A policy aims to resolve a social problem that has been defined as politically relevant to the public arena (see Chapter Seven). Once it has been programmed and implemented (see Chapters Eight and Nine), a policy is – or should be – subject to systematic evaluation. During this final stage of the policy life cycle, the analyst focuses on the effects generated by the state measures. In concrete terms, this means establishing the benefits and costs of policy, including where applicable whether groups have effectively modified their behaviour. In summary, policy evaluation involves the empirical testing of the validity of the causality model on which the policy is based. Thus, the analysis concerns both the relevance of this ‘theory of action’ and the scope of its practical application.

This chapter deals with ‘evaluative statements on policy effects’, which is the sixth policy product in our analysis model (see the introduction to Part III). We start by providing operational definitions of the concepts of policy ‘impacts’, which take account of changes affecting target groups (including changes in their behaviour), and policy ‘outcomes’, which describe the effects actually generated among the end beneficiaries (Section 10.1). There may be related effects on third parties. Based on these two variables, we present the five criteria that are generally applied when evaluating the effects of a policy: extent of impact, effectiveness, efficiency, relevance and productive economy (Section 10.2). These two preliminary stages then enable us to identify the form and content of the various evaluative statements that can be observed in political-administrative reality (Section 10.3). Finally, we identify the principal actors of the evaluation stage, their direct and indirect games and the resources and institutions mobilised during the process of the production of these evaluative statements (Section 10.4).

10.1 Definition of policy effects

While the concept of the formal implementation act (‘output’) identifies the final products of political-administrative processes (that is, the tangible results of implementation), the ‘impacts’ and the ‘outcomes’ concern the real effects of a policy in the social arena. Thus, at this level, what is involved is the empirical testing of the relevance of the intervention hypotheses (did the target groups react as anticipated?) and causality hypotheses (do the end beneficiaries see their situation improving?). To facilitate the empirical analysis of these effects, we present below the operational dimensions that make it possible to identify and qualify policy impacts and outcomes.