of implementation was to expand and several alternatives to the 'top-down' view
457
PUBLIC POLICY
were to be advanced (4.3.3-4.3.8). During the 1980s there were ^ important developments in the 'real world' of public policy. In Brit -° and America, and later in most other industrial countries, new tudes towards the role of government and the 'public sector' ^ ": problem-solver and service-provider came to the forefront of the litical agenda (see Massey, 1993; Hughes, 1994). Market and manage rial approaches offered a radical alternative to the hierarchical frame--work within which analysis of 'implementation' and 'evaluation' had \ for so long taken place. The financial constraints of the 1980s resulted! in attempts by governments throughout the industrial world *o tame! spending and control the growth and influence of bureaucracy. As the^ pressure on the state to 'live within its means' and 'cut its coat accord.1 ingly to its cloth' increased/ so did the demand for techniques which! could facilitate more control over the fiscal crisis of the modern state * The emphasis on the delivery end of public policy signalled a revolution in the discourse of governance; the language of economics and! management came to replace that of 'professionalism', 'administration' and the 'public interest' (Massey, 1993: 12-29).
This movement towards new forms of public-sector 'management' and privatization has not been confined to the Anglo-Saxon world (see Martin, 1993; Letwin, 1988). This was reflected, for example, ml the decision by the OECD in the 1980s to strengthen their co-operation in matters of public management so as to facilitate an exchange offl experience in the public sector. What the OECD has found is that, notwithstanding the obvious differences between member countries, the problems which they confront seem to show a remarkable convergence (see 2.12). As it reported in 1990: |f
Governments of all OECD countries now recognize that improving public management is an integral part of the structural adjustments needed fot/# better economic performance in a changing global environment. There is a*; growing conviction that a radical change in the 'culture' of public admin-r^ istration is needed if the efficiency and effectiveness of the public .sector is;;g to be improved. :% (OECD, 1990) •$
The framework of public policy is consequently being shaped by con-Io tinuing and increasingly common concerns about cost-effectiveness, : delivery of policies and services, improving human resource management, and better monitoring and evaluation (OECD, 1993a). J-urther-more, as we discussed in Part Three, a key trend in policy-making in, industrial societies, which many scholars have identified, is that the growth of government has taken place less in the expansion of central government bureaux - as the budget-maximization model would have it - than in the growth in new interorganizational and intergovurn- ;
DELIVERY ANALYSIS
459
mental arrangements at the implementation level (see Dunleavy, 1991: 223-5). The experience of the 1980s and 1990s has thus been in marked contrast to earlier periods in Western democratic systems: whereas 'overload' (see 2.11) was claimed to be the main feature of policymaking in the 1970s, subsequent decades have witnessed a process of 'downloading' - or what Dunleavy terms 'deinstitutionalizing' - via market-type mechanisms (such as competitive tendering and contract-ing-out) and the expansion of the role of the voluntary 'non-profit' sector. The analysis of the delivery side of public policy therefore increasingly involves taking account of techniques (or fads) of management control developed in the private/profit sector being applied with missionary zeal in the public/non-profit sectors (Anthony and Herzlinger, 1980; Handy, 1988; Drucker, 1990.)
if Trends in public management
Hede (1991) surveys the trends in four Anglo-American systems towards managerial reforms. ■\e concludes that, in the four countries surveyed, 'all are more performance-orientated, and place greater emphasis on mobility and executive development than previously. Though stili disparate in size, there is a trend towards flatter structures' (p. 507). Aucoin (1930) surveys the widespread international swing of the pendulum towards a managerialist 'paradigm'. More-recent surveys, such as by the OECD, confirm that this trend is not confined to Anglo-American political systems.
David Osborne and Ted Gaebier, Reinventing Government, 1992
"his book has proved a very popular and influential approach to government reform (especially on the Clinton administration). In its first year it went through some nine printings and was the subject of much discussion. Osborne and Gaebier attack the bureaucratic approach to deliver-ig public policy, which they regard as being ill-suited to the late twentieth century. As they "idmif, the book owes a good deal to the critiques of bureaucracy put forward by Drucker, Reich, "eteis and Waterman, and Toffler.
Osborne and Gaebier urge that government should be 'reinvented' to take account of the information-rich possibilities of the 1990s and exploit entrepreneurial spirit, greater decentralisation and more responsive forms of public organizations. They draw on examples from the US, Britain and other industrial countries which illustrate the 'reinvention' which is actually taking place.
le reinvention should be built on a number of principles: " governments should steer more than they row;
• policy-making should be about empowering communities, rather than simply delivering services;
• governments should encourage competition in the delivery of services rather than monopoly;
460     PUBLIC POLICY
DELIVERY ANALYSIS 461
• public organizations should be driven by a sense of mission rather than rules; 0 funding should be focused on outcomes rather than inputs;
• the needs of the customer should be the priority, not the needs of bureaucrats;
• public organizations should concentrate on earning, not just spending;
• invest in prevention, rather than cures;
• authority should be decentralized;
- solve problems by leveraging the market place, rather than by simply creating public grammes.
OECD, Administration as Service, 1987
The Organization for Economic Co-operation and Development was established in 1961 to promote economic growth, employment, financial stability and living standards. Its membership is composed of all the major market economies. The publication of the report on public administration marked the degree to which concern about the management of the public sector and the citizen as 'client' had become an international phenomenon. Here is the summary of the main conclusions:
8  The root of the problem: The public and the government want more responsive government.
• Designing policies for impact Policy and programme design can affect the ability of the administration to be responsive. Particular attention must be given to the selection of policy instruments and to the specification of administrative tasks.
• Adapting institutional arrangements: The broad institutional arrangements of the public service at all levels must be considered as the essential context for improving administrative responsiveness.
• Organizing for delivery. The responsiveness of units responsible for implementation depends upon finding an appropriate balance between management components of personnel, organization, control, procedures and communications.
• Increasing administrative responsiveness: Administrative behaviour (organizations and individuals) is largely determined by controls and incentives which are specified through performance standards. Performance measures should be adjusted to incorporate responsiveness.
The summary for policy-makers offers words of wisdom drawn from the policies of member states:
'Engage the participation of users in the design of administrative processes' (Austria). 'Simplify and improve relations between the citizen and the administration' (France). 'The recognition of the primacy of the customer must be paramount' (Ireland). 'Public administration must respond to the public's rights, wishes and needs' (Norway). 'Lifting the burden of government [on private enterprise]' (UK).
'The success of new economic policies depends on the transparency and effectiveness of administrative operations' (Turkey). ♦
In addition, to this replacement of a 'public administration' paradigm by the 'public-sector management paradigm', students of the policy process have also been busy revising their frameworks to take more account of the changing structures and institutional arrangements in modern political systems. As proponents of 'network' and 'sub-system' approaches to policy-making have noted, government is no longer a matter of triangular relationships and tiers of decision-making: new metaphors are required to explain the dynamics of policy formulation and implementation in a more complex 'post-modern' society (see 2.10). It is against this background of changing ideas and institutional forms that the study of the 'output' side of policy-making and of policy analysis must be viewed in the 1990s.
4.2   The arrangement of Part Four
The remainder of the book is divided into five sections plus conclusions (4.8):
• Implementation (4.3): reviews different approaches to the analysis of how policy is put into action or practice.
• Delivery systems (4.4): looks at how we can analyse implementation in terms of the mix of instruments, institutions and values which are used in providing public policy.
• Evaluation (4.5): examines how public policy and the people who deliver it may be appraised, audited, valued and controlled.
• Change and continuity (4.6): considers various approaches to studying the way in which policy change takes place.
• Promise and performance (4.7): focuses on the evaluation of policy impacts and outcomes.
4.3.1   Approaches to implementation
The implementation problem is assumed to be a series of mundane decisions and interactions unworthy of the attention of scholars seeking the heady stuff of politics. Implementation is deceptively simple: it does not appear to involve any great issues. (Van Meter and Van Horn, 1975: 450)
A study of implementation is a study of change: how change occurs, possibly how it may be induced. It is also a study of the micro-structure of political life; how organizations outside and inside the political system conduct their affairs and interact with one another; what motivates them
462
PUBLIC POLICY
to act in the way they do, and what might motivate them to act differs : (Jenkins, 1978: 203) aiITerexUly.
Policy-making does not come to an end once a policy is set out approved. As Anderson nicely expresses it: 'Policy is being made as it is being administered and administered as it is being made' (Anderson 1975:98). Implementation is policy-making carried out by other means - (to paraphrase Clausewitz on war). However, traditionally we have tended to view the political system in a way which reinforces the problem, by demarcating between policy and administration. Administration, according to this (Wilsonian) viewpoint, takes over where policy ends. The job of the administrator is to carry out policy formulated by decision-makers, and the role of the service provider is to carry out the policy administered by the bureaucrat. The interplay and interaction between politicians, administrators and service providers has, until comparatively recently, been a neglected area of analysis and research: a 'missing link' in the policy process (see Hargrove 1975). '
To a great extent, the lack of concern about the problems of 'post-policy-making' was due to the dominance of the models and maps which structured inquiry. The black-box model, for example, provided a powerful framework to analyse policy qua 'system', but tended to assume much about the processes which took place within the system, and within the 'output' and 'feedback' activities. Analysts of policies tended, until the 1970s and 1980s, to bypass the impact of bureaucracy and service-providers on the effectiveness of a policy. A policy was judged in terms of the decision-makers rather than by the 'street-level' implementation of fine-sounding ideas from national and local leaders. At the same time, the tradition of Anglo-American public administration has tended to lay great stress on the different functions of the administrator and the politician. These distinctions, although a necessary part of the liberal-democratic idea of the state and accountability (see Massey, 1993: 200-1), were in practice somewhat unrealistic and bore little relation to the political reality in which bureaucrats were not just neutral servants, but also had ideas, values, beliefs and interests which they used to shape policy. This distinction between policy as politics and administration as implementation, which was fundamental to the Anglo-American notion of public administration, was perhaps less evident in other continental European political systems, where the civil servant has long been characterized as having a more dynamic 'policy' role (see Aberbach et al, 1981).
DELIVERY ANALYSIS
463
^ An outline of the development of implementation studies Key readings
, 1 g40s: the work of Selznick on the TVA (1949). imp[ementation stage 'discovered':
. The analysis of failure: Derthick (1972); Pressman and Wildavsky (1973); Bardach (1977). . Rational (top-down) models to identify factors which make for successful implementation:
Van Meter and Van Horn (1975); Hood (1976); Gunn (1978); Sabatier and Mazmanian
(1979).
, Bottom-up critiques of the top-down mode) in terms of the importance of other actors and organizational interactions: Lipsky (1971); Wetherley and Lipsky (1977); Elmore (1978, 1979); Hjern era/. (1978).
. Hybrid theories. Implementation as: evolution (Majone and Wildavsky, 1978); as learning (Browne and Wildavsky, 1984); as a policy-action continuum (Lewis and Flynn, 1978, 1979; Barrett and Fudge, 1981); inter-organizational analysis (Hjern, 1982; Hjern and Porter, 1981); and policy types (Ripley and Franklin, 1982); as part of a policy sub-system (Sabatier, 1986a); and as 'public sector management' (Hughes, 1994).
In the sections which follow we examine several of these contributions.
Collections of readings or commissioned chapters on implementation which provide useful material are: Williams and Elmore (eds) (1976); Younis (ed.) (1990); Hill (ed.) (1993).
Williams and Elmore is essential to get a feel of the disillusionment with the reforms of the 1960s and 1970s out of which implementation was to grow. Younis is one of the few books in the field which takes a comparative perspective and includes studies of implementation from Europe, America and the developing world. Hill's collection also provides a variety of case studies drawn from European experience.
The best review of implementation theory is Sabatier (1986a). ♦
4.3.2   Top-down rational system approaches
This model was the first on the scene. The neglect of the politics of implementation was brought to an end with the publication of a study by Martha Derthick of urban policy New Towns in Town: Why a Federal Program Failed (1972), and Implementation by Pressman and Wildavsky (1973). Although Derthick's study was an important breakthrough in the development of a new focus on implementation, it was Pressman and Wildavsky's book which has had the most impact. Not least of the reasons for this is because it must hold a record for one of the longest subtitles of any book in public policy, including government reports: Implementation: How Great Expectations in Washington Are Dashed in
464
PUBLIC POLICY
Oakland; or, Why It's Amazing that Federal Programs Work At All, 7^ Being a Saga of the Economic Development Administration as told by 7^ Sympathetic Observers who Seek to Build Morals on a Foundation of Ruined Hopes! In 1968 the authors became interested in the efforts by Oakland Economic Development Administration (EDA) in California to irrtple. ment a programme of economic development for the city. As they explored the programme they became aware of the fact that, although often discussed, the problems of implementation were rarely analysed. The study involved interviewing the actors concerned, and examining policy and other documents over a three-year period to find out what went wrong. A policy, they posited, is 'a hypothesis containing initial conditions and predicted consequences. If X is done at time tlt then Y will result at time t2' (Pressman and Wildavsky, 1973 xiii). Implementation therefore is a 'process of interaction between the settings of goals and actions geared to achieve them' (p. xv). It is essentially an ability to 'forge links' in a causal chain so as to put policy into effect. Implementation will, they argued, become less and less effective as the links between all the various agencies involved in carrying out a policy form an ' implementation deficit'. Goals have to be clearly defined and understood, resources made available, the chain of command be capable of assembling and controlling resources, and the system able to communicate effectively and control those individuals and organizations involved in the performance of tasks.
In so dissecting the mistakes of the EDA, the authors hoped that the experience of Oakland would serve as an example to other policymakers: implementation requires a top-down system of control and communications, and resources to do the job. The moral is, that decision-makers should not promise what they cannot deliver. If the system does not permit of such conditions, then it is best to rein in the promises to a level which is more attainable for the implementation process:
Promises can create hopes, but unfulfilled promises can lead to disillusionment and frustration. By concentrating on the implementation of programs, as well as their initiation, we should be able to increase the probability that policy promises will be realized. Fewer promises maybe made in view of a heightened awareness of the obstacles to fulfillment, but more of them should be kept. (Pressman and Wildavsky, 1984: 6)
In later years the original position taken by Pressman and Wildavsky was modified to take account of the growing literature which their study had helped to stimulate. Wildavsky, in conjunction with Majone and Browne, put forward a more developed theory of implementation which form new chapters to later editions of the Oakland study:
DELIVERY ANALYSIS
465
» Majone and Wildavsky (1978): 'Implementation as evolution'
• Browne and Wildavsky (1984): 'Implementation as mutual adaption'
• Browne and Wildavksy (1987): 'Implementation as exploration'.
The main theme of these pieces is that implementation has to be understood as a more evolutionary, 'learning' process, rather than as the kind of policy-implementation sequence which was originally put forward. In the adjustments to the first edition, the study acknowledges that implementation is a process which involves implementers in making policy as well as in carrying out, or putting into effect, policy from above. These chapters written with Majone and Browne serve as a bridge between the 'top-down' model and later 'bottom-up' critiques.
The original study, however, set out an essentially 'top-down' view of implementation. Effective implementation required, they argued, a good chain of command and a capacity to co-ordinate and control which was sadly lacking in the case of the ODA. This 'top-down' notion of a rational system model - or ideal type - of implementation was later developed in the work of Andrew Dunsire (1978a, 1978b, 1990), Christopher Hood (1976) and Lewis Gunn (1978). These analysts proposed models which asked what would 'perfect' implementation Took like'. The idea, of course, has much in common with Weber's construction of the ideal type of bureaucracy. In his book Limits to Administration Christopher Hood (1976) set out five such conditions for perfect implementation:
• that ideal implementation is a product of a unitary 'army'-like organization, with clear lines of authority;
• that norms would be enforced and objectives given;
• that people would do what they are told and asked;
• that there should be perfect communication in and between units of organization;
• that there would be no pressure of time.
❖ Lewis A. Gunn, 'Why is implementation so difficult?', 1978
In this seminal article Gunn asked a very pertinent question. He draws attention to the general neglect of the issues and sums up what the state of play was with regard to the theory and practice of implementation. Gunn then set out ten conditions (or perhaps commandments) which could be said to provide a framework of questions that might be asked about a programme:
1   Circumstances external to the implementing agency do not impose crippling constraints.
466
PUBLIC POLICY
10
Adequate time and sufficient resources are made available to the programme.
Not only are there no constraints in terms of overall resources, but also at each staa ■
the implementation process the required combination of resources is actually available
The policy to be implemented is based on a valid theory of cause and effect.
The relationship between cause and effect is direct and there are few, if any, intervene
links. ' 9
There is a single implementation agency which need not depend upon other agencies fo
success. If other agencies must be involved, the dependency relationships are minimal T
number and importance.
There is complete understanding of and agreement upon the objectives to be achieved-and these conditions persist throughout the implementation process. ' In moving towards agreed objectives it is possible to specify, in complete detail and perfect sequence, the tasks to be performed by each participant.
There is perfect communication among, and co-ordination of, the various elements or agencies involved in the programme.
Those in authority can demand and obtain perfect obedience. ♦
This 'top-down' approach has a view of the policy-implementation relationship which is summed up in Rousseau's Emile: 'Everything is good when it leaves the Creator's hands; everything degenerates in the hands of man.' The rational model is imbued with the ideas that implementation is about getting people to do what they are told, and keeping control over a sequence of stages in a system; and about the development of a programme of control which minimizes conflict and deviation from the goals set by the initial 'policy hypothesis' (Pressman and Wildavsky, 1973: xiii). Andrew Dunsire (1990:15), for example; puts forward a 'rationalist model' which deliberately excludes all considerations of love, hate, envy, in fact any motivational factors whatsoever'. Where implementation has failed - that is a policy objective has not been met - it may be said to be due to factors such as the selection of the wrong strategy, or wrong 'machinery' or 'instruments'; the 'programming' of the bureaucracy was incorrect; operationalization was poor; something went wrong at the 'shop-floor level'; or there was a poor response to problems. However, if we are to arrive at an explanation of what goes wrong, Dunsire argues, we must start (a la Weber) with the rational ideal - 'when things go right':
We have to understand the nature of a bureaucracy, the essence of which in structural terms, is an advanced specialisation function in both horizontal and vertical planes, so that each member has a well defined function... the essence of which, in process terms, is advanced routinisation of performance ... We have to understand the nature of 'implementation' itself, comprising, basically, a repeated and widespread exercise of a thought process called 'operationalising'; ... and secondly of an organising of 'engineering' or 'designing'... called 'programming'... The number of points
DELIVERY ANALYSIS
467
at which things could 'go wrong' is clearly immense ... the wonder is that things ever 'go right' ... (Dunsire, 1990: 26)
The rational model is essentially a prescriptive theory in the sense which we may find in Taylorism and scientific management, and may be subjected to the same kind of criticisms. Too much emphasis is placed upon the definition of goals by the top, rather than on the role of the workers on the line. It assumes a great deal about goal definition and human interaction and behaviour, or, as in the case of Dunsire, it just blatantly excludes any consideration of how real people actually behave, all the better to understand the logical relationship between input, process and output. However, the distinction between 'policy' as input and implementation as the administrative output is specious. As 'bottom-up' critics argue, the implementation process involves 'policy-making' from those who are involved in putting 'it' into effect. Implementation is not a process in which x follows y in a chain of causation. Unlike a sausage factory, the output of public agencies is not so well-defined and quantified - or 'evaluated' - and what actually counts as success and failure is a matter of controversy and conflict.
4.3.3   Critiques of the rational control model Bureaucratic 'street-level' behaviour
The top-down model has been greatly criticized for not taking into account the role of other actors and levels in the implementation process. A major source of this criticism actually pre-dates the top-down model. In an article published in 1971 Michael Lipsky argued that students of public policy had to take account of the interaction of bureaucrats with their clients at a 'street-level' (Lipsky, 1971). Later, he developed his ideas more fully for a collection of papers on urban policy (Lipsky, 1976). He concluded that:
To better understand the interaction between government and citizens at the 'place' where government meets people, I have attempted to demonstrate common factors in the behaviour of street-level bureaucrats. I have suggested that there are patterns to this interaction, that continuities may be observed which transcend individual bureaucracies, and that certain conditions in the work environment of these bureaucracies appear to be relatively salient in structuring the bureaucrat-citizen interaction ... If this analysis has been at all persuasive, it suggests that in significant respects street-level bureaucracies as currently structured may be inherently incapable of responding favorably to contemporary demands for improved and more sympathetic service to some clients. Street-level bureaucrats respond to work-related pressures in ways that, however understandable or well-intentioned, may have invidious effects on citizen impressions of governmental responsiveness and equity in performance. If, indeed, gov-
468
PUBLIC POLICY
ernment may be most salient to citizens where there is frequent interact* with its 'representatives' and where the interactions may have import.,0** consequences for their lives, then these conclusions should evoke symn thy for current proposals for urban decentralization of authority. Wl^t* ever their other merits or difficulties, these proposals commend thenj selves at least for their concentration on fundamental alterations of th work environment of street-level bureaucrats. (Lipsky, 1976: 208-10)
Following this piece, in 1977 a study by R. Wetherley and M Lipsky . 'Street level bureaucrats and institutional innovation: implementing special education reform', showed how the rational model was not effective in practice, or convincing in theory. The authors examined the implementation of a law in the state of Massachusetts (Comprehensive Special Law, 1972), which depended on a shake-up in the attitudes and practices of teachers, welfare workers and others in the public sector who had direct contact with the public (service-providers). The conditions which the rational model sets out (a good chain of" command, well-defined objectives, and so on) were in large part in place: the law was well defined; there was plenty of support; ample resources; and a communications and monitoring system. Yet the law was not implemented in the way which policy-makers intended: indeed, the changes actually made matters worse as the workload on schools and staff increased dramatically. The services were maintained by 'coping strategies' employed by the dedicated and committed people involved at the 'street level' (see Hudson, 1989, for a review and interesting application of the Lipsky model to public policy and the disabled). The implication of this study was that control over people was not the way forward to effective implementation. Instead of regarding human beings as chains in a line of command, policy-makers should realize that policy is best implemented by what Elmore (1979, 1985) termed a 'backward mapping' of problems and policy, which involves defining success in human or behavioural terms rather than as the completion of a 'hypothesis'. Forward-mapping - or the top-down approach - Elmore regards as little more than a myth which was 'increasingly difficult to maintain in the face of accumulating evidence on the nature of the implementation process' (Elmore, 1985: 20). Elmore suggested that we should begin
with a concrete statement of the behaviour that creates the occasion for a policy intervention, describe a set of organizational operations that can be expected to affect that behaviour, describe the expected effect of those operations, and then describe for each level of the implementation process what effect one would expect that level to have on the target behaviour and what resources are required for that effect to occur. (Elmore, 1985:28)
DELIVERY ANALYSIS
469
What is really important, argue the bottom-uppers, is the relationship of policy-makers to policy deliverers. It is Alice in Wonderland in reverse. Alice says: 'Begin at the beginning and go on to the end.' The idea of backward mapping is to begin at the phase when the policy reaches its end-point, then analyse and organize policy from the patterns of behaviour and conflict which exist. The bottom-up model is one which sees the process as involving negotiation and consensus-building. These involve two contexts or environments: the management skills and cultures of the organizations involved in implementing public policy (schools, hospitals, police forces, welfare agencies, armed forces, government departments), and the political environment in which they have to work.
Bottom-up models lay great stress on the fact that 'street-level' imple-menters have discretion in how they apply policy. Professionals have a key role in ensuring the performance of a policy: teachers, doctors, planners, engineers, social workers, architects, all have opportunities and responsibilities of control and delivery of a service. This means that, as Dunleavy notes, the policy formulation process may be 'skewed' by policy implementation which is dominated by professionals (Dunleavy, 1981, 1982). Teachers, for example, may develop ways of teaching or implementing 'government policy' which actually result in outcomes which are quite different to those intended or desired by policy-makers. The same could be said for other professions charged with carrying out law or policy. Of course, this raises the question of analysis and prescription: street-level implementers may be shaping policy - but is it right for teachers or the police to be making up 'policy'? (See Linder and Peters, 1987.)
That this may come about is because implementation involves a necessarily high margin of discretion. The analysis of discretion in public policy is a subject addressed by students of social policy (Hill, 1969; Titmus, 1968), administrative law (Wade, 1967; Davis, 1969), law enforcement (Cain, 1973; Lambert, 1967; Wilson, 1970); and organizational sociology (Simon, 1945; Dunsire, 1978b; Gouldner, 1954). Policies, regulations, laws and procedures contain an interpretative element. As Davis expresses it: 'A public officer has discretion wherever the effective limits on his power leave him free to make a choice among possible courses of action and inaction' (Davis, 1969:4). Whether the mode of implementation is top-down or bottom-up, those on the front line of policy delivery have varying bands of discretion over how they choose to exercise the rules which they are employed to apply. On a larger scale, the existence of discretion within international law and policy-making makes very clear the problems of differences in interpreting and applying general policy to specific circum-
470
DELIVERY ANALYSIS
471
PUBLIC POLICY
stances. This is especially the case in the European Union, where law (directives and regulations) show considerable variation in how they are implemented by member countries (see From and Stava 1993).
Implementation as a political game
Models of organizations which see policy being made and implemented in situations of human interaction, rather than as a machine or system, focus on the nature of those interactions. In the Lipsky model implementation, for example, is something which involves the recognition that organizations have human and organizational limitations, and that these must be recognized as a resource. Effective implementation is a condition which can be built up from the knowledge and experience of those in the front line of service delivery.
This theme of interaction is also the focus of models which view implementation as a process which is structured by conflict and bargaining. Rational models, of course, also recognize that conflict and deal-making will take place in implementation. However, this conflict is seen as something which is essentially dysfunctional and in need of co-ordination (Pressman and Wildavsky, 1973: 134) or resolution (Dunsire, 1978a). In these models conflict and bargaining take place within shared goals, in which case implementation is effective when groups resolve their differences and put a policy into action. An effective implementation process will have methods and systems of controlling such conflict so as to bring about compliance (Dunsire, 1978a). It is all a matter of control.
However, if we have a view of organizations which is based less upon the notion of control as of structures which are composed of groups and individuals all seeking to maximize their power and influence, we may see such conflict as an essentially political process involving different strategies for acquiring and maintaining power. Implementation from this perspective is about self-interested people 'playing games'.
This game model was advanced by Bardach in 1977 in his book The Implementation Game. Implementation, he argues is a game of 'bargaining, persuasion, and manoeuvring under conditions of uncertainty' (Bardach,1977: 56). Implementation actors are playing to win as much control as possible, and endeavouring to play the system so as to achieve their own goals and objectives.
The Bardach model is essentially one which suggests that politics extends beyond the formal 'political' institutions. Politics does not
stop once a bill becomes law. It does not stop in the political process, nor does it cease in the decision-making process. Models of the kind which Bardach proposes are urging us to redefine the boundaries between politics and bureaucracy, and between the decision-making process and the delivery of those decisions. Implementation is therefore simply another form of politics which takes place within the domain of unelected power.
As we noted above, the rational-control model of implementation tends to see interests as capable of being united for common goals, and conflict as a manageable source of friction. However, models of implementation which stress power, conflict and interests as the stuff of which implementation is made, divide on how it may be interpreted. The various models of power that we discussed in Part Three (3.3) would see the conflict and power struggles within and around an organization responsible for implementing policy in very different ways (for a review of these, see Morgan, 1986: 362-5).
4.3.4  Policy-action frameworks: implementation as an evolutionary process
A problem with both the top-down and bottom-up frameworks is that they tend to over-simplify the sheer complexity of implementation. Two early models which incorporated and developed the insights of both approaches were developed by Lewis and Flynn (1978,1979) and Barrett and Fudge (1981). Lewis and Flynn, in an examination of urban and regional policy, put forward a behavioural model which views implementation as 'action' by individuals which is constrained by the world outside their organizations and the institutional context within which they endeavour to act (figure 4.1).
Figure 4.1   Lewis and Flynn's model
Perception Channels of action
The world outside the organization
	The individual within an organization		The institutional context
			
			
Source:  Adapted from Lewis and Flynn {1978:11)
Policy implements inhabit a world which bears little resemblance to the rational ideal:
472     PUBLIC POLICY
DELIVERY ANALYSIS 473
I
mm
I
ill
H
HRe
■Hi
W/Bp
H
In reality there are disagreements about policy goals and objectives; Vagueness and ambiguity about policies and uncertainty about their operation alisation in practice; procedural complexity; inconsistency between p0w" ers available and existing problems; and conflict arising from public participation, pressure group activity and political dissensus. (Leans and Flynn, 1978: 5)
The interaction with the outside world, the organization and its institutional context means that policy objectives are not the source of guides to action. In the cases they examined it was more often to be found that
actions result from the resolution of conflicts between two sets of priorities and policy areas; may precede the formulations of a procedure for dealing with similar cases in future and therefore the policy; or may result from what is feasible in the circumstances rather than the fulfilment of the original objectives. (Ibid.)
This theme of analysis in context is also present in the ideas of Barrett and Fudge, who argue that implementation may be best understood in terms of a 'policy-action continuum' (figure 4.2) 'in which an interactive and negotiative process is taking place over time, between those seeking to put policy into effect and those upon whom action depends' (Barrett and Fudge, 1981: 25).
Figure 4.2  Policy-action continuum
Policy
Reformulation
Time
Action
»- Reaction
Source:   Barrett and Fudge (1981: 25)
Power is central to the dynamics of this relationship. Implementation in this policy-action model is an iterative bargaining process between those who are responsible for enacting policy and those who have control of resources. In their model, 'more emphasis is placed on issues of power and dependence, interests, motivations and behaviour' (Barrett and Fudge, 1981: 29) than in either the top-down or the
bottom-up frameworks. Furthermore, the policy-action perspective focuses on the factors which affect the scope for action and behaviour of individuals and agencies, as well as how perceptions of that scope are formed. The policy-action model shows that policy is not something that happens at the 'front end' of the policy process. Policy is something which 'evolves' or 'unfolds'. The 'front end' of policy in this sense produces potentialities and principles which change and adapt in practice. In the words of Majone and Wildavsky (1984: 116): 'Implementation will always be evolutionary; it will inevitably reformulate as well as carry out policy/
The 'action' approach emphasizes the extent to which implementation involves more than a chain of command - it is a process which requires that we understand the way in which individuals and organizations perceive 'reality' (cf. Vickers in 3.7.5 on this point) and how organizations interact with other more-powerful or less-powerful organizations in order to attain their goals. This is an issue which is addressed in more depth in theories of interorganizational behaviour (see 4.3.7, 4.6.3). The policy-action model is also supportive of the idea in garbage-can theory that organizations do not have goals in the rational sense, but define them in the process of attaching problems to solutions (see 3.4.5).
4.3.5  Implementation in a managerialist framework
The line between the prescriptions of implementation theory - especially top-down models - and managerialist approaches to the problems associated with implementation 'failure' (sic) is a very fine one indeed. Sabatier and Mazmanian (1979), for example, offer a guide to perplexed managers on how to accomplish policy objectives (see 4.3.8). Many other models of implementation are redolent of the kind of advice and 'how-to' approach which is the stuff of management textbooks. Managerialist approaches to implementation have come to form the dominant 'operational' paradigm in the administration (qua management) of public policy. As the management of the public sector has endeavoured to become more 'business-like', so techniques which were once thought of as 'private-sector' methods have been adopted. We may consider these in terms of three kinds of approaches:
• operational management;
• corporate management;
• personnel management.
PUBLIC POLICY
!
DELIVERY ANALYSIS 475
Operational management
We touched on operational management techniques when we reviewed the use of OR in decision-making (3.8.6). Operational research is als0 applied in the delivery process in terms of 'project management'. It ^ significant that the OR approach to project management develops out of the public sector. In the 1950s the US government adopted two; techniques to manage the Polaris missile project (see Sapolsky, 1972)-
• Critical Path Method (CPM);
• Project, Evaluation and Review Technique (PERT).
Both methods employ the idea of managing large-scale projects in terms of 'networks'; for this reason they are known as forms of 'network analysis'. The aim of CPM and PERT is to 'control' the execution of a project by controlling the network of activities and events which compose the stages of implementation. CPM is a method which aims to identify those activities which are 'critical' to the successful implementation of a project on time. A network is mapped to show the start of the project and the estimated timings involved in moving from one critical activity to another. PERT is a method which predicates that the duration of critical activities are uncertain. PERT systems are programmed on the basis of three types of calculation of this uncertainty: the most probable duration of moving from one activity to another; the most optimistic (shortest) forecast of the duration of a given activity; and the most pessimistic (longest) forecast of duration. PERT analysis is used in the implementation of large-scale projects where there is a high level of uncertainty surrounding the completion of a project. In national and local government network analysis has most applicability in construction projects which involve problems of managing a process with many areas of uncertainty, not least associated with completing the project on time and within budget, given the contingencies of weather, geology, labour relations and inflation, to name but a handful of the problems which can prevent the successful completion of roads, tunnels, power stations, buildings and the rest.
In the OR framework we must also include 'systems analysis'. This considers implementation problems as something that should be analysed in the context of the 'system' which as a whole is involved in the delivery of public goods and services. The systems analyst is interested in how a total sequence of activities, inputs and ouputs, and information flows contributes to the success or failure of projects. Implementation in human 'soft systems' (as opposed to 'hard systems' such as plant and machinery) is viewed as a problem of control and co-ordination. Effective implementation in this model is therefore dependent on such elements as:
Monitor
• defining objectives and formulating a plan;
• monitoring the plan;
• analysing what has happened in terms of what should have happened according to the plan;
• implementing change so as to remedy the failure to realize a goal.
Carter et al (1984: 96) suggest that a successful implementation system involves four types of control:
• co-ordination over time;
• co-ordination at particular times;
• detailed logistics and scheduling;
• defending and mamtaining structural boundaries.
The systems approach lays stress on attaining good levels of co-operation within a 'soft system' by focusing on the importance of 'teamwork' for successful implementation (Carter et al., 1984:106-7).
Corporate management
Unlike OR techniques, the 'corporate management' approach to implementation is a framework which was developed in the private business sector and has been adopted by public-sector managers. As we noted in our discussion of PPBS, corporate management techniques found their way into government through the door opened by reforms to the budgetary process. The corporate management model has, therefore, much in common with PPBS, with its emphasis on the analysis of management problems in a strategic fashion proceeding
Figure 4.3   The corporate management planning cycle
Define goats and objectives
Strengths    — Forecasts — Opportunities
Weaknesses
Threats
Develop strategies
Implement
Formulate action plan
476
PUBLIC POLICY
DELIVERY ANALYSIS
477
through a cycle of defining objectives, planning, organizing, directs and controlling (figure 4.3). ^
Planning involves a process of defining objectives and thence pro. ceeding to develop strategies to attain them. A key phase in the model is so-called SWOT analysis, that is, identifying the internal Strengths and Weaknesses of the organization, and the external Opportunities and Threats. Out of this review the objectives are then considered irr terms of the strategies which are necessary in the light of SWOT. Having set the strategies, the next phase is to consider a plan to implement the objectives. A key consideration here is the relationship of strategy to the structure of the organization and its personnel. For successful implementation the adaptation of the organizational structure may be required. Strategy shapes structure, which in turn shapes strategy (Thompson, 1990: 522). The corporate approach gives a high priority to ensuring that the organization's structure, culture and style geared to the attainment of the 'mission' of the organization, and making clear to everyone in the organization what this 'mission' is, and how they relate to it (Foster, 1993). The monitoring of the implementation of the action plan provides the 'feedback' to management on how objectives are being met and what problems need to be addressed to facilitate success.
Corporate management has been an influential approach in 'new public-sector management', especially in the environment of the 1980s onwards when the pressure for cost-cutting, value for money and a more business-orientated culture has been evident in many industrialized countries (Kooiman and Eliassen (eds), 1987). As in the 1960s and 1970s, the budgetary process has been used to bring about change in the public sector. Performance measurement, for example, is advocated as a technique wherein the disciplines of financial control can be used as a way to achieve objectives more effectively (Thompson, 1990: 582).
❖ Culture as a model of implementation failure and organizational improvement
The culture model has many exponents. Amongst the more widely cited in books about public-sector management are those by Peters and Waterman (1982) and Charles Handy (1976).
Charles Handy, Understanding Organizations, 3rd edn, 1985
Handy argues that there are four main types of culture: power, role, task and person. These turn organizations into coherent tribes, with values, private languages, tales and heroes. And
strong cultures make for strong organizations which do not suit all times, purposes and indi-''duals- Eacn can- ne argues, be good and effective cultures, but they can lead to ineffectiveness and failure.
Each type of culture has a patron god which symbolizes its cultural values and structures:
• power Culture: a web in which decisions are the outcome of power and influence. God: Zeus who rules by impulse and whim.
, Rote Culture: a Greek temple whose culture is one of rationality and function. Bureaucracies have come to typify these cultures. God: Apollo, the god of reason.
. Task Culture: a net-like culture which is concerned with getting tasks done. It is efficient, and adaptive to the demands of consumers. Handy believes it has no patron god.
, person Culture: a cluster culture in which individuals predominate over structure and organization. Rarely found, save in communes. God: Dionysus, the god of the self-orientated individual.
What influences cultures?
. History and ownership;
• size;
4 technology;
t goals and objectives;
• the environment;
• the people.
The Plus Programme and the Metropolitan Police
A sign of the way in which management discourse is being incorporated within what were once thought of as activities far removed from business criteria and culture is found in policing. The policing of London has been subjected to high levels of criticism in the past decades. The London bobby, an international symbol of London and 'community policing', is no longer what it was in the 'good old days'.
In response to the changing 'customer perceptions' of the police, the Commissioner of the Metropolitan Police, Sir Peter Imbert, commissioned a study by management consultants Wolff Olins. It was a decision which he later described as a 'bit of blood-letting in order to ensure that the patient becomes healthy'.
They interviewed over 250 people, examined practices, visited police stations, and surveyed all aspects of police work. What they produced was highly critical of the Met. ft questioned their clearness and effectiveness of purpose, organization, management, attitudes, communication and presentation. The report argued that the police force had a job on its hands to make the service more effective. The Met had five tasks:
* lo feel more united and be clear about what it is there for;
* to improve leadership as well as management systems;
* to adopt a more positive attitude towards the concept of a service;
478      PUBLIC POLICY
DELIVERY ANALYSIS 479
• to improve its communication techniques, both internally and with the outside world; » to improve its appearance in terms of buildings, uniforms, equipment and associated matters.
The report, A Force for Change, was distributed to all personnel and was followed up with a strategy of implementation, the Plus Programme, to make it happen. It included a 'mission statement' designed to encapsulate the aims and purposes of the police. The Commissioner's 1990 strategy statement was to make it clear to both the public (consumers) and to the police service that the Met was determined to see the report's recommendations through. Significantly, this has involved building up from the canteens, rather than down from the Plus team at Scotland Yard. For a review of changing management approaches in the police service see Loveday(1993). ♦
Personnel management
The 'cultural' aspects of the corporate management approach takes us to another important aspect of managerialism in the public sector: the management of people. The quintessential method of corporate management is the setting and the getting of objectives. How people in public organizations and services respond to these objectives which they are being asked (told) to implement is of the greatest importance. Two techniques have been utilized to improve the human side of implementation, performance appraisal and management by objectives:
• Performance appraisal is a method of appraising individuals in terms of his or her 'performance' as set against the objectives of the organization and in the context of the development of the potential of the individual.
• Management by objectives (MBO) is a technique in which objectives are agreed between manager and managed so as to set clear and well-defined goals (Drucker, 1954). The aim of MBO is to facilitate the integration of the goals of the individual and the goals of the organization (Drucker, 1964).
As methods of approaching implementation problems, appraisal and MBO are designed to address the issues associated with changing cultures by changing people. They are a mix of carrots and sticks to create an environment in which administrators/managers and 'street-level' implementers can be encouraged to modify or adapt their behaviour so as to attain corporate as opposed to departmental, individual or professional goals.
The managerial approach as creating 'self-regulating' systems
I fhe development of techniques in delivery which are no longer 'purely' public or administrative ! rnarks me Predorri'nance of managerialist values (concerned with efficiency and effectiveness) in shaping the mix of delivery instruments. Advocates of the need to 'reinvent government' argue that such changes involve a more market-driven 'decentralizing' policy process, in which there is a shift from hierarchy to participation and teamwork in order to manage a more complex society (Osborne and Gaebler, 1993). However, this remixing of government, sectors and instruments may also be viewed as measures aimed at increasing the capacity of government to maintain control (and legitimacy) in conditions of greater social, economic and political complexity and financial constraint. Decentralization in this sense may be viewed as essentially an attempt to 'download' control to more built-in, self-regulating (Kickert, 1993) delivery systems. At one level this has involved the use of new mixes of government, sectors and instruments to promote more 'self-regulating' systems (through markets), and at another level, to create more 'self-regulating' people by applying new management techniques. We may classify these controls in terms of micro and macro self-regulation.
Micro delivery controls
These controls have sought to control the discretion and to change the attitudes of providers. They include:
» human resource management;
• performance-related pay;
• appraisal schemes;
• peer review;
• limitations on tenure.
These techniques signal a move away from improving the rationality and expertise of bureaucrats and professionals as a way of improving services, to an approach which seeks to improve the methods of controlling the non-policy-making stratum so as to make it more responsive to the demands of 'taxpayers' qua customers. This move from rationality of production to the sovereignty of consumers and control by decision-makers may be read as subjecting bureaucrats and professionals, long held to be different from workers and managers in the private sector, to forms of control which have been familiar in the profit sector. Thus the fllpside of consumerism is the 'proletarianization', deskilling and 'de-professionalization' (see Braverman, 1974; Clegg and Dunkerly, 1980) of bureaucrats and professionals.
Macro delivery controls
These techniques have sought to change the organizational and wider social and political contexts of service provision. We have reviewed these elsewhere (see 3.10.6, 4.3.5). They include:
• increasing the number of providers by measures to increase competition: — privatization;
480      PUBLIC POLICY
DELIVERY ANALYSIS 481
— contracting-out.
• measures to create a more business-like structure of delivery:
— internal markets;
— charging for services.
• direct controls over costs and size of public sector:
— incomes policies for the public sector;
— budgetary controls through performance measurement;
— cuts in numbers of civil servants and professionals;
— financial management.
• monitoring and improving performance:
— enforcement of performance indicators;
— institutional audits;
— departmental audits;
— league tables of performance;
— organizational learning.
Micro techniques aim to change the context within which individuals and groups work in the public sector, whereas the macro techniques are aimed at changing the organizational environment by bringing market, or quasi market, forces to counteract the supposed bias in bureaucratic hierarchical organizations and professional and 'street-level' discretion, whilst at the same time making policy delivery more open to the demands and requirements of 'consumers'. ♦
4.3.6   Implementation and policy type
Is implementing a missile development programme the same as a human service programme? Is it easier to put a man on the moon than put a homeless family in decent accommodation? The trouble with so much of the managerialist and 'rationalist' models of implementation is that they derive from a notion of decision-making which fails to take account of the fact that human problems are varied in their nature and complexity. Putting a man on the moon or developing a missile system does not serve as a convincing model for policy areas-in which defining goals, building consensus and acquiring resources-are infinitely more problematic. A more useful approach, therefore, may well be to focus on the relationship between the type of policy and the factors which may impact on the implementation process. Lowi's (1964, 1972) categorization of distributive, regulatory and redistributive policy types, for example, has been applied to implementation by several analysts (see Van Meter and Van Horn, 1975; Ingram, 1978; Ripley and Franklin, 1982; and Hargrove, 1975).
One of the first attempts to analyse implementation, by Van Meter and Van Horn (1975), took the view that the study of implementation needed to take account of the content or type of policy. Drawing on Lowi's work on policy types (see 2.5) Van Meter and Van Horn argued
that the effectiveness of implementation will vary across policy types and policy issues. The key factors operating in implementation -change, control and compliance ~ suggested to them that where there was a high degree of consensus and a low or marginal amount of change required, policy implementation was likely to prove more successful.
Ripley and Franklin (1986) suggest that relative difficulties of success in implementation is low where distributive policies are concerned, moderate in regulative policies and low in redistributive policies. The various policy areas have different patterns of relationships, which means that in the redistributive areas there is more bargaining and politicking than in the distributive areas, where a higher measure of control may exist. Ingram (1978) has related this variation in the pattern of relationship in terms of decision-making costs, structure of statute, appropriate approach, criteria for evaluation and critical variables.
*> Gordon Chase, 'A framework for implementing human service programmes: how hard will it be?', 1979
i the arguments of Gunn we see a general model of 'perfect implementation'. However, for the j   provision of human services, welfare, education, health (that is services directed at people), how can this general model be made more specific to policy types? Gordon Chase provides a framework to examine the obstacles confronting the implementation of human services: lead ontrol, methadone maintenance, and prison health reform. Using these three examples he onstructs the following framework which, he suggests, may be used by implementers as a ■nap to enable them to identify the key problem areas for their programmes. What is significant I   about his analysis is that he demonstrates that in implementing human service policies, the |   context is complex and uncertain, even in such matters as space and equipment resources.
It'
I   Difficulties arising from operational demands
'  Who are the people to be served?
What is the nature of the service to be delivered? ]   3 What are the potential distortions and irregularities in the population?
-  Is the programme controllable? Can it be measured? Are any parts not controllable?
Difficulties arising from the nature and availability of the resources required to run the programme
Money: what are the limits on funds, and what are the prospects for more? 2  Personnel: are they in place, and with the right qualifications? Does the programme have enough?
482      PUBLIC POLICY
DELIVERY ANALYSIS 483
3 Space: has the programme got enough? Will it need more?
4 Supplies and technical equipment: are they available and usable? Is technology important')
Difficulties arising from the programme manager's need to share authority with or retain the support of other bureaucratic and political actors
1 Overhead agencies: how many will a manager have to deal with, and will they be support-ive?
2 Line agencies: how many are there involved and can the personnel work together? Are lines of responsibility clear?
3 Elected politicians: can they help or hurt?
4 What higher levels of government are involved?
5 Private-sector providers: how badly wifl the programme manager need private-sector pro-viders? How well will the programme manager be able to control the private contractors?
6 Special interest groups: what are their interests and political influence?
7 The press: will the programme have high visibility? Could the media do any good/harm? ♦
A key factor in policy type analysis is the issue of the organization of a policy subsystem (see Sabatier, 1986a; and 4.3.8). In the case of a simple problem - such as putting a man on the moon - the subsystem (NASA) was tightly organized, influential and well resourced. Putting a man on the moon was a triumph of rational comprehensiveness in both capitalist and communist societies. However, PERT, CPM, PPBS, Zero Budgeting and the rest, when applied to more complex social and economic problems and policies, was less than a spectacular success. Whereas putting a man on the moon was a well-defined goal and had a narrow policy subsystem, the problems of man (and woman) on earth have long exhibited fuzziness and subsystem complexity. As Mayntz (1993) notes, for example, policy sectors differ in the degree to which they have well-organized subsystems capable of resisting guidance or direction. Some organizations will be better informed, better financed, more powerful, or more independent than others. Implementation has therefore to be set in the context of types of policy and political priorities but also in terms of interorganizational relationships.
4.3.7   Inter-organizational analysis and implementation
A major focus of implementation studies has been the issue of how organizations behave, or of how people behave in organizations. However, if we accept that implementation is a process which involves a 'network' or multiplicity of organizations the question arises as to how organizations, interact with one another. Two approaches have informed and framed this debate.
Power and resource dependency
This argues that the interaction of organizations is a product of power relationships in which organizations can induce other less-powerful and more-dependent organizations to interact with them. In turn, those organizations which are dependent on other more resourceful organizations have to engage in strategies of working with more powerful organizations so as to secure their interests and maintain their relative autonomy or space within which to act (Aldrich, 1972, 1976; Kochan, 1975; Yuchtman and Seashore, 1967).
If A cannot do without the resource mediated by B and is unable to obtain them elsewhere, A becomes dependent on B. Conversely, B acquires power over A ... dependence is an attribution of the relation between A and B, and not of A or B in isolation. It is thus possible that A may be dependent upon B, while having power over C. (Aldrich and Mindlin, 1978:156)
Organizational exchange
This takes the position that organizations work with one another so as to exchange what is to their mutual benefit. Levine and White (1961) argued that the defining characteristic of exchange between organizations is that it is voluntary interaction which is undertaken for the realization of the goals and objectives of the participants. Whereas in the power-dependency model the organizational relations are based on dominance and dependence, interaction based on exchange is structured by mutual interests (Bish, 1978; Tuite, 1972; White, 1974). Even though an agency may be dependent on central resources, it may also be the case that the centre is dependent on the local agency for implementing policy goals. As Scharpf argues:
While the seemingly dominant party may exercise hierarchical authority or control over monetary resources, it may, at the same time, be fully dependent upon the specialist skills, the clientele contacts and the information available only to subordinate units ... In short, unilateral-dependence relationships which are stable over time may be more rare, and mutual dependency more frequent, than the ubiquitous nature of hierarchical authority and unidirectional flows of budgeted resources in inter-organizational relations might suggest. (Scharpf, 1978: 359)
♦ K. Hanf and F.W. Scharpf (eds), Interorganizational Policy-making, 1978
This book comprises a number of key readings in interorganizational analysis based on European and American case studies. At the time this book was published there was increasing concern with the apparent 'overload' taking place in liberal democracies and alarm at the prospects of 'ungovernability'. By examining the growth of organizational complexity the authors explore the structural features of policy formation and implementation which influence the
484
PUBLIC POLICY
capacity of governments to solve problems. The theme which the studies investigate is the w in which the existence of organizational networks of exchange and dependence impose iim"^ on co-ordination and control in the making of policy and the attainment of policy goals. ♦
Benson (1975, 1977a, 1977b, 1982) has argued that, in order to under-stand the way in which interorganizational relationships operate we need to consider the network of interests within a policy sector, which he defines in terms of: 'a cluster or complex of organizations con-nected to each other by resource dependencies and distinguished from other clusters or complexes by breaks in the structure of resource dependencies' (Benson, 1982: 148). Benson's approach also stresses the importance of 'deep structure' of the kind suggested by Bachrach and Baratz, which takes into account the bias which exists in a given policy sector and its constituent structures: administrative networks-interest group networks; and the rules of structure formation. Benson argues that the rules of structure formation within a policy sector -which incorporates the ideas of agenda-setting which we have encountered in non-decision-making theories and in critical theory -need to be explored in relation to the administrative and interest-group networks (see 2.7.2). (Benson's ideas have been applied by Rhodes in a network approach to central-local relations: see 210.2 and Rhodes, 1985,1988.)
Applying the theories to public policy
The application of interorganizational analysis developed by Aldrich, Benson and others has been applied in the study of public policy through the work of Hjern and Porter (see also Hjern et al., 1978; and Hjern and Hull, 1982). Adopting a bottom-up framework Hjern and Porter (1981) argue that implementation should be analysed in the context of 'institutional structures' composed of 'clusters' of actors and organizations. Programmes may be conceived as being implemented in 'pools' of organizations. A programme will involve a multiplicity of organizations of various kinds: national and local, public, private and voluntary, business and labour, and so on. Programmes are not implemented by single organizations, but through a matrix or set of organizational pools (see figure 4.4):
Organizations (Ol ...) participate in parts of several programmes (PI ■■• )■ Persons from within a single organization attempt to adjust these parts of programmes to meet overall organizational objectives. The vertical oval enclosing the parts of the three programmes implemented by 01 is the area within which an organization rationale is practiced. Persons identifying with a programme (PI) pursue a programme rationale within the horizontal ova! trying to adjust the contributions of the parts of a number of organizations to meet the needs of the programme. Persons occupying
DELIVERY ANALYSIS 485
Figure 4.4   Organizational 'pools' programmes
P1
P2
P3
01
Source:  Hjern and Porter (1981)
02
03
Organizations
positions at any of the nine points of intersection ... have to serve 'masters' both in programmes and in organizations. Failure to identify implementation structures as administrative entities distinct from organizations has led to severe difficulties in administering the implementation of programmes. When a new programme is enacted, it is assigned to a single organization ... If there is a failure, the programme is assigned to another organization, or the department head is fired, or both ... such reactions may be inappropriate and dysfunctional. (Hjern and Porter, 1981: 217)
Such arrangements whereby a programme is in the hands of a multiplicity of organizations give rise to a complex pattern of interactions which 'top-down' frameworks (and practices) fail to take into account, with the consequence that their theories do not satisfactorily explain implementation, and that in practice programmes are not successful. Implementation structures are not organizations so much as 'localized implementation structures'. Hjern and Porter believe that the implications for the theory and practice of public policy of this model are extensive, and should prompt us to revise our conceptualization of the implementation process in terms of both analytical and normative frameworks. As we shall see below (4.3.8), the theoretical implications of their model have been taken on board in more recent theories of implementation.
486     PUBLIC POLICY
I
DELIVERY ANALYSIS 487
4.3.8   Implementation - towards a synthesis
We have already encountered Sabatier's ideas in other parts of th' book. The model he advances for viewing the policy-making procesS may also be extended into the study of implementation. For Sabati S the policy-stages approach does not help in understanding the policy making process because it splits it up into a series of artificial and unrealistic sections. From this point of view, therefore, implementa tion and policy-making become virtually one and the same process His early contribution to the study of implementation came with a highly top-down account written with David Mazmanian (Sabatier and Mazmanian, 1979). This work has, perhaps, been more widely tested (see Sabatier, 1986) than most theories and is justifiably seen as one of the most developed top-down models. Their work endi-.,-,--oured to synthesize the ideas of both top-down and bottom-up theorists into a set of six sufficient and necessary conditions for the effective implementation of legally stated policy objectives. These comprised:
• clear and consistent objectives, so that they can provide a standard of legal evaluation and resource;
• adequate causal theory, thus ensuring that the policy has an accurate theory of how to bring about change;
• implementation structures that are legally structured so as to enhance the compliance of those charged with implementing the policy and of those groups who are the target of the policy;
• committed and skilful implementers who apply themselves to using their discretion so as to realize policy objectives;
• support of interest groups and 'sovereigns' in the legislature and executive;
• changes in socio-economic conditions that do not undermine the support of groups and sovereigns or subvert the causal theory underpinning the policy.
This model has been applied in a wide range of US and European case studies which demonstrated the usefulness of the framework to empirical research. Their framework was subsequently modified in the light of this research (Sabatier, 1986: table 1), and an evaluation of the bottom-up case as developed by Hjern et al Sabatier (1986) proposes that a synthesis of the two positions is possible drawing on the insights of Hjern et al into the interorganizational dynamics of implemention and its network/matrix form, and the top-down focus on how institutions and social and economic conditions constrain behaviour. This is accomplished through the use of the model we have discussed in Fart Two (2.10.4): implementation as taking place within the context of a
policy subsystem, and bound by 'relatively stable parameters' and 'events external to the subsystem'. Central to this model of implementation is the notion that it is an ongoing part of policy-making within ACs ('advocacy coalitions'), and that the deep, near core and secondary aspects of the belief systems of implementation coalitions should be a main focus of analysis.
The model takes into account, therefore, the concerns of the bottom-up approach in that it emphasizes the network which structures implementation, whilst at the same time stressing the importance of top-down considerations within the system, including the beliefs of policy elites and the impact of external events. Implementation in this sense may be conceptualized as a learning process (Heclo, 1974: 306; cf. Browne and Wildavsky, 1984). The aim of the approach is to analyse the way in which policy learning takes place amongst ACs, and to set out those institutional conditions which are most appropriate or conducive to 'learning' and change in core beliefs.
Although the incorporation of the bottom-up concern with subsystem dynamics makes for a comprehensive model of implementation, the advocacy-coalition approach does not confront the normative dimension of the bottom-up versus top-down argument. Whereas the bottom-up approach, for example, is interested in the 'street level', the AC model is primarily focused on the policy elite. Policy learning, for Sabatier, is something which essentially takes place within 'the system' and its policy subsystems- The framework is designed to analyse institutional conditions in which such learning changes the policy core. The more radical implications of the bottom-up approach question whether success in implementation is more to do with wider social learning than policy learning by subsystem coalitions. The top-down position adopted by Sabatier and Mazmanian in their original model is concerned with the relationship of decisions to attainments, formulation to implementation, and the potential for hierarchy to confine and constrain implementers to achieve those legal objectives defined in the policy. Bottom-up approaches, however, are predicated on the significance of the relationship between actors involved in a policy or problem area and the limitations of formal hierarchy in such conditions. As Sabatier and Mazmanian (1979) make clear, their model is concerned with effective control and compliance; bottom-uppers, however, are in various ways concerned with interactions, conflict, power and empowerment. In their desire to construct a comprehensive model, the authors ignore the possibility that what they are trying to combine are, in a Kuhnian sense, incommensurate paradigms. This is most apparent when we move from using the models analytically to using them in a normative way. Top-downers are essentially working within
488
PUBLIC POLICY
a frame which focuses on decision and power, and the potential f0r decision-makers to effect change in society is regarded as a problem of developing effective modes of control and elite learning. The bottom up models are suggestive of implementation as a process of policy making and (possibly) empowerment of those seen as targets of decisions. The preference of the top-down models is for tiers, hierarchies control and constraints, whereas for bottom-up models, spheres, net' works or markets constitute a more desirable state of affairs.
The clash of frameworks and the values and beliefs which they contain provides competing frames of analysis and prescription. The synthesis thus serves to produce a consensus which is not there. If the aim of policy analysis is (as Lasswell argued) to clarify values, then the synthesis advanced by Sabatier only serves to muddy the waters More satisfactory on this score is the synthesis which Sabatier (1986) commends as a guide to policy-makers, but as inapproriate as an explanatory model of the policy process (Elmore, 1985). Elmore argues that a variety of frameworks need to be deployed in analysis and implementation: 'backward-mapping' (bottom-up) and 'forward-mapping' (top-down); and that policy-making, if it is to be effective in implementation terms, must adopt multiple frameworks. What this means when applied to the study of the policy process is that analysis must aim not so much for a synthesis as a sensitivity to the frameworks (in Vickers's sense, the value, reality and action judgements) of theorists, policy elites and those at the 'street level'.
This way of looking at implementation may be derived from the kind of ideas proposed by critical policy analysis which we reviewed earlier in this book (2.7.3, 3.8.8). The use of models as Tenses' (in Allison's sense) through which we can explore implementation is suggested by Elmore's (1978) categorization of four sorts of implementation model: systems management; bureaucratic process; organizational development; and conflict and bargaining. Elmore argues that models of implementation should not be regarded as rival hypotheses which could be empirically proved, but as ambiguous and conflicting frames of assumptions.
This idea of models as essentially incomplete, partial perspectives on complex problems and realities is an issue which has been most fully addressed in the work of Gareth Morgan (1986/1993). Morgan maintains that if we are to understand complexity, we are required to adopt a critical and creative approach to thinking in terms of models - or 'metaphors'. From this point of view an attempt to bash different models together to create a synthesis based on the strengths of two different frameworks is a somewhat misguided exercise. Developing
DELIVERY ANALYSIS
489
his approach under the influence of 'postmodernist', 'constructivist' theory, Morgan takes the view that the analysis of complexity involves not a vain quest for synthesis but, on the contrary, a recognition of the differences, partiality, incompleteness and distortion which is inherent in human knowledge and discourse. Metaphors/models/ theories create insight, or ways of seeing, but also ways of not seeing. For Morgan there can be no single metaphor which provides a general theory. In terms of implementation, this means that the problems of implementation may be constructed in a variety of different ways. Each approach or theory gives some insight into a particular dimension of the reality of implementation, and, as in the case of the somewhat constrained debate on top-down versus bottom-up, both approaches and their hybrids and variants provide us with part of the picture. As students of public policy we should aim to become skilled in the art of reading the frameworks which are employed in the theory and practice of implementation in the contexts in which they occur. As Sabatier (1986) concedes, the different approaches have comparative advantages as explanations in different contexts. Applying Morgan's metaphors, for example, we should recognize that frames will disclose or illuminate various dimensions of implementation. No one metaphor provides all the answers.
In terms of the policy sciences, the approach which approximates most closely to the method which Morgan uses in organizational analysis is Lasswell's idea of contextual mapping (see p. 447). Thus, for
Figure 4.5  Metaphors of implementation failure
Machine metaphor
Result of poor chain of command - problems with structures and roles
Domination metaphor
Result of labour/ management conflict
Psychic metaphor
Result of subconscious forces - groupthink/ ego defences/repressed sexual instincts
Organism metaphor
Result of 'human relations' or the 'environment'
Implementation failure'
Autopoietic metaphor
Result of a
'self-referencing'
system
Brain metaphor
Result of poor information flows - or 'learning' problems
Culture metaphor
Result of the 'culture' of the organization
Power metaphor
Result of power in and around the implementation process
490      PUBLIC POLICY
example, a Lasswellian approach to implementation would involve mapping the participants/stakeholders, their perspectives, situations values and strategies, and their desired outcomes and actual effects The method recognizes that implementation has a specific context in terms of the values and institutions involved in a given problem Furthermore, as with the Morgan approach, the Lasswellian orientation would also stress the idea that analysis is fundamentally a 'learning' activity which should lead to the enlightenment of participants, Mapping the context of problems offers the possibility of understanding the multiple dimensions of knowledge, beliefs, power, meaning and values which frame policy-making and implementation. The quest is not for Sabatier's 'promised land' (sic) - a general theory - but a clarification of the values of theorists and practitioners.
❖ Sandra Kaufman, 'Decision-making and conflict management', in Bingham ef a/., 1991
We examined Kaufman's ideas in 3.7.7. Her model seeks to explain decision-making and implementation in terms of theories of cognition and conflict. As we noted, her model does provide - like Sabatier's - a way of synthesizing a number of approaches to analyse implementation. Her main point is that implementation takes place in a situation in which there is necessarily conflict between numerous divergent interests, actors and organizations. This agrees with the arguments we encountered in 4.3.7. She believes that the conflict which takes place between different stakeholders can be both positive and negative. However, if such conflict is not managed, Kaufman argues that it can
have destructive consequences on individuals, an agency, and the constituents it serves. At a minimum resources are wasted and high opportunity costs can be incurred. At its worst, conflict renders individuals and organizations dysfunctional and deprives beneficiaries of these organizations' services. The more complex a decision situation - in terms of the number of parties, issues at stake, consequences unfolding over time and affecting large groups - the more important it becomes to all involved that conflict be managed. (Kaufman, 1991: 129)
She offers a number of recommendations to improve implementation and decision-making in such conditions (see Kaufman, 1991; 131-2). ❖
DELIVERY ANALYSIS 491
4.4   Delivery systems
4.4.1 Mixes
Modes of delivery or 'systems' of policy delivery have become a central concern of analysis of and in the modern public sector. This focus on the increasingly diverse intergovernmental and interorganizational network of delivering public goods and services has been considered elsewhere in this book (see 3.4.6). As Kaufman (1991) observes (above), implementation now involves a large number of stakeholders and the potential for a good deal of conflict and 'dysfunctionalities'.
From being a neglected area of interest in the 1970s and early 1980s (see Hogwood and Peters, 1983:165), the study of the 'technology' of delivery has come to assume a dominant position in the literature of public sector management. This is especially the case in the context of the 'complex' conditions facing modern governance (Kooiman (ed.), 1993) and the changing 'architecture' of the state in modern society (Dunleavy, 1989). As Self notes, for example, with regard to the delivery of welfare services:
The provision of welfare can be regarded as a complex mixture of contributions from four sources: government, market, voluntary organizations and individual households. (Self, 1993:121)
(See also Rose, 1986, on the idea of the welfare 'mix'.)
In this section of the book we examine these delivery systems in terms of the way in which public goods and services are now provided through an ever more complex and diverse set of institutions and instruments. We term these sets or combinations delivery 'mixes' to convey the idea that policy fields are composed of a plurality of actors, institutions and organizations, modes of enforcement and values. Policy fields - such as health, housing, economic development and so on - may be viewed as a mixture of relationships which change and vary over time and space - in both unitary and federal political systems. As Rhodes observes of the former:
governments increasingly resort to a variety of instruments for pursuing their policies. Functions are not allocated to general purpose governments ... but to special purpose authorities. Institutional 'adhocracy' is the order of the day, a process which generates conflicts between agencies compet-i ing for 'turf and between central and local governments which resent
J being bypassed. Government has not been rolled back but splintered and
1 politicized, a process which can only frustrate the attempt to control through
492      PUBLIC POLICY
DELIVERY ANALYSIS 493
centralization. Such fragmentation not only thwarts control and fuels pol" slippage (or deviancy from central expectations) but it also increases p ^' ernmental complexity. ° v~
(Rhodes (ed.), 1992: 330)
This fragmentation also, as he notes, creates new problems for control and accountability in a democratic society {see 4.4.8). Citizens now confront an often bewildering array of agencies responsible for the provision of public services. Simple hierarchies and tiers have given way to delivery systems which use a mix of governmental relationships, new 'partnerships' between the public and private sectors, market mechanisms and 'marketized' public policy; and new roles are being defined for the voluntary sector and 'the community'. In place of the relatively ordered patterns of relationship which existed ten years or more ago - like a cake made of weU-defined sections - industrial societies have become more jumbled up, and resemble not so much a Battenberg as a marble cake. As we take a slice through our cake we find that the arrangement or mix varies considerably between policies. Another metaphor which captures the idea of a mix is the notion of a 'field' or 'space' as used by Lewin (see 3.7.7) in which, over time, the relationship of parts of the policy field change and interact with one another.
In this section we examine four 'mixes':
• governmental mix (4.4.2);
• sectoral mix (4.4.3);
• enforcement mix (4.4.4);
• value mix (4.4.5).
The first two provide the institutional and organizational setting of policy delivery - that is, governmental forms and the interaction of the public, private, voluntary and community sectors. The enforcement mix is concerned with the mix of approaches to the problems of gaining compliance in public policy. (Making it stick!) Finally, we examine what Colebatch and Larmour express as the underlying values and framework of meaning (Colebatch and Larmour, 1993: 108), which interact with institutional, organizational and enforcement mixes. The resultant 'mix' in delivery systems may be viewed as a blend or compound of 'market', bureaucratic'/'hierarchy', and 'net-work'/'commuruty' models of organization (see 1.8). By using the term 'mix', therefore, we are suggesting that thinking in terms of clearly defined sectors or modes of co-ordination is not, in practice, helpful, since in the real world there is considerable ambiguity and overlap between them. As Colebatch and Larmour point out, the world
can't be divided into 'market', 'bureaucratic' and 'community' organizations, because there are likely to be elements of all these in any organization. It will not be the same in all organizations ... and an organization in the 1990s may operate quite differently from the way that organization worked in the 1950s. The task is to identify the nature of the mix, not to place the organization into one box or another. (Colebatch and Larmour, 1993: 80)
So, too, in public policy our task is not to put policies into boxes, but to analyse and map the mix of elements that make up particular policy fields. If we envisage delivery systems as a function of the mix of market, bureaucratic and community forms, we might represent them in the context of a triangular relationship which changes over time and space (figure 4.6).
Figure 4.6   The delivery mix
Market
Hierarchy/ bureaucracy
Community/ network
Thompson, et ai. (eds) (1991:15-16) suggest that we may envisage this in terms of three 'analytical lights' which focus on different aspects of governance (figure 4.7). In practice the mix of analytical lights overlaps and combines to provide 'hybrids' or 'plural forms' (see Bradach and Eccles, 1991).
Figure 4.7  Analytical lights on delivery
Market-Hie rarchy-Network Source:  Adapted from Thompson el al. (1991)
494      PUBLIC POLICY
DELIVERY ANALYSIS 495
4.4.2  The governmental mix
In the first instance, we must consider the territorial dimension. What part or level of government is responsible for the delivery of a pro-gramme? How is responsibility shared in terms of administratíve and financial arrangements? The mix of levels will, of course, vary from policy arena to policy arena. The mix will also be determined by the political and constitutional traditions and arrangements between centralized unitary states and decentralized federal systems. Although political systems may share policy goals, the level of government which is deemed responsible for actually providing a service varies widely between:
• national;
• regional/state; 9 local;
• neighbourhood.
OECD surveys (1990,1992,1993a), for example, which profiled the level of system delivery in its member states, illustrates that, in the case of education, health and welfare policies, a very diverse pattern of governmental responsibility exists even between countries that share unitary or federal characteristics. A service such as education may be highly decentralized, as in the USA, or highly centralized, as in France and Greece. Some countries have sub-national tiers of government and administration which may have responsibility for regions or states. The role of local government also varies considerably, from having tight control on service delivery to others where the local authorities' powers are weak. Furthermore, the delivery mix (local-public-private-voluntary) has, in the European context, long been highly diverse compared with Britain. However, as Batley and Stoker note in their review of changes in service delivery in European local government:
the language of reform has much in common in the countries studied (France, Germany, Italy, Britain, Holland, Ireland, Denmark, Norway, Spain and Portugal) ... In different ways the issues of choice, proximity to the citizen (or customer) and reduction of bureaucratic control, complexity and uniformity are being addressed. In some cases these reforms are being pursued in the context of an ideology of efficiency of competition, market and business ... (Britain, Holland, Germany and to some extent Italy)... In other cases the ideology is more to do with the citizen's access to a democratic state (Sweden, Norway, Spain, Portugal). Most have elements of both sorts ... the monopoly provision of certain defined, mainly social services by British local government has been unusual ... [since] in every other European state there has been a long history of partnership ... between levels of government, between communes and between the private and public sector ... (Batley and Stoker, 1991: 215)
An important trend in the delivery mix has been with regard to the neighbourhood and community decentralization which began in the 1980s (see Hoggett and Hambleton (eds), 1987) and has continued in the 1990s in Britain and other countries (see Cannon (ed.), 1990; Osborne and Gaebier, 1992; Batley and Stoker, 1991; Burns et ah, 1994). Two notable examples of this decentralization of service delivery are the London boroughs of Islington and Tower Hamlets who have under respective Labour and Liberal control developed multi-service neighbourhood offices to manage services, and forums/committees which provide for local participation, access and accountability. These two authorities have led the way in moving from centralized bureaucratic structures towards local management networks which are closer to citizens/consumers/clients/voters. At the time of writing, a considerable number of British local authorities are endeavouring to devise schemes which decentralize the management of service delivery (see Burns et al.r 1994). Given the experimental nature of the reforms, the range of models which have emerged is very pluralistic. The leading researchers in the field have suggested a four-point 'ideal type' of neighbourhood decentralization (see Burns et ah, 1994: 88):
• localization: the physical relocation of services from a centralized to a neighbourhood or 'patch' level;
• flexibility: the promotion of more flexible forms of management and work organization through multi-disciplinary team working, multi-skilling, local general and corporate management;
9 devolved management: the devolution of decision-making powers to service delivery managers and staff;
• organizational culture and change: the reorientation of management and staff values to promote quality of service and local empowerment.
Hoggett (1987) has argued that this shift towards neighbourhood modes of service delivery must be viewed in the context of a wider process of 'post-Fordism' in modern society. Control through a Weberian/ Taylorist/Fordist hierarchy, it may be argued, is giving way to less-bureaucratic, flatter, more-fragmented 'post-modern' structures (see Clegg, 1990:180-207). Hoggett and his co-workers believe that in this new context public-sector organizations will go the way of business and industrial organizations:
the conditions now exist for entirely new forms of service organization which allow both for much greater degrees of operational freedom and for centralised strategic control. The key question is how these conditions are shaped by political choices and strategies, for it is the struggle around these issues which will decide the relative weight given to internal as opposed to external decentralisation, where strategic command is to be (in
496      PUBLIC POLICY
DELIVERY ANALYSIS 497
local or central government) and so on. Our vision of a reformed puhr service seeks to replace complex bureaucracies with far more internal? devolved structures ... Within the context of a local authority, interna
devolution gives service managers and staff the power to deliver a simultaneously frees the centre from absorption in administrative detail In place of departmental hierarchies a new kind of organization emere in which there is a strong but lean centre with an outer ring of devolve service delivery units. (" (Burns et al., 2994: 272)
The mix of governmental levels in a given policy arena will clearly structure the context within which interorganizational relations take place. One level will be dependent on another for resources, and the pattern may be such that the notion of 'levels' and 'tiers' may be inappropriate to describe a relationship which is more akin to 'spheres' and 'networks' (Rhodes, 1981). Increasingly the pattern which has emerged in the mix of governmental organizations involved in public policy in Europe, America and elsewhere is intergovernmental and interorganizational in form (Dommel, 1991; Batlei and Stoker, 1991; see 3.4.6, 4.3.7, and 4.6.3). As with other mixes, th< patterns of interaction and power relations between levels in industrialized societies is becoming ever more complex and diverse within and between political systems - as the experience of multi-service neighbourhood delivery illustrates. Indeed, it is this complexity which perhaps constitutes the main common denominator in the experience of public policy implementation in modern (perhaps post-modern) conditions. The managerial and political pressures which influence the mix of responsibility creates some difficult and intractable problems. As Brian Smith notes:
The managerial needs of national organizations can only be met by delegating authority to field officers. Politically, threats to integration from culturally distinct communities can only be met by a measure of devolution ... However, having decided that an administrative presence in the regions and localities is needed does not conclude the process of political choice. A decision still has to be made on how the administration is to be carried on, and whether it is to be politicized. There are too many cases of state intervention in the fields of income maintenance, transportation, public utilities and health care that have been centralized after a period during which they were the responsibility of regional and local governments, and too many instances of government functions which are under central administration in one state and local in another, to believe that there is some politically neutral formula for the territorial allocation of governmental powers. The distribution of power between levels of government, as well as the choice of institutions for decentralization, are the outcomes of political conflicts at the centre which originate in group and class interests which sometimes have a territorial identity but which also unite and mobilize people regardless of region. (Smith, 1985: 201-2)
The mix of governmental levels of responsibility is therefore an issue which prompts the need for both centralization on the one hand, to secure control - of national finances in particular - whilst, on the other, requiring decentralization in order to secure managerial benefits. Centralization in order to decentralize, and decentralization as a mode of centralization thus constitutes a key paradox in the allocation of responsibilities in the modern state (Metcalfe and Richards, 1992: 77-94). As Smith notes (above), this dilemma is not open to one neat neutral solution. In part, the reason for the attractiveness of focusing on the sectoral mix is that it provides a mode of delivery which overcomes the problem of which level of government should be responsible.
4.4.3   The sectoral mix
The mix of levels (or spheres) of government must also be considered alongside the sector which is involved in the delivery of public goods and services. Here again, the pattern is complex. The relationship between the public, private and voluntary sectors, for example, is one which has undergone considerable change in the last decade or so. Services may comprise a mix between public and private responsibility, as well as between the voluntary sector and 'community' agencies which may have a role in delivering a service.
The public-private mix: partnerships
The setting-up of partnerships between the public and private sectors (or PFPs) has occurred in a number of policy areas. These include:
• infrastructure developments;
• urban renewal;
• regional development;
• training and education;
• the environment.
The reasons for the expansion of PPPs involve understanding the advantages which government obtains in terms of private-sector finance and management expertise, and the financial and other benefits which may be gained by the private sector. Kouwenhoven (1993: 125-7) suggests that we consider the development of successful PPPs in the context of three sets of conditions:
Start conditions for a PPP
— interdependence between the two sectors;
— convergence of objectives.
498      PUBLIC POLICY
Given these conditions in which both sides need one another an(j share objectives, two secondary conditions or 'interlinking' mecha nisms are required:
• Secondary (mterlinking) conditions
— the existence of a network of communication channels between the public and private sectors concerned;
— the existence of a broker to facilitate negotiations.
• Project conditions
In terms of the project in process, the PPP must meet a formidable list of conditions:
— mutual trust;
— unambiguity - and recording - of objectives and strategy;
— unambiguity - and recording - of the division of costs, risks and returns;
— unambiguity - and recording - of the division of responsibilities and authorities;
— phasing of the project;
— conflict regulation laid down beforehand;
— legality;
— protection of third parties' interests and rights;
— adequate support and control facilities;
— business- and market-orientated minking and acting;
— 'internal' co-ordination;
— adequate project organization.
This model - as with models of perfect implementation - draws attention to the sheer number of things that can, do and will go wrong. The PPP is a 'learning process' for both the private and public sectors; the experience of partnerships which have not proved as successful as both sides hoped occupy a good deal of the literature on the topic. The problems which arose with the London Docklands development scheme illustrates only too well that PPPs do not present an easy route out of the apparent dead-end of top-down implementation in public policy (Massey 1993: 167-8). Even so, as long as the two sectors find mutual advantage in the risks which PPPs involve for both sides, the extension and development of partnership models of delivery will, continue to grow. This is particularly the case in big projects for which partnerships are possibly the only way in which problems (such as urban regeneration) can be addressed and opportunities (such as securing the Olympic Games) may be grasped. For smaller-scale social projects, the use of partnerships in local government offer possibilities of securing additional funds and expertise from the private sector, which stand to gain in profits as well as the PR that results from the
DELIVERY ANALYSIS 499
acceptance of social responsibilities and an ethical approach to business.
The voluntary sector
The involvement of the voluntary sector in social and other policy areas is a matter of growing interest. The voluntary sector has been historically very important in the history of the welfare state. Voluntary groups seeking to do good works for the poor and needy have a long and distinguished history (Butler and Wilson, 1990: 9-14). Religious organizations in particular have made an enormous contribution to the development of charitable institutions which, until the emergence of the welfare state, served as a principal provider of many 'social services' as well as education. Some services have remained in the voluntary sector: notably fire services and sea rescue services in a number of countries, and religious institutions have retained their involvement in education and other social activities. The growth of the welfare state and mass education made the existence of the voluntary sector apparently less necessary in the twentieth century. However, in recent years, the role of the voluntary sector has come to the fore as the state is no longer capable or no longer desires to provide the range of services that was once expected of it (Mellor, 1985).
As a sector the terms 'voluntary' and 'non-profit' are something of a misnomer. Their role has been described more extensively and more accurately in terms of being private agents of public policy (Streeck and Schmitter, 1985). Voluntary organizations employ people on a permanent basis, and they have to 'make' money and run on a sound financial footing. Charity is big business. This has led some to propose that the idea of the 'voluntary sector' should be replaced by the notion of a 'third sector' (Osborne and Gaebler, 1992: 44) which is recognized as having a major role in the delivery of goods and services. The responsibilities which have been devolved to this sector as a result of cuts and anti-welfare-state policies mean that it composes a key ingredient in the mix of delivery systems in the modern state, even in those societies which have traditionally had a large welfare state, and even more so in those (such as the USA) where state welfare programmes are less developed,
How this restructuring of the welfare state and the role of the voluntary sector is interpreted depends on where analysts and commentators are located on the ideological spectrum. On the 'right', the rise of the welfare state and the demise of voluntary services and self-help was a detrimental development, and from this viewpoint the future of social policy turns upon the growth of a less state-dominated system in which the voluntary sector has a wider role in actually delivering
PUBLIC POLICY
services under contract (see Gutch, 1992). On the 'left', the growth voluntary organizations as agents or instruments of government p0]1 °f is seen as symptomatic of the crisis of the liberal state in capitals society (Loney et al. (eds), 1991: 206-13). 'Communitarians' W0uld view the use of the voluntary sector as contributing to a greater sense of community and personal responsibility (Willets, 1994).
Whatever the framework used to explain the rise of voluntary organizations as service-providers, the fact remains that in designing the delivery of public services, the contribution of the voluntary sector is a factor which cannot be excluded from an analysis of the development of modem delivery systems. The mix of voluntary to private and public sectors and the 'partnership' forms is diverse. Private-sector finance may support activities of voluntary groups in association with government; government may choose to deliver via the voluntary sector and fund through grants; or there may be a web of funding and support which facilitates an exchange of finance, expertise and commitment. The model which best approximates to an explanation of such a mix is that of interorganizational analysis (4.3.8), with its stress upon the dependency of organizations such as voluntary groups on the resources of funding bodies in the public and private sectors. In the British context, for example, Butler and Wilson (1990) propose that the relationship gives rise to a neo-corporatist pattern of interaction between the state, the National Association of Voluntary Organizations, local government and charities (figure 4.8).
Figure 4.8   Government-voluntary relationships
Central government
Local government \
National Association of Voluntary Organizations
Charities
Funding relationships -1>~
Mediation relationships--
Source:   Adapted from Butler and Wilson (1990)
DELIVERY ANALYSIS 501
In this relationship state agencies and government have considerable influence over the strategic decision-making of the voluntary sector through its funding and mediating powers.
Contracting-out services to voluntary sector organizations has expanded rapidly in the UK, the USA and elsewhere, with the consequence that we are witnessing the emergence of
a new type of relationship between governments and private organizations, which changes the behaviour of both parties, increases their interde-pendency and blurs the traditional distinctions between them. (Self, 1993:124)
Voluntary agencies seek to raise finance from the public by fund-raising campaigns so as to minimize their dependency on governmental and non-governmental funding. The growth of the voluntary sector has raised issues as to the role of 'management' in the organization and strategies of 'non'-business/profit organizations (Butler and Wilson, 1990; Handy, 1988; Drucker, 1990b). In becoming more dependent on private-sector expertise and resources it may be that the voluntary sector will come/has come to resemble the resourcing organizations in their field (Di Maggio and Powell, 1991; 4.3.8, figure 4.9).
Figure 4.9   Voluntary sector funding and activities
Partnerships with public/private sectors a
Membership of policy networks/communities
Public sector resources
(grants, tax breaks etc.)-
Private sector resources
(money, management expertise)
Individual & community support/membership
Voluntary sector organizations/
activities"
Service provision
campaign
Self-help
Intermediary bodies
•Handy (1988)
At the same time, public agencies may become dependent on the voluntary sector to deliver services which they find either difficult to provide or have not the resources to provide or which have traditionally been provided by the voluntary sector. The level of dependence of the private sector on the voluntary sector is comparatively small.
502
PUBLIC POLICY
DELIVERY ANALYSIS
503
However, the participation of the private sector in sponsorship voluntary organizations indicates that the private sector does see th^ such activities offer PR opportunities and advertising potential which can only do the ethical image of the organization a power of g00cj The spillover effects of private-sector management involvement in the voluntary sector may also be seen as having a wider benefit to company or business (see Lloyd, 1993). a
However, as Knight (1993) argues, the danger is that the close relationship which increasingly exists between the state and the private and voluntary sectors will have a damaging effect on the independence of voluntary agencies qua service-providers competing for re-sources. In recent years the voluntary sector has, argues Knight, taken a 'safety first' approach and opted for a cosy relationship with the public and private sectors. The consequence of this cosiness has been a loss of direction, and a preoccupation with financial opportunities and managerialist values with the result that it feels 'a strong sense of ... low morale and spoiled self-worth' (Knight, 1993: xi). The contribution which they have to make in terms of being important critics of government is gravely undermined by the existence of a 'contract culture' in which criticism may result in the loss of funds and contracts (see Gutch, 1992).
Furthermore, if voluntary sector organizations become ever more entwined in the state and the private sector, there is also the question of whether they can serve as the kind of intermediate bodies envisaged by the various advocates of 'communitarianism'. The voluntary sector has long had a vital role in providing 'mediating structures' which can stand between the individual and large state and business institutions (Berger and Neuhaus, 1977). If delivery mixes continue to use the voluntary sector as a service-providing instrument, their role in acting as means of 'citizenship' and 'empowerment' or 'social learning' and their critical function in society may well be put at risk.
❖ Community 'Communitarianism'
This is a theory which has proved attractive across the political spectrum: Clinton, Gore, the Labour Party in Britain; the 'new right', Helmut Kohl and Jacques Delors! As we noted in 1.6, communitarians argue that modern society has seen the destruction of a sense of community and voluntary organizations, and an over-development of the demand for individual rights at the cost of a sense of responsibility and obligation. In policy terms it suggests that policy-makers should seek to build up intermediate social institutions so as to foster social cohesion, civic
aride, a*10" c°unter tne effects of individualism and social fragmentation. Think-tanks such as perrtos and the Social Market Foundation have been widely credited with advancing the cause 0f 'communitarianism' in Britain. See, for example, Willetts (1994) and Etzioni (1968, 1993). pave Willetts, a leading 'think-tanker', makes the 'new right' case by arguing for 'civic con-'    servatism' which looks to strengthen community and voluntary organizations:
Conservatives understand that intrusive and overweening government poses far more of a threat to our traditions and institutions than the free exchange of goods and services ... Looking back we can see that the destruction of old communities and the building of j enormous housing estates ... did more damage. The insidious expansion of regulation ... I has clearly become a threat to local communities, imposing unacceptable burdens on a host I of independent groups, from local charities to childcare providers and lunch clubs ... The | weakening of our civic institutions, as government has encroached on them ... is responsible i for many of our social discontents ... A concern with the strength of Britain's institutions ... is ■t at the heart of the Tory tradition. (77?© Times, 17 June 1994)
The 'community' as an alternative to markets and bureaucracies
The concept of community has a diverse theoretical background. The concern for the loss and renewal of 'community' has been an enduring theme of social and political thought in industrial societies (see Nisbet, 1974). The idea of 'community' as providing an alternative mode of governance may be seen at its most developed form in the work of M. Taylor (1982, 1987), who has formulated a philosophy of community derived in part from anthropological observations. Laver (1983: 147-66) considers the community as the basis of an anarchic society.
A 'community approach' to public policy may be traced back to the colonial administration of Britain, France and Belgium, whereby administrators aimed at making a community more self-reliant and responsible for problems in their locality. In the 1960s the idea was taken up in the USA through the Economic Opportunity Act, 1964, and the Model Cities Program (amongst others), which aimed at increasing ('maximizing', sic) the participation of the community in solving its social and economic problems (see Marris and Rein, 1967). In Britain the 1969 Community Development Act initiated a range of schemes to promote the community and 'neighbourhood' approach (on the development schemes of the late 1960s and 1970s in Britain, see Loney, 1983). By this time, of course, much of the damage to communities - such as East London - had already been done in the name of 'redevelopment' (see Young and Willmott (1957) for a timely, but sadly unheeded sociological warning). Significantly, the newest wave of 'community'-based public policy in the 1990s has come in the aftermath of the massive destruction of communities in the industrial regions and inner cities which had taken place in the previous decade. ♦
The 'community' and public policy
> Another important component in a sectoral delivery mix is the 'com-
j munity' - as opposed to 'the market' or 'the state' or 'hierarchy'. As
j we noted above (see 4.4.2), community-based public policy strategies
504     PUBLIC POLICY
DELIVERY ANALYSIS 505
have been a significant development in new approaches to local p0r making and implementation in the 1980s and 1990s. As a concept "th~ idea of 'commxinity' is notorious for the variety of constructions which are deployed by policy-makers and academics. As Willmott notes-
Those advocating a new initiative, or those attaching or denning a Darti lar point of view, may invoke the community in support of their ca^" without making it clear which community they mean, in what sense th? refer to it or how far they have established what its opinions or intere</ are.
(Willmott, 1989: 5)
Hillery (1968) suggests that there are some 94 definitions. So whenever two or three social scientists or policy-makers are gathered together to discuss 'community', there are probably many more defrni tions than people! In general terms, the idea refers to groups who share a location or physical space or who have common interests traits or characteristics. Thus, in 'community policy' the concept i applied in either a 'territorial' or a 'non-territorial' sense. Community policy may, for example, be directed at a neighbourhood or a part of a town, or it may be directed at a group of people who share a problem or an interest: 'young people', the 'gay community', the 'disabled community', or the 'arts community'. Forms of community policy can be divided into three major approaches (Glen, 1993):
• community development: this approach is concerned with helping the community to help itself. A good example would be that of programmes designed to promote the economic development of a community or to improve housing. The aim is to create a T>ottom-up' process in which people 'in the community' participate in voicing (defining) and meeting their needs and goals;
• community service: this is an approach which is directed at improving the relationship between the outputs of a service-provider and its users or clients. The aim is to make a service more responsive to the community, and increase the involvement of the community in the way in which services are delivered. Example of a 'service' community approach is the idea of Neighbourhood Watch schemes which are designed to improve the relationship between the police and the public, and to involve the community in policing their immediate environment;
• community action: is an approach which is focused on the problems of power and the mobilization of interests. Community action is a form of 'voicing' which is concerned to campaign for the interests of and policies for those who feel excluded from the political agenda/process. An instance of the 'action' approach is the way in which the 'gay community' has sought to engage in campaigns to
get their views about Aids to the attention of the public and into the policy-making process.
As there are numerous definitions of 'community', it is only to be expected that there are many frameworks within which we may understand the notion of 'community policies'. Butcher and Mullard (Butcher et ah, 1993: 217-37) suggest that we can differentiate between three main approaches to 'community policy' by focusing on their different attitudes towards 'citizenship':
• the public citizen framework: which places an emphasis on participation and rational judgement, rights and obligations. In this framework community policy is seen as a way of enhancing democratic participation, extending democracy and devolving power beyond the decision-making processes of bureaucracy, legislatures and political executives. More democracy, at a 'community'/neighbour-hood level is seen as a way of involving people in the decisionmaking processes which impact on their lives. This view of community policy covers liberal', 'pluralist' strategies to improve local democracy and increase participation, as well as more radical policies which relate to the 'entitled citizen' model;
• the entitled citizen framework: here the emphasis is on the distribution of outcomes in a fairer way. The thrust of 'community policy' in this sense is towards greater social and economic justice through the use of strategies to empower the weak and marginal members of society. The entitled framework is committed to a radical view of the 'community' as an agent of social change and reform/revolution. The strategy is to use the community as a means of defending the interests and rights of individuals and groups who are threatened by the power of bureaucracy, capitalism and professionals;
• the dutiful citizen framework: in this approach the prime values are those of order, tradition and the organic nature of society. The aim of community strategies is therefore to strengthen the intermediate organizations and bodies in society and strengthen traditional social institutions as an alternative to state intervention. This is a framework favoured by the conservative right, which sees the use of community as a means of combating the growth of state interference and 'nannying'. The aim of policy is therefore seen as the encouragement of a sense of civic or public duty, community service, mutual aid, self-help, and voluntary work.
Butcher et ah, (1993) propose a synthesis of community policy which incorporates these frameworks (see figure 4.10).
506     PUBLIC POLICY
DELIVERY ANALYSIS 507
Figure 4.10  The community and public policy
Community policy; a mode of public policy-making and implementation that entails:
Pursuit of goals, outcomes and processes that embraces one or more community values (solidarity, participation coherence)
Relating to the targets/beneficiaries of the policy as members of a 'territorial' or 'interest' community
Authentic engagement with groups and organizations active at the community level
• Priority given to the needs and, interests of disadvantaged, oppressed and marginalized segments of society
Source:  Adapted from Butcher (1993)
Examples of the 'community' as an instrument of policy-making which display some potential for realizing this model include:
• community policing (Osborne and Gaebler, 1992:49-52; Weatheritt, 1993);
• community health care (McNaught, 1987);
• community care for the mentally ill and physically handicapped, people with learning difficulties, child care (Walker, 1982; Hoyes and Means, 1993; Bornat et at, 1993; Dallos and Boswell, 1993, Nixon, 1993; Fimister and Hill, 1993);
• community development programmes (Higgins et at, 1983; Loney, 1983; Broady and Hedley, 1989; Roberts, 1979; Robinson and Shaw, 1991; Smith and Jones (eds), 1981; Rasey et at, 1991; McShane, 1993);
• neighbourhood and decentralization of government (Smith, 1985: 166-84; Sharpe (ed.), 1979; Habeebullah and Slater, 1993; Hoggett and Hambleton (eds), 1987; Burns et ah, 1994);
• local economic development policies (Lynn, 1993; MacFarlane and Mabbot, 1993; Young and Mason (eds), 1983; Community Development Foundation, 1992; Bennett and Krebs, 1990);
• the community and race relations (Hill and Issacharoff, 1971; Saggar, 1991).
The community dimension of policy delivery adds to the 'network' and interorganizational character of the 'mix' which may pertain in a
given policy field. It is also a development which further subverts and challenges the power/autonomy of professionals, such as social workers, in their relationship with the service users as more active 'partners' (Cochrane, 1993b).
* Alternative delivery systems: a menu for change
Osborne and Gaebler (1992) suggest that there are (in the US context) over 30 alternatives to service delivery by public employees:
1 creating legal rules and sanctions;
2 regulation or deregulation;
3 monitoring and investigation;
4 licensing;
5 tax policy;
6 grants;
7 subsidies;
8 loans;
9 loan guarantees;
10 contracting; If franchising;
12 public-private partnerships;
13 public-public partnerships;
14 quasi-public or private corporations;
15 public enterprise;
16 procurement;
17 insurance;
18 rewards, awards and bounties;
19 changing public investment policy;
20 technical assistance;
21 information;
22 referral;
23 volunteers;
24 vouchers;
25 impact fees;
26 catalysing non-governmental efforts;
27 convening non-governmental leaders;
28 jawboning;
29 seed money;
30 equity investments;
31 voluntary associations;
32 co-production or self-help;
33 quid pro quos;
34 demand management;
35 sale, exchange, or use of property;
36 restructuring the market.
508     PUBLIC POLICY
HI
mm
DELIVERY ANALYSIS 509
In one of the most comprehensive surveys (OECD, 1993a) of the trends in public-s management the OECD reported that the most popular initiatives amongst member stnt* # 1992) comprised: es W>
• deconcentration of central government (6);
• development of 'agencies' (7);
• reorganization of public enterprises (9);
• limits to the public sector (9);
• privatization (10);
• decentralization (14);
• new role of central management bodies (18);
• market-type mechanisms (18)
What is driving this re-mixing of structures and instruments? The OECD argues that the new forms of public management are, for the most part, the consequence of economic conditions which are bringing about a 'heightened awareness, both in governments and by citizens, of the size and performance of the public sector' and growing pressures to reform the structures of the public sector so as to make it 'leaner and more competitive'. Thus, throughout the OECD countries, the picture is of a panoply of policies designed to make the public sector emulate the private sector and bring the delivery of public services closer to citizens (OECD Survey 1993a-9).
OECD, Managing with Market-type Mechanisms, 1993
The OECD's review of the use of 'market-type mechanisms' (MTMs) is one of the most comprehensive analyses of its kind. The use of MTMs in 17 member countries is reviewed. It focuses on several of the main instruments in widespread use in those countries, including:
• vouchers (US and UK);
• contracting-out (Australia, Canada, France, Finland, Netherlands, Sweden, USA and UK);
• internal markets (Australia, Denmark, Ireland and UK).
The report notes that the practice of MTMs is very broad and varied between member states, and that, as a consequence, generalizations are somewhat dangerous. However, the authors come to three general conclusions as to the experience of MTMs:
• in the design of MTMs, the deliberate inclusion of as many basic features of competitive markets as possible is likely to maximize efficiency gains. Opting for MTMs that closely resemble market arrangements is a less-risky strategy than grafting on only a few aspects of market arrangements or private-sector policies;
• obtaining improvements through MTMs is never automatic, fast or painless. Implementation is not easy;
• fears of the distributive effects between different income and wealth groups seems less than many have anticipated.
In the light of the somewhat limited experience of MTMs, the OECD report believes that they can serve to increase efficiency in the public sector, but that it requires investment in the design, resourcing and implementation of the methods. They conclude that:
public administrators should be encouraged to introduce MTMs in order to improve resource allocation and make savings on public expenditure. But if MTMs are poorly planned, underfunded and, as a consequence, badly implemented, they will almost certainly fail to achieve their objectives. Devoting the necessary financial, human and technical resources to the design of MTMs has to be considered a sound investment, but it needs to be complemented by campaigns to inform the public and stakeholders about their nature and operation. (OECD, 1993b: 103) ♦
4.4.4  The enforcement mix
Policy is all very well, but without an enforcement or compliance capability, the delivery of public policy is unlikely and uncertain. We only have to consider, for example, the difficulties which European Union policies (and other international agreements and laws) have encountered to realize that a good policy, if it is to be carried out, must have effective means of enforcement. The mix of enforcement methods may range from brute force and fixed bayonets to information broadcasts which seek to change behaviour. Markets, bureaucracy and community, for example, may be viewed in terms of different ways of enforcing. The market does it through supply and demand, prices, and the interaction of buyers and sellers; bureaucracy relies on rules; whilst community enforcement relies on such modes as shared values, reciprocity, trust, and gossip (see Colebatch and Larmour, 1993: 23). Here we shall examine four approaches to the mix of enforcement modes. One approach focuses our attention on enforcement and power (Boulding); another examines forms of administrative enforcement and their effectiveness (Hood); and a third provides a framework for analysing the relationship of enforcement to policy type and regime (Burch and Wood). Finally, we examine Etzioni's schema for analysing kinds of power as providing a link between the enforcement of public policy, and the problem of compliance within the organizations responsible for implementing it.
Boulding
We have already discussed the work of Kenneth Boulding (3.5.2). As an economist, his ideas about power have great relevance to understanding the enforcement 'costs' of delivery and the 'boundary possibilities' of different mixes of enforcement strategies. Boulding distinguishes between three kinds or categories of power: threat; exchange; and love (figure 4.11). He then considers these against their consequences: destruction; production; and integration. The consequence of destructive power is to destroy people and valued things. The consequences of productive power is to be seen in the 'fertilized egg', blueprints, ideas, tools and machines, and the activities of construc-
1
510      PUBLIC POLICY
Figure 4.11   Boulding's categories of power
Threat
Exchange
Love
Destructive
Productive
Integrative
D (Destructive) P (Productive) I (Integrative)
Source:  Adapted from Boulding (1990)
tion and manufacture. Integrative power 'may be thought of as an aspect of productive power that involves the capacity to build organizations, to create families and groups, to inspire loyalty, to bind people together, to develop legitimacy' (Boulding, 1990: 25).
Significantly, Boulding does not see his categories of power as 'pure' forms: as in the case of 'love', threat and exchange have the quality of being 'fuzzy sets', in that they contain elements of other types of power consequences. Enforcement through threats can have productive and integrative consequences, as in the case of the exercise of threats by income tax authorities or policemen. The threat carries destructive consequences, but if successful may be socially productive and 'integrative'. The use of exchange power to enforce is productive in that for x, a gets y from b, and b gets z from a (reciprocal) or it may be contractual: a gets b if he does c. It may involve bargaining to establish a satisfactory rate of exchange or terms of trade. However, it also contains elements of destruction - to enforce contracts - and trust and courtesy (integrative traits), to make exchange possible. The power of love is such that it calls for compliance on the basis of 'respect' and legitimacy', out of social and persona) responsibility. The use of 'love' as a mode of enforcement involves an appeal to a moral sense or a sense of community or citizenship. It is the request for compliance not based on a threat, or on 'if you do x, you will get y (exchange)' but on
DELIVERY ANALYSIS
511
the appeal to a social sense or loyalty to the nation. However, as with the other categories of power, love also contains a capacity to generate its destructive opposite: to create enemies, to alienate people; it has a destructive as well as a productive aspect. And also as a power, the integrative consequences of 'love' as a mode of enforcement may prove highly productive: compliance takes place because of concern, care or a sense of duty, rather than for money or avoiding the long arm of the law.
As a model for a delivery strategy, Boulding's classification points towards the mix of enforcements which societies seek to use to gain compliance (integrative enforcement), or to force compliance (threat enforcement), or to produce compliance (exchange enforcement). Some policies may contain more threat than exchange, others more love than threat.
Hood
Hood (1986: 48-60) identifies four options or modes of enforcement:
• set aside/modify rules: in cases of non-compliance, government may choose either to set the rules aside or to modify the rules in order to bring about compliance;
• spread the word: government may choose to use publicity and persuasion. This amounts to a kind of 'please keep off the grass' notice: an attempt to gain compliance by giving information or advice which aims at modifying behaviour;
• pursue and punish rule-violators: the use of legal and police action to deter non-compliance;
• make it physically difficult, impossible, inconvenient to break rules: in this case, the enforcement method is not to use notices to inform people about the grass, but involves making the grass physically difficult or inconvenient to get at in the first place. Hood examines the problems of different modes of enforcement in different conditions. We have simplified this in figure 4.12.
In the case of incompetent (1A) non-compliance, the application of mode 1 enforcement has no effect on non-compliance. Mode 2 may change behaviour, whilst mode 3 is unlikely to diirmnish an incompetent form of non-compliance. However, modes 3 and 4 (3B and 4B) may depend on whether the prevention of non-compliance is feasible and affordable. In the case of an opportunist non-compliance, mode 1 is likely to increase the oppportunities for non-compliance; mode 2 is unlikely to have much effect on an opportunist; modes 2 and 3, however, are more likely to have an effect on the opportunist, through fear of getting caught (3B) or inability to do it (4B). In the realm of princi-
512     PUBLIC POLICY
DELIVERY ANALYSIS 513
Figure 4.12  Hood's model of enforcement
Enforcement response
Type of
non-compliance
uwrence Waliack, ' inflice and Atkin, 1£
Mass communication and health promotion: a critical perspective', 988
	Soft		Hard	
	(1)	(2)	(3)	(4)
Incompetent	A	A	A	A
Opportunist	B	B	B	B
(a)  Specific rule/law	C	C	C	C
(b)  Authority in general	D	D	D	D
- Unprincipled
- Principled
Enforcement response options:
1 Set aside / modify rules.
2 Spread the word.
3 Pursue and punish rule violators.
4 Make it physically difficult, impossible, inconvenient to break rules.
Source:  Adapted from Hood (1986)
pled objection to a rule or a specific application, mode 1 may influence dissent, without doing too much damage to basic rules. Mode 2, in conditions of an objection in principle, is unlikely to change behaviour (2C); 3C may lead to 'martyrdom', whilst 4C may make matters worse. In the case of a principled objection based on a rejection of authority in general, ID amounts to admitting defeat, 2D is unlikely to change behaviour, 3D has a potential for martyrdom, and 4D is again dependent on the feasibility and affordability of the preventative measures.
❖ Evaluating the effectiveness of campaigns
With the growth of mass communications and the use of information campaigns in many areas of social and environmental policies, an increasing emphasis is placed on individual responsibility for problems. Campaigns on road safety, for example, urge that we take responsibilty for crossing roads which are becoming more dangerous because there are more cars. Why there are more cars is not something which such campaigns tend to address. If, for example, we take the view that road deaths are essentially due to public policies which favour cars against public transport, then the road-safety issue is less a problem of personal fault or carelessness than an inevitable consequence of public policy.
WaHack notes that the way in which society sees problems has a great impact on how it reSp0nds in terms of public policy. The media presentation of health issues and government campaigns have, he argues, a distorting effect on this process. Chauncey Gardner, in the film Beir>9 There, is all too typical of the growing trend for citizens in Western societies to receive most of their information about the world via the TV screen. Public health is a complex problem Which is bound up with a wider social and political context, and yet this context is ignored by the presentation of health issues on TV:
One of the ways to ignore the debate and take what appears to be meaningful action is to talk about life-style factors as being significant influences on health status. This is a basic marketplace concept that suggests that health problems are 'purchased' as a by-product of goods consumed, (p. 360)
He points out that the evidence strongly suggests that
. life-style changes do not have long-term effects on health status;
» large segments of the population have little chance to participate in changes;
• health status is determined far more by external factors (such as economic status) than by life-style;
• the TV images of health and medicine may be bad for your health.
The manufacture of illness as a personal/knowledge problem rather than as a complex public problem may therefore be actually making for a less-healthy population as a whole because it results in the fundamentals of health status being ignored!
In many areas of government policy the aim is to change attitudes and behaviour. This is particularly the case in areas such as health and crime, where government seeks not so much to provide a 'service' as to alter life-styles through information and persuasion. In a sense it is a form of public policy which lays the burden of change on individual behaviour rather than on public choices. Life-style campaigns stress the responsibility of the individual for problems which we may well argue are essentially social. Aids, smoking, drugs and crime prevention are some of the most common forms of campaign as instruments of public policy.
Evaluating the impact of such campaigns is fraught with difficulties, and there are a variety of models on offer. For reviews of this material, see Rice and Atkin (eds) (1988).
It may be argued that such campaigns must be viewed as forms of social learning. But how that can be measured is even more problematical. ♦
Two crucial aspects for Hood are: how much enforcement should take place? and by whom? In the enforcement of a policy, government has to make a decision about what levels of enforcement they consider to be acceptable - 100 per cent, or 20 per cent, say - and what standard
PUBLIC POLICY
DELIVERY ANALYSIS 515
or quality should be enforced. The delivery of policy will va terms of what levels and standards will be enforced. It may not h cost-effective to put extra resources into enforcement for a mar ■ & improvement deemed not worth the effort. Delivery agencies decide that enforcement will take place to attain a given level***' compliance. Or they may choose to hunt down and prosecute ev °* case of non-compliance. Standards of compliance may be low in sorri areas and very high in others. The levels and standards of enfor^ ment may be related to who is doing the enforcement, and on what basis. For instance, we may decide that the enforcement of policy ^ one area is a matter for public bureaucrats; in another, 'professionals' may be responsible for ensuring compliance; it may be left to a pri vate-sector agency; or enforcement may be policed internally (self-policing); or may be implemented via the community.
The issues of how much compliance and by whom adds a further layer of complexity in the mix of delivery systems. The low levels ot enforcement (say of measures to counter racial discrimination or sexual harassment or social security fraud) or the low standards of enforcement (in health and safety, for instance), and the fact that the enforcement is in the hands of professionals (who use their expertise to make judgements), or that enforcement is in the hands of insiders rather than outsiders (in the police force, or in the medical and legal professions) may generate new issues and problems. Who does the enforcement is a contentious matter. Here again we must examine the other dimensions of a delivery mix. If government chooses to move a public-sector service out of the hands of professionals into the hands of private-sector businesses - such as prisons or policing - then we as citizens may complain that the standards of enforcement in the delivery of these services is driven by 'profits' rather than 'the public interest'. Yet again, as we have seen elsewhere in this book, the process of deprofessionalization and a distrust of professionals reflects a broad-based social concern with the power of professional groups.
Burch and Wood
Burch and Wood (1990) provide a framework on which to map the choices of enforcement methods in terms of 'negative sanctions' which prevent people doing things. This corresponds with the 'hard' sanctions in Hood and the use of 'threat' power in Boulding. At the other extreme is the use of 'positive' sanctions, which are more indirect: they are aimed at inducing a change in behaviour. Positive sanctions include the 'soft' sanctions of Hood, and the use of 'love' and 'exchange' in Boulding. Against these Burch and Wood set modes of controls in terms of 'formal' and 'informal' controls. The modes of control which are employed in policies are set in the context of forms
of political regime. The domain of positive sanctions and informal controls permit of a political regime which is pluralistic and in which enforcement takes place through bargaining. The domain of formal controls and positive sanctions may be characterized as being highly legalistic, but manipulated to induce compliance. (This is the domain of Hood's mode 1: set aside and manipulation.) The domain of formal controls and negative sanctions is an area of authoritarian legalism, in which rules are rules, and enforcement takes place in the use of law and policing and threat; whilst the domain of negative sanctions and informal controls takes us into the areas of tyranny and autocracy. Discretion is high in the domain of positive sanctions and informal controls, and low in the domain of negative sanctions and formal controls (figure 4.13).
Figure 4.13   Enforcement and regime type
Low discretion
Negative direct sanctions (restrictions and modifying behaviour)
(Authoritative^ legalistic J
c
Formal controls (clear rules, legally binding)
Centralized autocratic tyranny
5
^Petty tyranny^
^Manipulative legalistic^
Positive indirect sanctions
(behaviour-inducing)
^Bargaining, pturalistky
^ Regime type
Informal controls (standards, conventions)
High discretion
Source:  Adapted from Burch and Wood (1990)
As the authors admit, such attempts to classify the complexity of the interaction of control, sanctions, discretion and types of regime is invariably 'crude'. However, this said, Burch and Wood do provide a way of understanding the mix of factors which shape the delivery process. As they observe:
In the top left quarter, we see the combination of formal and negative systems. This can produce policies that fail to take proper account of local needs and changing circumstances and that are highly unpopular and so carried out by the use of force or directives. Citizen allegiance to the regime and its policy products may be undermined. By contrast, the bot-
516
PUBLIC POLICY
DELIVERY ANALYSIS
517
torn right quarter illustrates the mixture of informal and positive p0tye Those who carry out policies have great powers, but at the cost of their rw being fully accountable for their actions, which will not necessarily mJJj the objectives of the original policy makers. And the treatment of j^^ ; vidual cases will vary from official to official, from place to place, ar^ from day to day. Possibly the worst of all is the bottom left quarter comb' nation of negative sanctions and informal controls. In such 'petty tyranJ nies', harsh and unpredictable behaviour by uncontrollable officials r&J suits in neither allegiance to the regime nor the effective carrying out of? policy. i (Butch and Wood, 1990: 184) \
The Burch and Wood classification also brings out the fact that the mixes of enforcement vary across policy areas, and types of political systems.
-J
the perspective of an interorganizational framework it might be argued f*0^ [jcy area (network) involves notions of 'co-operation' and 'negotiation'
if! af<;uch as legal authority, 'command' and 'obedience'. ♦ ideas ^
that enforcement gotiation' as well as formal
❖ Enforcing environmental policy
The problems of enforcement come into prominence when we consider the issues relating to the implementation of environmental policy. This is a field which is not short of good intentions and legal regulations. However, making them stick is quite another matter. Voge! notes, for example, that in the case of the EU's environment policy:
as the scope of Community regulations has grown, the problem of enforcing them has become more acute. The increased number of Community regulations not only makes the monitoring of their enforcement more difficult, but also more urgent, since significant variations in national compliance threatens to disadvantage industries in some EC member states. The Community itself has no police or enforcement powers. It only knows of a violation if someone complains. If it finds the complaint is justified, its final legal recourse is to sue the member state in the European Court. This, however, is a time-consuming procedure ... The problem of enforcement is further complicated by the fact that while EC Directives are intended to bind national governments, in many cases it is local governmental officials who are responsible for enforcing them. (Vogel, 1993: 124-5)
The case of the EU's enforcement of environmental policy in the context of Burch and Wood's highly pluralistic and discretionary quadrant (figure 4.13) illustrates that enforcement has to be viewed in terms of 'negotiation' rather than top-down command. From the perspective of nation states, the experience of 'top-down' enforcement in environmental policy has not proved a spectacular success (Downing and Hanf (eds), 1983). Hanf (1993) suggests that we ought to revise conceptions of enforcement to take more account of the way in which the practice of enforcment involves forms of 'co-production':
In real terms it is not legislation that determines what deviant behaviour is, but rather the inspectors and investigative officials who do. It is these officials who create cases of violation of environmental law. Consequently, the process of regulation is not simply one where the regulators command and regulated obey. (Hanf, 1993:91)
Etzioni: mixing love, fear and money
Etzioni (1961) takes the view that there are three basic reasons why people in organizations comply with rules, disciplines, orders or policies: to begin with, they may do so out of a sense of agreement, love, or morality. They do not need to be forced into doing something; they do it because they want to do it. Secondly, people may comply because of fear. If they do not do what they are asked/told they fear the consequences of non-compliance. Finally, they may comply because it is in their monetary or remunerative interests. Although they may disagree on moral grounds, or hate what they have to do, they do so because of the monetary reward which compliance will bring. Enforcement may therefore be the result of normative, coercive or remunerative power. As we have seen from the other theorists above, in seeking to make sure a policy is carried out, implementation will require a mix of different enforcement modes. A policy may rely on coercion or monetary rewards/sanctions, or it may ask people to exercise a moral choice (sticks, carrots or kisses). When we consider that, in order for a policy to be carried out, the implementing agencies must also comply, we realize that enforcement has an organizational as well as a 'public' context. In making policy, decision-makers have to choose a mix of instruments which it is hoped, will ensure that a bureaucrat or professional actually does what is required. As we shall see in 4.5.3, this is the problem which has been addressed in public-sector reforms which draw on the techniques of human resource management (HRM).
Etzioni argues diat there will tend to be a balance between different kinds of power and kinds of involvements:
• coercive—alienative;
• remunerative-calculative;
• normative-moral.
Effective organizations, argues Etzioni, are those which attain a balanced mix between low levels of 'fear' (coercion and alienation) and high levels of 'money' (remuneration and calculation) and 'love' (normative and moral) involvements. Significantly, from the perspective of HRM and recent management reforms in the public sector, Etzioni posits that organizations with 'similar compliance structures tend to have similar goals, and organizations that have similar goals tend to have similar compliance structures' (Etzioni, 1961: 71). As goals be-
518     PUBLIC POLICY
DELIVERY ANALYSIS 519
come more congruent, so do the compliance structures of organj^ tions (see Clegg, 1990: 42-4, for a review of Etzioni).
We could argue on this basis that, as schools, hospitals, universities and other 'public' organizations come to share similar goals to that of the 'private sector' - efficiency and profit for example - so the comply ance structures will shift to resemble the structures found in business and industry. As Clegg notes, the evidence to support Etzioni's modef; is 'considerable' and has withstood much empirical scmtiny (Clegg,;: 1990: 43). The model would predict, therefore, that as institutions and/ organizations which deliver public goods and services become more^ 'businesslike', their mix of compliance structures will increasingly^ resemble those in the 'profit' sector. The public and private sectors may well come to look and behave in such a similar way that the distinction may become quite meaningless (see, for example, Malkin and Wildavksy, 1991; see also the arguments regarding institutional isomorphism: 4.6.3). 1
Enforcement and organizational context
The enforcement or compliance mixes which are used have, of course,, an organizational setting. The choice of market modes of organization, for example, will perceive the problem of gaining compliance as : one which is rooted in self-interested behaviour (figure 4.14). Policy in a market model may be viewed as being enforced through market-;? type arrangements, the benefits of exchange, the incentive of remuneration or the use of contracts to ensure that 'agents' do what they are contracted to do by principals (see 3.6.3, on agency theory). ;
Hierarchy or bureaucratic modes involve the notion that enforcement will require effective methods of command and the use of coercion or
Figure 4.14  Modes of enforcement and modes of organization
Modes of organization	Modes of enforcement/compliance			
	Rigby (1964/1990)	Etzioni (1961)	Bouiding (1990)	Bradách & Eccles (1991)
Market	Contract	Remunerative	Exchange	Price
Hierarchy/ bureaucracy	Command	Coercive	Threat	Authority
Network/ community	Custom	Moral	Love	Trust
threat to ensure compliance with authoritative rules. The selection of 'network' or ' community' organizational forms will rely on the operation of custom, tradition, common moral codes, values and beliefs, love, a sense of belonging to a 'clan' (see Ouchi, 1991), reciprocity, solidarity and trust.
The selection of enforcement rnix is therefore a matter of values, and it is this issue of the mix of values that shape and inform choices which we examine next.
4.4.5   The value mix
The value dimension Key texts
On the problems of equity, see Le Grand (1991); Le Grand and Bartlett (eds) (1993); Rawls (1971). For the utilitarian approach, see Sen and Williams (eds) (1982); Smart and Williams (1987); Allison (ed.) (1990). The case for a minimal state is addressed in Nozick (1974); Buchanan (1977).
In general, see Frohock (1974, 1979); Beauchamp (1975); Lane (1993: 205-19); Bobrow and Dryzek(1987: 101-16); Weale (1983); Miller (1990).
On the role of values in general, see the work of Vickers (3.7.5) and Young (3.7.7). ♦
The final mix to be considered in our analysis of delivery systems should really go first: what is the distribution of values which frame and inform the delivery mix of a given policy or programme? The governmental, sectoral and enforcement mixes are ultimately manifestations of the values - or the 'assumptive world' - (see Young, 1977; 3.7.7) of policy-makers. In Vickers's terms, 'action' judgements are the outcome of the interaction of reality judgements and value judgements. The value mix will involve choices and priorities regarding the allocation of resources between policy and problem areas as well as between different programmes directed at common problems and policies. These choices in the distribution of resources between policy/ problem areas will have a major impact on the other choices of mix between governmental, sectoral, instrumental arid enforcement. Resource distribution reflects values and beliefs, power and interests, and, as we have seen in 4.3.7, will shape the way in which organizations within policy fields relate to one another. As Coiebatch and
520      PUBLIC POLICY
Figure 4.15  Institutions, values and meanings
institutional settings
Framework of meaning
Underlying values Source:  Adapted from Colebatch and Larmour (1993:108)
Larmour point out with regard to organizational mixes, we have to understand the choices between market, bureaucracy and comrnunity as an interactive process involving institutions, values and meanings (figure 4.15).
The framework of meaning they define as 'the way people make sense of the organizational activity; what action is appropriate and why' and how 'underlying values' form a pattern which 'informs the frameworks of meaning people employ and the institutional settings within which they operate' (Colebatch and Larmour, 1993:108-9). These different dimensions interact and frame meaning for participants and what institutional or organizational mix they think is appropriate for dealing with a problem or delivering a policy response and valuing the outcome:
For instance, a state's youth unemployment is recognized as a major policy problem, and it is argued by some that the best way to combat it is to remove the statutory restraints on the employment of young people ... this claim is hotly contested, and the conflict reflects the clash between the market and hierarchical models. The market model contains a framework of meaning which sees life as a set of self-interested exchanges between rational actors, an institutional setting where these exchanges can take place ... (market place and law and police) ... and a way of valuing outcomes ... (agreement between employer and young person) ... The hierarchical models rest on a different set of values, seeing the young person as not having bargaining power ... to preserve his/her best interests, and therefore needing protection of a set of institutions ... In other words each model offers a distinct way of organizing the employment of young people.
[Colebatch and Larmour, 1993:110)
Colebatch and Larmour's model may be combined with two models we have encountered in Part Three: Vickers's judgement approach and Young's idea of the assumptive world. Colebatch and Larmour's dimensions of organizing are illustrations of Young's assumptive world (see 3.7.7). They may also be read as the products of an interaction
DELIVERY ANALYSIS
521
Figure 4.16   Values, reality and action
Institutional settings
Framework of meaning
Underlying values
(Assumptive world) Young
"Colebatch and Larmour
Value judgements	
	
Reality judgements	
Action judgement
Delivery mix
Market
Bureaucracy
Community)
Vickers
Source:  Adapted from Colebatch and Larmour (1993), Vickers (1965) and Young (1977)
between Vickers's 'value', 'reality' and 'action' judgements. We combine these ideas in figure 4.16.
Philosophy and public affairs
The approach which has been most commonly used to explore the value or normative aspects of public policy may be broadly stated in terms of one its leading journals: Philosophy and Public Affairs. The philosophical approach to delivery systems would, in the context of the mix between markets, hierarchies and networks, ask questions about the assumptions, values and beliefs which underpin policy choices:
• utilitarian analysis would be concerned with whether a given mix in a policy area is 'efficient';
• analysis which is concerned with issues of social justice would ask questions about whether the implementing of a public policy is done in a fair or equitable way and/or whether the outcome of a given mix is just;
• an analysis predicated on the ideas of a 'minimal state' might be more interested in the issue of whether individual rights have been extended or reduced.
For example, a policy may be geared to the attainment of greater 'efficiency' rather than greater 'equality' or 'equity', in which case, it
PUBLIC POLICY
DELIVERY AM A LYSIS
523
may be argued that markets (real or quasi) best deliver efficiency-value for money. Even though a policy may include a concern a0o^ issues such as equality, the mix of delivery techniques may be biase<j ^ more towards securing efficiency than equality. A programme m ~ ^ well endeavour to distribute values by setting as the primary valu^ greater efficiency, but it may also take account of the demands fojJ equality and equity. How policies are delivered - or their organic í tional mix - will consequently reflect the priorities and values of ^ policy-makers. An instance of this might be the way in which a value! setting in health policy may place 'efficiency' at a higher priority than equality of access or equity of treatment (Bartlett and Le Grand, 1993-1 18). A delivery system in this sense is composed of a bundle of value i and the choice of a mix is the result of the way in which policy-makers have designed or selected a mix to realize one set of values over another. Using LasswelTs idea of 'enlightenment' as a value in pubV! policy, we might argue that 'community'-based delivery may well be predicated on a belief in developing or using institutions in which people can learn about problems. On the other hand, although the usel of quasi-market mechanisms may include values of enlightenment, i those of learning, of participation, and so on, will be ranked fairly low in evaluating the 'effectiveness' or 'efficiency' of the delivery system. %
At the same time, the mix between efficiency and equity may vary within and between policies and programmes. Equality, for example, may be defined in terms of equality of opportunity and access, rather than equality of outcome or consumption. Freedom in the context of 1 policy delivery may be framed in positive terms: that is, delivery should i be motivated to make users free from conditions such as 'want, ignorance and disease'. On the other hand, the delivery of a policy may be ^ more constrained by the negative sense of freedom as the absence of > constraint. The former (positive) clearly involves the idea that the deliv-1 ery process should be concerned with outcomes and provision; whereas = the latter (negative) suggests that the delivery process should be designed to provide a procedure which is just and which 'enables' rather than provides. The theories of Rawls and Nozick which address the questions of fairness in outcome and procedural justice advance rival philosophical frameworks within which the choices of delivery systems may be understood in a normative context. For example, in the case of . the decentralized systems chosen by Islington and Tower Hamlets (see 4.4.2), the actual mix of organizational arrangements and policy instruments has been framed by the different values which have informed their respective policies. As Burns et ai, note:
it is the values of the political parties and not the organizational structures ... that have determined the outcomes which the decentralization pro-
duces ... The Labour Party in Islington has a universalistic conception of social justice (Rawls, 1971). Hence their stress upon equal opportunities and anti-poverty ... The Liberals in Tower Hamlets, on (he other hand, have a conception of rights founded on historical relationships to territory and property - which would not be uncomfortable wifh the radical libertarian philosophy of Robert Nozick (1974). Thus their stress on 'sons and daughters schemes' and endorsement of council house sales alongside their enthusiasm for decentralization. (Bums et al., 1994:219)
tf Whose costs and benefits?
What is optimal in the distribution of resources between one policy/programme/option and another? The economist Pareto (1848-1923) provided an important model to explain and explore welfare problems. His argument was this. An optimal distribution of welfare is when everyone is better off: costs = benefits / benefits exceed costs. Cost and benefit analysis was framed in terms of this notion of being able to make calculations as to the optimal distribution of costs and benefits. Does the wonderful world of public policy inhabit this domain of Pareto optimality? The answer is that it depends on a number of factors. If the size of the cake is bigger over time, then it is possible for everyone to have an increase in resources for their policy/programme. And, if decision makers believe in equity across the board (or horizontal equity) then a bigger cake means everyone is better off. However, if policy-makers do not believe in equality of welfare then the allocation of the cake will be vertical, and unequal. Bigger cakes do not mean bigger slices for all: costs and benefits are not distributed equally. It depends on the nature of the distributional values which inform the allocation. Of course, the really bloody position is where the cake is getting smaller. If the decision-makers have a horizontal welfare position then the model will be equality of misery: everyone is worse off. However, if they are of the view that equality should not be applied, there will be a vertical allocation in which case, there will be winners and losers (although according to the so-called Kaldor-Hicks modification of Paretian optimality, winners may compensate losers to satisfy an optimal distribution). The problem of the distribution of welfare in CBA raises the issue of the relationship between what is efficient and what is equitable. It may be that the budget or a series of policy decisions represents an 'efficient' use of resources, but they may not result in an equitable allocation. This problem has been addressed in John Rawls's A Theory of Justice (1971). Rawls argues that the values we should apply to resource allocation are not those of costs and benefits, but that of 'fairness'. His argument uses the metaphor of a veil of ignorance under which we are placed to consider what kind of world we would choose to live in. Under this veil we know all about the world, but nothing about who we are. What, asks Rawls, would we vote for? The answer, he says, is that we would opt for a situation in which gains for the best off could only take place if the least well off were in a better position after the winners had gained. Obviously this notion has many implications for the way in which we see the aims and purpose of public policy as a whole. If we take the view that such an arrangement would make for a fair society or a fair decision, then the policy process would be structured by the principles that winners can be allowed to win only it they benefit the least well off. In figure 4.17, the 45 degree line (e) provides for equality. Every increase in A's income is met with the same increase in B's. The distribution here is positive: y, represents a position in which A's increase (1) is matched by B's. However, in the case of the negative (i) line, an increase in A's income actually makes B worse off (y2 to y3). In a Rawlsian sense, we may see public policy as to do with ensuring line f,
524      PUBLIC POLICY
Figure 4.17  Fairness and public policy
B's income
or fairness provides the context of decision-making. Here we allow A to increase his income but as it rises, so does that of B (y4 to y5). The more we move the nine to the right, the greater the advantage to B. It may therefore be necessary to keep a programme or a policy going, even though it is not efficient or cost-effective so to do in order to be 'fair'. A Rawlsian view of the provision of public goods will mean that changes to policies, or the development of new policies will have to be set against the need to maintain existing allocations for reasons of equity rather than for reasons of efficiency. ♦
The desire for 'positive freedom', 'equality of outcome' and 'fairness' has long been regarded as strongly biased towards bureaucratic or hierarchical systems, as opposed to negative freedom, procedural justice and 'efficiency', which have been viewed as conditions best delivered by markets. However, this polarization of markets as more suited to 'efficiency', and hierarchy as more fitted to provide equality and equity, is grossly simplistic. In practice, market forms of public management often aim to arrive at a mix in the delivery process which utilizes a range of instruments to improve 'efficiency', whilst seeking to balance these values against the demands for 'equity' and 'equality' (and enlightenment) (Osborne and Gaebler, 1992: 348).
From a Lasswellian perspective, what is at issue here is how a delivery system may be mapped in terms of the values which institutional and organizational arrangements seek to promote. In this context there will be inevitable conflicts in values between the different agen-
DELIVERY ANALYSIS 525
cies involved in the organizational matrix (see Hjern and Porter, 1981), and between the values of high-level policy-makers (at the Ministry of Health) and the values of 'street-level' implementers (health professionals) who may well disagree with the delivery mix within which they have to do their job. In analysing the way in which values are distributed in the delivery of public policy we are therefore concerned with power: whose values predominate, when, and how (Lasswell, 1936)?
As we argued in 4.3.8, one of the problems with a 'general' or synthetic theory of implementation is that it serves to make very distinctive sets of values which operate in the delivery of public services less clear. Bottom-up approaches are, in the main, predicated on values which stress openness, decentralization, participation and so on. Thus if the value which predominates in a given delivery mix is 'control', the institutional arrangements and the modes of enforcement will reflect this value. If, on the other hand, the delivery of a policy is viewed in terms of 'social learning' (4.6.4) or as a means of 'community development', the values which shape the choice of implementation are likely to be far more predisposed to a bottom-up framework.
4.4.6   Evaluating delivery systems: exit, voice and loyalty
The ultimate test for a delivery system is whether citizens are satisfied with the goods and services which it provides. They may feel unhappy with the value outcomes and be concerned about the falling standards of service. When this happens, what options do citizens have? This issue of response to deterioration was the subject of an influential study by the economist Albert Hirschman (1970b). Hirschman's arguments as to the response of firms, organizations and states to decline provides a framework and the language uses both economic and political concepts to explore the relationship between policy-makers and implementers and those who are actually on the receiving end. As with other economic analyses, Hirschman employs the notion of rational actors; however, his model also has a cognitive dimension which is suggestive of other theories of group behaviour which we reviewed in Part Three which address the 'non-rational' dimensions of decision-making (3.7.1, 3.7.2). For this reason, the Hirschman model provides a useful integrative framework for the analysis of organizational behaviour and the interaction of members within and outside groups.
As a theory it also has a special relevance in those situations in which policy-makers have sought to make public goods and services more
526      PUBLIC POLICY
DELIVERY ANALYSIS 527
ini
like those in a market place. The model illustrates the problems and issues which may arise in market-driven public policy. Hirscrm^ argues that there are two options which may be considered availabl to the consumer in the face of a decline in the quality of goods and services: exit and voice.
Exit
The capacity to exit is the essential ability of the consumer in a market place. If the consumer is no longer satisfied with the goods or service he or she can choose to buy from another firm or use another organization. For the exit mechanism to work, firms should have a rrux of inert and alert customers; alert customers provide feedback on the quality of goods and services, whilst the inert provide the 'cushion' needed for recuperation to take place.
Voice
Voice corresponds to the articulation of interest and protest in a political sense. Hirschman notes: 'To resort to voice, rather than exit is for the consumer or member to make an attempt at changing the practices, policies, and outputs of the firms from which one buys or the organization to which one belongs' (Hirschman, 1970b: 30). Just as alert customers who opt to exit serve to improve quality (although not always: Hirschman, 1970b: 27), so active, articulate customers may serve to improve the quality of output: 'Voice has the function of alerting the firm or organization to its failings, but it must give management, old or new, some time to respond to the pressures that have been brought to bear on it' (Hirschman, 1970b: 33). The readiness to resort to voice will depend on the extent to which the certainty of exit advantages can be traded off against all the uncertainties of changing things through voice. It will also depend on the strength of voice in terms of number and organization. The greater the capacity to exercise an exit, the louder and more effective will be the voice option.
Voice may serve to complement the capacity to exit, or it may be an alternative - and the only alternative for some - to exiting. Hirschman argues that in general, however, the capacity to exit drives out voice. If alert customers have alternatives, they are likely to take the option to exit rather than to express dissatisfaction. However, the efficacy of exit as a response to decline may not have a beneficial effect in counteracting decline: the firm may simply be content with its inert customers and their reluctance or inability to exit. The exit option may not be exercised for several reasons, including the customer's belief that voice may have the effect of changing the situation, or the customer/member/citizen may have a sense of loyalty to the organiza-
tion and its output. The exercise of loyalty constrains the exit option, and strengthens the voice option.
In the case of public goods and services, the exit from a public good -such as transport, education or health services - does not mean that the citizen gives up voice: he or she still has a vote and may express dissatisfaction about the service from which he or she has been constrained to exit. At the same time, public goods in a pure sense cannot be exited from in the sense in which consumers of private goods can 'vote with their feet' (although public-choice theorists would argue that by marketizing public services consumers will be free to choose: see Tiebout, 1956). Citizens cannot move out of society; although, of course, they may ultimately leave and seek membership of another country - assuming that they are fit and healthy and sufficiently well-resourced to do so, and that immigration laws let them move about (see Soysal, 1993).
Loyalty
Loyalty is clearly a primary factor in the choice between exiting and voice. Loyalty restricts and retards exit and increases the propensity to choose voice as the mode of response to decline. Even though a citizen may have the capacity to exit from a public programme, he or she may decide that to leave it would be disloyal and 'wrong'. They may reason that if 'everybody' exited from the service, society as a whole, which includes them, would be worse off, and thus they choose voice, rather than exit. Another factor is the cost of exiting set against the costs of entry. The higher the costs of entry the lower the disposition to exit; and the greater the loyalty, the lower the desire to exit. This aspect of the model was subjected to a number of tests by psychologists who examined the relationship between severe initiation procedures for admittance into groups and the degree of self-deception. The more severe the initiation, the greater the group loyalty and exercise of voice, and the lower the rate of exiting (Hirschman, 1970b: 146-55).
This finding is parallel to the idea of cognitive dissonance observed by Festinger, whereby facts which disprove belief do not serve to disprove, but to reinforce, beliefs. Even when beliefs are disconfirmed, believers become even more vigorous in proselytizing (Festinger, 1957). Loyalty, therefore, may well predominate in conditions where there are high initial costs and high exit costs: in which case citizens may choose to exercise loyalty and voice, rather than exit completely. Or they may choose to exit for a temporary period in order to bring pressure to bear - by the use of a boycott as an instrument of voice. Loyalty, Hirschman argues, 'holds exit at bay'.
528     PUBLIC POLICY
DELIVERY ANALYSIS 529
McPublic sector
A plausible story
George Ritzer, The McDonaldization of Society, 1993
Ritzer is a sociologist interested in the process of Weberian 'rationalization'. He argues that th McDonalds chain of fast-food outlets constitutes a model of the contemporary process \ rationalization. He describes the spread of this as 'McDonaldization'. Max Weber believed that the process of bureaucratic rationalization would result in an 'iron cage' from which industrial societies cannot escape. McDonaldization, however, represents less the triumph of an 'iron cage' so much as a 'velvet' or 'rubber' cage. In this world, people, especially those born since the advent of McDonalds, 'like, even crave' McDonaldization:
This is the world they know, it represents their standard of good taste and high quality, and they can think of nothing better than an increasingly rationalized world. They prefer a world that is not cluttered by choices and options. They like the fact that many aspects of their lives are highly predictable. They relish an impersonal world in which they interact with human and even non-human robots. They seek to avoid, at least in the McDonaldized portions of their world, close, human conduct. For such people, and they probably represent with each passing year an increasingly large proportion of the population, McDonaldization represents not a threat but nirvana. (Ritzer, 1993: 160-1)
Things to do!
Go and buy a hamburger, and while you are dining (sic) make a list of what you regard as the key features of the McDonalds approach to delivering your yummy repast. This might include:
• training people to do a limited task, rather than to be 'cooks';
• a limited product range;
• application of assembly-line principles;
• quick turn-over of customers;
• making customers do some work; « an emphasis on speed of delivery;
• the same wherever you go;
« an emphasis on calculation and measurement;
• a stress on 'quality' (sic) and 'quantity' (illusion of?);
• customer-driven (illusion of?).
Next, think of a public-sector service and consider how it might be made to be more like a McDonalds. Invent a McHospital, a McSchool, a McUniversity or a McPolice service. Then, after this flight of fancy, check out the reality: books which suggest that McDonalds may have a lot to teach the public sector (Osborne and Gaebler, 1992:114,167,182-3, 229; Wilson, 1989: 135-6).
QnCe upon a time, in a land where generations had grown up with McDonalds as an icon of rationality, it was decided to 'Ronaldize' the education system. Over a period of five years all teachers were sacked and fast-food chains were given the job of delivering education policy. In place of public education, private-sector-funded schools were set up. McTeachers were trained jn a limited range of skills and dispatched to McSchools, where the range of subjects taught was restricted to reading, writing and 'rithmetic.
The schools were identical, and the subjects were prepared and delivered in the same way. Schools employed McCentre-trained staff, motivated by the award of 'golden apples' which were displayed on the rather fetching McUniform. These 'Mclnstructs' used key McTexls designed (by a team of 25 experts) to give lessons on a programmed, predictable, efficient and controlled basis.
Where possible, computers were heavily involved in teaching. The parents of customers, or stunits, were encouraged to participate in the 'fun' of the school by helping to keep it clean, and could register for a McTeaching Unit qualification. The school Manager's task was to ensure that the quality and quantity of output conformed with the McManagers Manual. To this end each teacher had to complete a Star Date Manual on the completion of a task, such as meeting a reading target. Attainment measures (McTargets) were evaluated by the area manager and published on a weekly basis; they were also displayed on a screen outside the school.
A drive-thru service was available for parents who wanted to have a 'take-out teach' {or McHome video, CD, floppy etc.). Parents could choose (in a 'Customers' Conference') another company for the next franchise period. Or, if they had the capacity, they could simply drive to another educational outlet in the (vain) belief that another 'fast-education' school might do better. ♦
4.4.7  The limits of consumerism
The ability to exit from a public service is manifestly limited by other factors which Hirschman does not consider: economic and financial. Take the cases of private medical insurance and private education. If I consider that state provision is deteriorating to the point at which I review options, I may decide that enough is enough and I am going private. However, not everyone is able to do so. The capacity to exit from a public service to private provision is limited by financial resources and socio-economic status.
In such conditions the capacity to exercise exit is unequally distributed. Those who find themselves in such a position may view their captivity, in which they have little or no exit force and in which voice may not be effective (Hirschman, 1970b: 44-54), as nothing to do with loyalty to the national health service, so much as an expression of
530      PUBLIC POLICY
DELIVERY ANALYSIS 531
their dependency upon the service. They might want to go private but they cannot afford to do so. In this case dependency/loyalty suits in driving out exit and a total reliance on voice: the vote, political parties, interest groups, the use of the media. Even so, the voice option is one which, as Lindblom argues, suffers from considerable impairment (Lindblom and Woodhouse, 1993: 104-13).
The problem with exit as a mode of response is that, whereas formally all citizens have equality of voice, although their effectiveness will not be so equally distributed, exit is a response which is not available so readily to all, and, in many instances of public policy, may not be available to anyone except the very, very rich. There may, therefore, be no alternatives, or the alternatives may be too expensive for all but relatively few. In some instances, it may be impossible to exit from the 'externalities' of a public good: some goods and services impact on all even if some members of the community choose not to use them. I may have a private security force, but I cannot exclude myself from crime in society as a whole or riots which erupt on my doorstep, even if that doorstep is two miles away behind a security fence.
Two issues arise from this. The first is that policy-makers may seek to increase the capacities of citizens to get the best out of the public sector. They may see the 'marketization' of services in this light. Markets, it could be argued, are more responsive to voice than are hierarchies. The strategy here will be to make the citizen a more effective 'consumer' and create a relationship in which a 'consumer' voice is enhanced by improving 'access' to public agencies (Lamb and Schaffer, 1981). Measures to improve the 'access' of citizens as a means of enhancing voice could include:
• citizenship (Ranson and Stewart, 1989; Hall and Held, 1989; Coote (ed.), 1992);
• consumerization of public services (Gyford, 1991; Stewart and Clarke, 1987; Winkler, 1987; Martin, 1993);
• equal opportunities policies (Lupon and Russell, 1990).
The effectiveness of access and voice may, however, still be ineffective in communicating the responses of citizens to the values and instruments of policy. Policies to improve access may, in Edelman's sense, be simply rhetoric which does little more than window-dress rather than deal with the substance of the problems. In this case, the failure of voice and conditions of dependence (rather than loyalty) may induce disillusionment: inert exiting from the political process (figure 4.18). In these conditions citizens do not vote with their feet: they may (cf. Lindblom and Woodhouse, 1993) simply not bother to vote. The
Figure 4.18   'Active' and 'inert' exit
Income
Loyalty/dependence
'Active' exit
'Inert' exit
perception that public policy is not something which has much to do with them may well induce apathy and non-participation. Thus at two ends of the spectrum exiting may take place: those who can, 'opt out'; and those who cannot, 'drop out'. All of which puts further strain on those members of the community who stay put!
The problem of an optimal mix
In concluding his study, Hirschman draws attention to the elusive mix: a balance between exit and voice:
In order to retain the ability to fight deterioration those organizations that rely primarily on one of the two reaction mechanisms need an occasional injection of the other. Other organizations may have to go through regular cycles in which exit and voice alternate as principal actors. Finally, an awareness of tendencies toward instability of any optimal mix may be helpful in improving the design of institutions that need both exit and voice to be maintained in good health. (Hirschman, 1970b: 126)
The challenge of policy-making in today's world is to arrive at a balance between the need to facilitate choices that work in market conditions where exit and voice can be effective, and the use of modes of delivery in which the provision of services to 'customers' can also be equitable. Furthermore, in addition to the arguments about whether consumers (sic) can opt out of public goods and whether the resultant distributions are fair, there are a number of other objections which may be made to the model of consumer sovereignty (see Potter, 1988). Hood (1986: 173-80) , for example, puts toward the following limitations on the powers of exit and voice:
PUBLIC POLICY
• Is there a choice of suppliers? The lack of choice of suppliers significantly reduce the power of voice.
• Is there a choice of brands and quality?
• Is the consumer or user the best judge of his/her own interests? ^ They may not know enough to make informed choices. Some wants may have more merit than others. I
Clearly, consumer sovereignty has its limitations. The market for pub, \ lie goods and services is different to that for pork pies and toothpaste ; and such limitations do restrict the possibility of realizing a pUre| consumer model. In practical (as well as in theoretical) terms citizen- i ship is more than consumership. Hood (1986: 181-6) suggests that \u= ■ can identify a number of other lines of arguments against the consumer model: 1
• paternalist arguments that we cannot let the average citizen make? choices on things about which he or she knows little or nothing;
• competition, it may be claimed, will drive down standards to the lowest quality and level of safety;
■ stress and strain on providers as a result of consumer sovereignty will make life more difficult, if not impossible, for those involved in providing services;
• average costs of production will be driven down to make the product uniform.
These objections to the theory and practice of operationalizing a consumer model of delivery prompts consideration of what we know about how new modes of delivery actually work in practice. This brings us to the issue of policy evaluation. How can we evaluate the success or failure of new modes of delivery? In a study of change in the provision of public services in Britain, Le Grand and Bartlett oficr the following framework for evaluating 'quasi-markets' in social policy (the health service, community care, education and housing). The goals of 'quasi-markets' in welfare services, they suggest, are a mix uf values: improving efficiency, responsiveness, choice and equity. If the -improvements are to be attained they propose that they must meet the following conditions, some of which have already been covered earlier:
• market structure: if quasi-markets are to achieve their goals, there should be a situation of competition between many providers and purchasers. If there is a monopoly, then it should be broken up, but where this is impossible, there should exist a countervailing power;
« information: there should be accessible independent information about costs (for providers) and quality (for purchasers);
DELIVERY ANALYSIS
533
• transaction costs and uncertainty: transaction costs, especially with regard to uncertainty, should be kept to a minimum. Extra transaction costs must not be higher than they would be in the administrative systems they are replacing;
• motivation; providers ought to be informed by financial motives. They must be motivated to respond to market signals. Purchasers ought to be motivated to pursue the welfare of users;
• cream-skimming: there should be no incentives for providers or purchasers to discriminate between users in favour of those who are least expensive (Le Grand and Bartlett (eds), 1993: 33^)
The conclusions reached by the contributors to the project are, as we might expect, mixed:
The quasi-market reforms are in their infancy and it is too early to predict their long-term consequences ... but we could summarise the argument so far by saying that reforms which involve the district health authority as a purchaser of health care and the social service department as a purchaser of social care do not seem to hold out much prospect of gains in terms of efficiency, choice and responsiveness, but may not have much adverse impact on equity either; whereas the housing, education, GP fundholding and care-management reforms seem to hold out the prospects of real improvements in efficiency, responsiveness, and choice, but, unless the incentives for cream-skimming are reduced, may have a detrimental effect on equity.
(Le Grande and Bartlett (eds), 1993: 218-19)
However, as they point out, evaluation of this kind is problematic. It involves comparisons with other imperfect systems, and is constrained by the fact that reforms take time. State or hierarchical systems of providing health, housing and education have a much longer time period in which they may be judged. To be fair and methodologically sound, the evaluation of new systems of delivery needs more time to see if they work better or worse - or the same. But as a former British Prime Minister (Harold Wilson) once noted: 'a week is a long time in polities'. To which may be added Keynes's dictum that 'in the long run we are all dead'! Thus, although we need to make evaluations of policy and programmes, the business of evaluation is riddled with difficulties, not the least of which is time. It is to an examination of the difficulties of evaluation to which we turn in 4.5.
4.4.8   The control-consensus mix
In analysing the mix of instruments, mechanisms and modes of government which may compose a given policy field, an important consideration is the relationship between two sets of goals and values: the
534
PUBLIC POLICY
DELIVERY ANALYSIS
535
»(1968)
4
1
4
"4
desire for control and the formation of social consensus. Etzioni defines this problem in the following terms:
Control - The process of specifying preferred states of affairs and revisi on-going processes to reduce the distance from these preferred, state! Consensus formation - the process by which perspectives of the member -: of a societal unit are transmitted to the controlling overlayers and th differences among them are reduced, (p. 668) -..i
While there is no simple, straight line substitution, by and large, increase ing one element reduces the other, (p. 481) 5
The different mixes we have discussed above may in practice be said to represent formulations along this line of substitution. The deveM opment of market-type mechanisms and other public-sector reforms from the 1980s onwards were the result of increasing pressures to" control the costs of public goods and services and increase the effi-; ciency and effectiveness of their delivery. New public-sector management has been primarily control-orientated in its preference for 'hands-on management', measurement of performance, output control, breaking up units, competition, discipline and cost-cutting (see Hood, 1991: 4-5). A major feature of this control process has been the move to create a relationship between the citizen and public services and agencies similar to that which pertains between the consumer and a market place.
As we have noted above, there are limitations to this notion that the consumer of health or education can be like a shopper in a store. The danger may well be that the new public-sector mix, in achieving control, has undermined and subverted the other aspect of the mix: the building of a social, civic or political consensus. In emphasizing control through quasi-market approaches, the issue of how citizens relate to the political or public order has been somewhat neglected. Rtzioni himself has argued that this problem needs to be addressed through policies which rebuild the role of community, family and associations in an active society (see Etzioni, in 1.6; and 1993a, 1993b, 1994).
Public policy is not only concerned with the three Es - economy, efficiency and effectiveness - it is also about the two Ps - participation and politics. It is also not simply to do with 'exit, voice and loyalty', but also 'voice, ear and respect' (see Healey, 1993). It is through the promotion of democratic skills and a 'societal consciousness' (Etzioni, 1968: 224), however, that consensus-building may take place in the policy mix. The development of individual and community political education has, however, been absent from so much of what has passed for the re-mixing or 'reinvention' (Osborne and Gaebler, 1992) of public policy The 'community' approach to local government is perhaps
an exception to this rule, but in so many other policy fields and delivery systems, the relationship of the citizen to policy process qua politics is neglected in favour of managerialist controls, rather than political accountability and consensus-building.
Citizens' juries
However, there are some indications that the control-consensus-building mix could evolve in a way which allows public policy to be a source of public learning. An example of this is the growth in several countries - especially Germany and the USA - of 'people's juries' in which citizens participate in decision-making by subjecting witnesses - experts, politicians, bureaucrats, managers, interest groups and others - to judicial-style cross-examination. The end result is a report which can make a valuable input into the policy process (see Stewart et al, 1994). As John Stewart points out:
It is often argued that the public are not interested in participating. Apathy is the enemy and many in government will tell tales of frustrated initiatives in public participation. But these initiatives have too often worked against the grain of how people behave. Citizens' juries build on familiar and robust institutions. A local authority could use a citizens' jury to test out propositions on a school closure or a planning issue ... Problems of law and order could be explored ... mediation groups which bring together conflicting interests, deliberative opinion polls, and consultative referendums and electronic town meetings - all of these, as well as citizens' juries, can be seen as part of a strategy to strengthen democracy. {Stewart, 1994)
Decision seminars and the case for social think-tanks
Another approach which can build on familar and robust institutions is suggested by Lasswell's decision seminar technique and his idea of 'social planetariums' (see Lasswell, 1959, 1960, 1970b; Cunningham, 1981; Ershkowitz, 1981; Gorrell et al, 1993; and 3.10.5). Citizens in this model are not on the end of a delivery line, but are active in guiding choices and selecting and evaluating the mix. The centre point of this process should - in an information age - be the library or another knowledge institution, such as a museum, in which the society/community/region or whatever social unit, can think about where it has been, its present realities, the likely trends in the future, and what kinds of guidance or control they would like to exercise over their problems and opportunities, strengths and weaknesses. Lasswell suggested that there should be exhibitions which can facilitate this open exchange of values, ideas, beliefs, information, fears, hopes and dreams. The social planetarium would be a place where a community could observe itself, imagine possibilities and explore problems. It could become a place where the schools, churches, voluntary organizations, community groups, universities, and so on, in a locality present argu-
536
PUBLIC POLICY
merits, images, ideas and schemes. The aim would be to increase the self-knowledge and awareness of society so as to create a forum in which to enrich the political agenda.
Citizens would be encouraged to use this information-rich environment for such purposes as citizens' juries; town planning; discussion on social and economic problems; neighbourhood forums and so or\. Policy-making will thereby become a way of politicizing and analysing: a process of enlightenment rather than a delivery and consump-I tion nexus. Now, more than ever before, we have the capacity tot utilize public institutions such as the jury system, libraries and museums as 'prototype' (see Lasswell, 1963) social 'think-tanks' in which? men and women, young and old, are engaged and are active in think-;; ing about their society, making choices, shaping values and building consensus.
*:* The idea of social think-tanks or 'planetariums'
Lasswell envisaged that 'social planetariums' could extend the use of decision seminars (see 3.1D.5) to the community as a whole.
Using the idea of a planetarium where people can observe the sky, Lasswell argued that we should adapt it for social or self-observation (cf. Etzioni's 'societal consciousness' and Lindbiom's 'self-guiding society'):
Participants are liberated from the perceptual caves of the present. The planetarium technique makes it feasible to give proportionate weight to alternative versions of the future or the past... In principle, every community can build its own social planetarium where stress is put upon local objectives, local history, and local prospects ... [It is]... a means of giving importance to institutions that in many places are struggling feebly for public recognition, f refer, for example, to art museums, museums of local history, and museums of natural history. The contextual frame of reference - the orientation towards the future and toward decision-making - is a 'shot in the arm'. It makes the past pertinent to the present and the future ... In many circumstances ... it will prove expedient to amalgamate the social planetarium with a program that encompasses all the museums and related cultural resources of the capital of the locality, city, region, nation, or the transnational areas involved ,.. Gradually society can be changed until people learn to live as much in imagination of the future as in reminiscence of the past ... In such a configurative setting ... individual choices can be made at a respectable level of rationality. {Lasswell, 1959: 106-12)
What might they look like, bearing in mind the approach taken to policy analysis by those influenced by Habermas and other critical theorists (see 2.7.3, 3.10.4, 3.10.5) and in the light of the experience of 'citizens' juries'?
DELIVERY ANALYSIS
a building complex in a community, which could house rooms in which citizens' juries would sit to examine and contribute to policy-making; offices for the clerk of the citizens' jury system; an office for the director and staff to administer the think-tank;
facilities for researchers engaged on research in and with the community (see, for example, Bice etai, 1994; Phillimore and Moffatt, 1994);
, offices for people seconded from other organizations (such as universities, planning departments, and business) to work in the think-tank;
, exhibition space to promote community thinking;
. video, film and drama facilities;
t communication (computer and video) technology to provide for networking between think-tanks;
. rooms for decision seminars in which policy-makers, people and policy professionals could
engage in sustained long-term analysis and discussion of issues and problems; . library resources;
■ technical support facilities to support activities;
. community advice facilities and personnel to act as citizen 'advocates';" . rooms for political surgeries (for elected politicians to meet citizens).
* A prototype for such a system is the idea of improving the advocacy capabilities of citizens. For example, Bill Montgomery reports in the author's own neck of the woods on such a scheme in the London Borough of Barnet:
The highlight of the well attended annual meeting was the signing of a three-way agreement between Barnet Council's new administration, the Barnet Health Authority, and Choice. It means a new and improved advocacy service will be operated by Choice in the borough ... Choice works to ensure that all users of community services have a voice in how they operate. In particular it assists disabled people and others to establish their needs and aspirations. And it will represent their interests in a way they may not be able to do themselves. In Barnet it plans to support advocacy for all users of services, with the long-term aim of establishing an entirely independent system owned and managed by the users. (Edgware and Mill Hill Times, 24 November 1994:20) ❖
Social think-tanks or 'planetariums' could contribute to the growth of a society which thinks about itself, clarifies its values and is aware and informed, rather than passive and ignorant. Social think-tanks could help to widen and enliven the policy agenda and counter the influence of other organizations in the ideas business. In such an 'active society', Etzioni suggests:
a higher ratio of assets would be invested in political action, and intellectual reflection would have a higher, more public status. The status of political and intellectual activity combined would approximate to the status of economic processes in modern society. {Etzioni, 1968: 7)
PUBLIC POLICY
In other words, the 'public and its problems' (as Dewey, 1927 pressed it) may over time become more directly connected and gaged in forms of 'communicative rationality' (see Dryzek, 1987- ^ 3.10). The citizen could become less a 'consumer' than a co-prod ^1 of public policy. Hupe, for instance, reports that in the developments forms of 'co-production' in the delivery of social services in the Neth erlands: 'clients ... can be seen no longer merely as calculating dtj" zens, but rather as integral parts of the implementation structure f public policy' (Hupe, 1993: 149). The 'co-production' of policy is° phenomenon which has also been noted elsewhere (Bouckaert, 1993. Mattson, 1986), and is being spurred on by reforms in the public sector which have introduced more market and network forms of organization and less Weberian-style bureaucracy. A communicative approach to public policy may therefore develop out of or build on the changing and increasingly complex mix of delivery systems which have emerged over the last decade and more. The policy sciences should - as Lasswell long ago hoped - function so as to facilitate this interaction between the public and its problems by helping to clarify values, formulate arguments, undertake research in and for communities (see, for example, Rice e.t al, 1994; Phillimore and Moffat, 1994-Lincoln and Guba, 1985; and 4.5.4), and inform and widen public debate. That is, it should assist in the production of what Stacey (1994; 89) terms 'people knowledge', as in the production of professional: and policy knowledge.
The citizen - unlike the consumer - is both a producer and consumer. The model of exit, voice and loyalty therefore omits another quality of political life in a democracy: input. The citizen should be seen as someone who can be involved in making a productive input to public policy. This requires that the state should aim to be as concerned about the promotion of what Etzioni (1968: 24) terms a societal con-sciousness as it has been in recent years with the promotion of efficiency, effectiveness and economy.
In re-mixing the relationship between control and consensus-building we also address another crucial area of public policy: evaluation. As we shall see below, the evaluative process has been primarily framed by managerial or technical values. However, if we regard control as having a political and wider social dimension, then evaluation should provide a significant opportunity for social learning and the promotion of democratic skills (see, 4.5.4, 4.6.4) or active 'citizenship' (Coote (ed.), 1992; Warren et al, 1992). Citizens in this model should be actively involved in making an input to the evaluation of public policy in ways which are more meaningful than voting either through the ballot box or with their feet. It should not be the monopoly of the
DELIVERY ANALYSIS
539
manager, expert or bureaucrat (see Lindblom, 1990). In practice, of course, it has been ever thus (for critiques of the dominant evaluative paradigm, see 4.5.4).
Control and accountability
An essential aspect of the control-consensus mix are the modes and methods of accountability. This has two dimensions: political and managerial (see Hughes, 1994: 240-55). In the last decade or so the main feature of the accountability mix is the extent to which managerial approaches have been seen as applicable to the public sector. The politics/administration divide, as framed by Wilson, Weber and others, posited that the primary task of the political process was to hold the bureaucratic workings of the state to account for its activities. A strong legislature and party competition was viewed as the primary way in which bureaucracy could be kept under control and steered (see Weber, 1991: 43). However, the arguments of those advocating privatization and managerialist reforms of the public sector in the 1980s cast doubt on the effectiveness of politics as a means of controlling bureaucracy. Day and Klein (1987: 51) argue that as society and government have become more complex, so the relatively simple notion that voters elect representatives who hold civil servants to account has become a less-convincing theory. In surveying accountability in five different public services (health, police, water, education and social services), they conclude that for accountability to be a reality it requires a mix of political/community accountability supplemented with the use of managerial techniques which focus on the performance of those who actually deliver public services. They conclude that accountability
must be seen in terms of individual institutions but as a system which is woven into the fabric of political and social life as a whole ... [however] this may, in turn, bring about excessive complexity in the machinery of accountability and at the same time create dead ends. So, why not concentrate less on formal links or institutions and engage more in a civic dialogue to recreate at least something of the high visibility and directness of the face-to-face accountability with which the story of the word began. (Day and Klein, 1987: 248)
In practice this weaving of accountability into the social and political fabric has taken place through privatization and the introduction of management systems, forms of budgetary control, and performance appraisal and measurement in most OECD countries (see Hughes, 1994: 247; OECD, 1991,1992, 1993a and b). Allied to this has been the idea of cultivating a 'client' or 'customer' orientation/culture in public-sector agencies and personnel (see Hughes, 1994: 248-50; OECD, 1987). However, the notion that we should also develop a more Athenian, 'face-to-face' form of accountability in which citizens can engage
Í
540
PUBLIC POLICY
DELIVERY ANALYSIS
541
in a 'civic dialogue' has been far less in evidence than managerj t-reforms. As we have discussed above in this section, control alsoy, to address the issue of how citizens can make a real and meanin t; input into the process, so as to improve the political or civic dime ' sions of accountability. Indeed, it may be argued that in seeking \ 'improve' on the political model of control, managerialism may Wclj^ be undennirung representative government, thus leaving little ground i for hope in terms of a more participatory - or in Etzioni's sense 'active' - democracy. As Andrew Massey eloquently puts it; ';
Public administration cannot be judged by the same standards reserved^ for retailers, nor can it be run along the same lines. There are no generic laws of management that apply equally to the public and private spheres -... Managerialism is a very weak bulwark to defend democratic govern-^ ment against malpractice, however many citizens' charters may be inA vented. The difficulty for observers in assessing the worth of the New-Right's reforms is that of timescale and information. Government's own performance indicators have changed, becoming more managerial, whilst' questions regarding morale, integrity and public service will take years to:? answer ... The emphasis of the New Right has been on management and? control as a way to ensure the accountability of the state, devolving power out to the market and individual consumers. In short it has bueii an attempt to get the state to wither away. This is not the way to protect liberty. The wisdom of America's Founding Fathers was clear upon this point: if only men were angels we could do away with government. (Massey, 1993: 200)
The danger is, as Massey maintains, that managerial controls which aim at eroding the state will end in destroying the very foundations of political systems (such as those of Britain and the USA) which have provided liberty for generations: |
To the New Right with their obsession with efficiency, accountability may appear like Aristo's serpent, 'a hateful reptile' that hisses and stings. Yet it is the foundation upon which the two countries have avoided overbearing ^ government: accountability may be inefficient, 'but woe to those who in; disgust shall venture to crush her'. (Massey, 1993: 200-1)
With this resounding defence of the theory of representative govern- ■ ment ringing in our ears, let us now examine the issue of how public policy in practice may be evaluated.
jvjee production, delivery and evaluation. It should also be read in the context of 4.5.4 and
*> Linking policy delivery, evaluation, research and people knowledge
J. Popay and G, Williams (eds), Researching the People's Health, 1994
This book provides an excellent link between the kinds of issues which we have been discussing in 4.4.8 with regard to the relationship of people, as 'consumers' and citizens, to
book examines two main issues: the role of social research in health policy, and the Ijonship of lay and expert knowledge to the shaping and evaluation of health care policy.
case of the role of research, the authors draw attention to the fact that social science has
In the
Tiportant contribution to make to health policy, yet it has been grossly underused, underfunded, undervalued and underfinanced (Klein, cited p. 19). Although several official pronouncements have argued that social scientists do indeed have a major part to play, in practice -■-arch into policy outcomes has had a relatively insignificant impact on the policy process. Hunter, for example, says that there is an urgent need for more interaction between policy analysts- practitioners and policy-makers - in Britain as in other countries (Hunter, 1994: 26-9).
One of the problems is that actually evaluating the outcomes of health policy is a complex methodological issue. As Andrew Long comments:
Although outcomes are high on the policy-making agenda, progress will not be straightforward. Despite the ultimate goal of health services to improve the health of patients, it is still difficult to measure health outcomes. Indeed, there is no agreed taxonomy. This is partly due to an underdeveloped theoretical framework and a paucity of people equipped to develop, apply and interpret outcomes measures ... Evidence on effectiveness and outcomes and an emphasis on health gain and health outcome provide an apparently value neutral, rational approach and means of rationing health and social care. Beneath the range of technical issues in assessing the outcomes are political and social values that need to be explicit. =•   [Long, 1994: 162, 175)
The long-term agenda, he believes, should be focused on the clarification of such values involved in evaluating whether a policy leads to improvements or to deterioration. This issue of clarification as an aim of research is taken up by the other theme of the book: the role of lay knowledge and participation, which has been discussed in 4.4.8. A number of contributors stress how getting more input from users/customers/citizens has increasingly been seen as a vital part of new research designs.
An interesting example of this is the Glasgow (Corkerhill) study into child accidents, which involved local people. Instead of doing research on a community, researchers worked with lay people to analyse the problem of child accidents, and developed recommendations to reduce them. The contribution written for the volume is jointly authored, and includes one of the local people involved in the project - Cathie Rice. They conclude that, although their work does not provide a 'knitting pattern' to improve child safety: 'The combination of the research and the parents' action group is one we feel may be an effective way of producing local data and exploring ways of making communities safer for children' (Rice etai, 1994: 132).
Another example of designing research so that people are participants and producers, rather than objects or guinea pigs, is the use of Rapid Appraisal (RA) methodology - a method of identifying and prioritizing community health needs. The authors argue that we ought to rethink the relationship between decision-makers and users, and make citizens an integral part of the evaluative process. This rethink will involve, they believe, a 'a paradigm shift whereby the
542
PUBLIC POLICY
community perspective will be used as the guiding principle for setting priorities in health care' (Ong and Humphris, 1994: 80).
As the editors of the book argue, there is a growing body of opinion that in health polta* research and evaluation there ought to be provision for real participation by lay people (Popav and Williams (1994a): 2, 6). Such a conclusion, it seems to the present author, has an urgent relevance to other policy fields. This means that - as we note above in 4.4.8 - the lay voice should be taken seriously and that the enhancement of that voice should be a matter of conce for both policy analysts and policy-makers. ♦ m
4.5 Evaluation
• Key texts
The textbook in the field is Rossi and Freeman (1993). This gives the essentials of 'traditional' evaluation research. It is also supported by an excellent workbook.
In order to develop a good grasp of the development of the subject the reader need go no further than a variety of books published by Sage publications (including Rossi and Freeman) such as:
• Dolbeare (ed.) (1975) for a review of the field in the mid-1970s;
• Saxe and Fine (1981) on experimentation;
• Palumbo (ed.) (1987) for a collection of papers reviewing the changes in evaluation and the new frameworks.
On new approaches to evaluation, see:
• Lincoln and Guba (1985);
• Guba and Lincoln in Palumbo (ed.) (1987) give an account of their 'naturalistic' approach;
• Miller (1984) and Bobrow and Dryzek (1987) for the 'design' approach;
• Cook (1985) on multiplism;
• Rita Mae Kelly's essay on the politics of research methods in Palumbo (ed.) (1987) is an important read on the development of new approaches to evaluation. The section on the alternative approaches (4.5.4) draws on her article.
For the impact of audit, see Henkel (1991) and Power (1994). ♦
Policy evaluation: linking theory to practice
DELIVERY ANALYSIS
543
Rist (ed.), (1995) is one of the most comprehensive surveys of evaluation available and is an essential reference book for studying the role and impact of evaluation in policy cycle terms. It
ontains some of the most important articles in the field published in the US and Europe. It is divided into three parts:
policy evaluation in the policy arena Theoretical perspectives policy evaluation and policy instruments policy evaluation and utilization policy evaluation and governance
, policy cycle policy formulation Policy implementation policy accountability
• Sectoral policy evaluation The environment Health insurance Education
Economic development Industrial relations Energy Child support Housing ♦
4.5.1   The modes and phases Evaluation has two interrelated aspects:
• the evaluation of policy and its constitutent programmes;
• the evaluation of people who work in the organizations which are responsible for implementing policy and programmes.
In this section of the book we shall examine evaluation in the context of two dominant frameworks: as a form of rational analysis (4.5.2); and as a tool for the management of 'human resources' (4.5.3). We shall also review critical and alternative views of evaluation in the formation and execution of public policy (4.5.4). These sections follow, in broad terms, the phases in the historical development of evaluation over the last thirty years or so:
• 4.5.2 The 1960s and 1970s saw the emergence of the subject in its modern form, as marked by the publication of books such as those by Campbell and Stanley (1966) and Suchman (1967) and the setting-up of the journal Evaluation Review in 1976. As Henshel notes with reference to the US, where the expansion of evaluation was