Houlihan, B. (2005) 'Public Sector Sport Policy. Developing a Framework for Analysis', in International Review for the Sociology of Sport, 40 (2): 163 - 85.

[This article argues first of all that policy must be analysed at a suitable  'meso' level, since policy-making institutions and personnel have dynamics of their own and are not simply determined by broader economic and social forces. Secondly, the article discusses four major recent models in policy analysis and evaluates them. ]

Actual sport policy remains largely unanalyzed, despite the importance of sport in a great deal of social policy, and despite considerable government funding, both in the UK and in other countries. There is no specialist journal for sports policy, and only a few articles in other journals. Instead, sociological cultural and economic analyses dominate. The sports policy process itself, the way different interests are articulated, and the way in which power works needs specific analysis at the level of  'national organisations  (ministries, national sports organisations and interest groups)' (165).

Classically, policy analysis elsewhere focuses on 'problem identification, agenda-setting and policy formulation', but this approach follows a  'neo-positivist/rationalist epistemology reflected in an uncritical assumption of the separation of fact from value, the privileging of quantitative methods, and the search for generalisable results' (165). The failure of many government policies focused attention on the implementation process and on evaluation. In turn, this led to a discovery of the 'role of politics, interests and values in shaping policy through a continuous process of negotiation and interpretation', and the discovery of persons with power at low levels in the official hierarchy (166). At the same time, debates in social sciences generally reintroduced matters such as the 'structure/agency debate... gender inequity... and the role of  "policy entrepreneurs"' (166). The emergence of important non-state agencies meant the consideration of a new role for government in control and regulation rather than direct policy formulation. The earlier, neo-marxist, position has also been contested by those who see ideas as variables in their own right, not reducible to interests or strategy.

As a result there are a number of competing frameworks for policy analysis. These can be evaluated against four criteria:
(1) whether they can explain both stability and change -- analysis of structural factors [economic or social, natural or historical constraints] needs to be combined with accounts that give due emphasis to the role of agents, including their capacity to lead and motivate, and an acknowledgment of  'the role of ideas and epistemic communities' (167). [it is surprising to find no reference to Foucault here].
(2) whether they can explain the full range of the policy process, take a holistic approach.
(3) whether they are applicable across a range of policy areas, while allowing for the distinctive features of sport policy: sport policy clearly overlaps with processes in health and education, for example.
(4) whether they can sustain historical analysis over a number of years and thereby separate out minor fluctuations from major explanatory variables.

The stages model
This presupposes the series of discrete stages  (deciding, defining issues, setting objectives, evaluation and so on), borrowing from rational action models. It has been useful in focusing on stages such as agenda-setting. Houlihan refers to his own analysis of public policy responses to football hooliganism -- it was used for end agenda-setting, but not so useful in describing the implementation and impact of policy, the details and the  'ideological imperatives that led to shifts in the definition of the problem over time' (169). Stages models also tend to be descriptive, over rational, formalist in its focus on official policy-making, and over-simple in assuming that policy making is a clear rational process. [highly reminiscent of the now forgotten critique of planning by objectives in education]. In terms of the criteria above, change is described without reference to underlying issues of power, it focuses unduly on rational aspects of policy, it has been applied widely, but also subject to extensive criticism in other areas, and it seems to be best at describing particular moments rather than whole sequences.

Institutional analysis
One strand emphasizes institutions as organisations, and another 'shared values norms and beliefs'  (170). Both approaches argue that institutions are relatively autonomous, important mediations between agents and structures, rather than simple arenas for policy. Cultural institutions explain the essential background to policy decision-making --  'operational rules... are nested within other more deeply rooted rules, namely, collective choice rules and, at a deeper level, constitutional choice rules which are more resistant to change' (170). Some examples of the approach are provided on p.170, where particular organisations are seen as having a definite impact on sports policy, while cultural institutions provide a context for the classic analyses of class, gender, disability and ethnicity. The approach seems to include both actors and structures, and points to important aspects of state provision. Nevertheless, the approach seems to work best as an  'analytic orientation or sensitising concept' (170). It often assumes a rational actor model, and can slip into institutional determinism. It can ignore the broader origins of ideas and interests. This clearly affects the performance of the model against the first two criteria. However, it has a wide range of applicability, and it does emphasise the history of institutions. Overall, though, it points to the importance of institutions  'without explaining the extent to which they matter and the circumstances in which they would matter more or less' (171) [I'm not sure this must be so-- eg the work on micropolitics in education is clearly linked to Foucault's more general approach to power and discourse?].

Multiple streams
This emphasises organisations as anarchic rather than rational, and makes links with the  'garbage can' model of decision-making  [much application here to educational organisations -- see Westoby]. Three streams are involved in policy:  (a) the problems stream contains issues that have been identified as problems [not ignored] requiring action, which arise from crises or  [moral panics] such as the obesity problem;  (b) the policy stream, a loose collection of various ideas  'sponsored by particular policy communities... or... adopted by policy entrepreneurs' (171), such as an enthusiasm for privatisation or school PE, which can get prioritised in particular circumstances, such as when they become technically feasible;  (c) the political stream, specific to organised political parties, government, or interest groups. These streams can coincide to provide  'a  "launch window"' which gets an issue into policy-making. Sometimes, ideas can spill over into other areas . Overall, the model stresses randomness, the role of ideas, and the absence of rational decision-making.

This approach can clearly link with other concepts such as policy communities, spillover and the policy entrepreneur which are particularly useful in explaining sport policy, although actual examples are not common  [some are given on page 172]. However, the emphasis on randomness and on ideas may underestimate  'structural factors and institutionalised power' (172); it stresses agenda-setting and is less successful with implementation; it seems most successful when applied to the current UK and not to other political systems which are more centralised. It does allow some historical analysis.

Advocacy coalition framework
Various coalitions compete for influence in policy. They are often held together by common belief systems --  ' "deep core"  beliefs which cover basic values... [such as]... gender relations and... property rights;...  "policy core"  beliefs which are the basic normative commitment within the [policy ] subsystem;... secondary policy core beliefs which are narrower beliefs... [such as]... the seriousness of particular issues and the details of resource allocation' (173). Coalitions can conflict and may be reconciled by a  'policy broker'. Beliefs may change over time, which is partly based on learning and partly on external changes such as personnel or external conditions. There have been a number of applications of this model  (examples on page 173). The model seems able to explain both stability (via core beliefs ) and change, although Houlihan finds matters such as policy learning to be  'less convincing' (173). Coalitions are not usually analysed in detail, and the role of power is still undertheorised in favour of an notion of rational adaptation and learning. Any irrationality in the system is seen as resulting from  'factors such as limited time, computational constraints, or cognitive dissonance rather than the manipulation of the policy agenda through the mobilisation of bias' (174). The model has been widely applied, and in analyses of sport, the outlines of the model and its utility have been confirmed  [examples on 174]. Thus the model meets the second and third criteria, and clearly requires historical analysis, thus meeting the 4th one.

The models reviewed above have different combinations of strengths and weaknesses. The stages model still seems to persist despite its weaknesses  [I think because it simply expresses the ideology of  managers and capitalists]. All agree on the importance of the institutional context, and all point to the complexity of sports policy in the UK with its combination of departments and agents at various levels. The last two models seem particularly suitable to grasp the complexity of policy-making in modern societies. They also seem to offer a valuable corrective to the rational action assumptions in many discussions of policy. Notions of policy entrepreneurs and policy brokers seem particularly relevant to sport.

The suggested approach involves taking the advocacy coalition framework and developing it, especially the way in which it analyses both political ideas and the systems and interests involved in the political system. For sports specifically, ideological  '"storylines"  (Fisher, 2003)' (176) seemed important as  'deep-seated policy predispositions', but the action of actual government mechanisms and interest group activity are also influential. One model suggests that policy operates at a number of relatively autonomous levels, from administrative arrangements at local level, through national agencies, to the  'deeper layers of the social formation' (177)  [with no determinism from top-to-bottom -- local agents can also  'develop relatively stable preferences' (177)]. [There is a useful model on page 178-9 ]

The UK system in particular features 'administrative dispersal' (177) across a number of agencies, and overlaps with other large areas such as education and health. Certain ministries had developed distinctive cultures which limit overlaps -- UK Sport has developed a 'dominant culture of elite achievement', which limits its capacity to take on  'the country's anti-doping efforts' (177). Resources, including expertise and capacity are also widely dispersed, and their provision involves power and dependency -- for example national lottery funding  'has increased markedly the influence of central government and its agencies', while  'UK Sport has used its resource control to force national sports organisations to  "modernise"' [i.e. rationalise] (171). As the a result the state has become extended and also  'congested', and has 'retreated from direct service delivery but has extended its capacity to steer policy sub-systems through the strategic control over resources' (180).

Interest groups are also important --  'demand groups... [consumers], provider groups... such as leisure services managers, PE teachers, coaches... sports clubs, direct support groups... such as national governing bodies, commercial sponsors, local authorities and schools, and indirect support groups  (related local authority services including land use planning, community development... and commercial sponsors)' (180). These groups both implement government policy and place demands on it. Government often intervenes by pursuing policies of  'modernisation'-- for example by forcing both demand and provider groups to merge. Such interest groups have grown steadily in recent years. Their main role may be  'to protect recent resource gains' rather than to actively shape agendas  (180, with a reference to a study by Houlihan and White, 2002).

The overall appearance of pluralism is also restrained by  'the dominant policy paradigm, the service specific core policy paradigm and... the rules of structure formation' (180) [still dominated by central government]. Examples of the dominant policy paradigm include the push towards privatisation in the Thatcher years, or the stress on social integration under Labour. The service specific policy paradigm 'nests' inside, and covers matters such as emphases on sport for all or elite development -- or the recent attempt to argue that elite development can lead to increased participation  [discussed page 181]. Ideas clearly affect these paradigms. There are not always based on evidence, but on storylines as above --  'decision making short cuts'(181) , bundling together factual information, value orientations and  assumptions -- that sport builds character, or that men and women have different needs. The storylines themselves are the subject of struggle between interests, and  'deeply entrenched and longer-established biases within the policy process' (181). They have a role in limiting policy actions, and work rather like cultural institutions.

Individuals such as policy brokers also have a role, especially in  'congested policy-making systems' [lots of competing but not dominant interest groups and a degree of loose coupling?]. Another study, forthcoming, promises to deliver some evidence on this.

Problems still remain, in separating different levels of analysis and explaining their articulation and relative autonomy  [we know how notoriously ambiguous these terms are from marxist analysis of the state]. The model does acknowledge complexity and lack of predictability, the role of both structures and agents. There is no obstacle in the way of sophisticated policy analysis in sport to complement the other kinds.

back to key concepts