CASE STUDY PROPOSAL: THE ROLE OF AGENCIES IN POLICY MAKING Salvador Parrado & Sandra van Thiel 6 February 2009 Introduction Agencies carry out all sorts of public tasks, such as inspection, financial payments, tax collection, quality assessment, settling disputes, registration, and so on. These tasks are based on policies (laws, regulations, directives), developed by the (central) government. It is often assumed that agencies only implement or execute policies but are not involved in the development of policies or the decision-making process on new policies (cf. Verschuere, 2009). This assumption the policy and administration divide was strengthened by the advent of the New Public Management, which led to an increased number of semiautonomous agencies, at arms length of the government. However, there are also examples in which agencies do have influence on or play a role in the development of new policies, either by design (see e.g. Norwegian agencies, Christensen & Laegreid, 2003) or on an haphazard basis depending on political saliency, agency size, knowledge intensity of policies, or agency management capacities (Verschuere, 2009; Moynihan & Pandey, 2006; Jacobsen, 2006; Pollitt et al., 2004; Svara, 2001). In this paper we propose a design for a comparative case study into the role of semi-autonomous agencies in policy making, to describe whether and how agencies play such a role and to offer explanations for possible differences in influence between agencies, countries, and types of policy. Research questions 1. (How) do semi-autonomous agencies play a role in the development of the policies they carry out? 2. Which conditions explain the type, intensity and shape of the role that semi-autonomous agencies play (or not) in the development of the policies they carry out? These questions are relevant because they deal with a number of aspects, such as: Accountability, transparency and legitimacy issues. The policy-administration divide was expected to create a clear division of tasks: ministries develop policies and semiautonomous agencies implement them (cf. steering not rowing by Osborne & Gaebler, 1992). However, in practice this divide is not easy to make. Moreover, good quality policy-making requires information from the street level agency employees, because they
are the experts (knowledge) and in touch with the customer or user of public services on a daily basis (information). The role played not only by the agencies themselves, but also by parliament, professionals, interest groups, and civil servants from the ministry. The interplay between multiple actors raises important questions about the primacy of politics and the neutrality of executive agencies (cf. Vibert, 2007). Critics of agencification have for example pointed at the risk of capture (Wilson, 1989) of agencies by rent-seeking interest groups. The possible influence of agencies on how their performance and operations are undertaken, controlled, and measured. A number of researchers have studied the emancipation of agencies and the horizontalisation of the relationship between for example parent ministries and semi-autonomous agencies ( t Hart & Wille, 2006; Veenswijk & Hakvoort, 2002; Spence, 1997). Not only do hierarchical steering and control mechanisms no longer apply when agencies become semi-autonomous, but semiautonomous agencies also spread their wings in discovering new markets and new partners (cf. Van Thiel, 2008). Comparative case studies A case study approach enables researchers to study complex phenomena and their context intensively (Gerring, 2007; Barzelay, 2007). The limited number of units of analysis (cases) is compensated by the large number of (explanatory) variables, which is an important benefit to the study proposed here. The role of agencies in policy making processes may take on different forms (techniques, instruments, project meetings, documents) and may be contingent upon a number of characteristics, for example agency characteristics, the policy itself, politico-administrative traditions such as institutionalised legal procedures, the saliency of a policy, the presence and representation of stakeholders, and the timing (e.g. in times of crisis). The goal of the proposed case study is to explain the role of agencies in policy-making and differences therein between tasks/policy sectors and between countries. The specific task or policy that an agency carries out is therefore an important criterion for the case selection. We propose to select semi-autonomous agencies for this project that carry out different types of policies. There are several typologies of type of policy available. For example, Lowi (1964) distinguishes between distributive, redistributive, constitutional and regulatory policies. Another typology, by Wilson (1989), differentiates agencies on the basis of visibility of their output: production organisations (e.g. tax agency), craft organization (e.g. inspection
agencies), and procedural organizations (e.g. mental health counselling service). Pollitt et al. (2004) applied a similar logic in their selection of agencies, carrying out the same task (prison, social security, meteorology, forestry) in four different countries (Finland, Sweden, UK, Netherlands). For the case study proposed here, the typology of Hood and Lodge (2006) on public service bargains has been preferred and adapted. Public service bargains are defined as explicit or implicit agreements between the civil service and those they serve i.e. politicians as representatives of citizens. When we apply this concept to semi-autonomous agencies, public service bargains are the explicit or implicit agreements between organizational heads of agencies or any kind of organization that deliver goods or regulates the way in which those goods are delivered and the parent ministry. Hood and Lodge (2006:21) identify eight types of public service bargains based on two main types. In trustee bargains public servants (or agencies) are expected to act as independent judges of the public good. They will serve the general interest in an independent manner, and not merely to take their orders from some political masters). Examples of this are pension funds, central banks, tribunals and independent regulators like competition authorities. Agency bargains are defined (Hood and Lodge, 2006:44) as agreements in which the agent works for a political master (the principal) 1 to carry out specific tasks, within legal boundaries or a legal framework (like statutes, or more generally the Constitution). As agents will be held accountable, they have a fiduciary duty to give full disclosure about their activities to the principal. Hood and Lodge (2006:50) refine the concept of agency bargains into two sub-types. In delegated agency bargains agents work within a general framework agreed with the principal, with discretion to operate within that framework. Examples are agencies that perform their tasks through framework contracts with the parent ministries and consequently have much discretion and limited steering such as research institutes or the tax service. In directed agency bargains agents work on the principal s instructions. Examples are agencies that have to follow exact instructions when carrying out their task and/or agencies that are under direct supervision by the government, such as the prison service or agencies that pay certain highly regulated benefits like unemployment, social welfare, or student loans. Each type of bargain will lead to a different type of institutional arrangements i.e. relationship between agencies and ministries and hence to differences in the role of agencies in policy-making. For the selection of cases we propose a strategy in which per country three similar agencies are selected, each with a different public service bargain structure (i.e. different cases but 1 Please note that the principal is not always also the customer; citizens of private businesses may for example be the customer of a service that is paid for by the political principal (from tax revenues or other public sources).
similar across countries; Seawright & Gerring, 2008). Although we aim to select semiautonomous agencies only, we cannot guarantee that this will be possible for all countries. Prior research (by the CRIPO team) has shown that the cross country variation in organizational types of agencies is very large (see also Pollitt et al., 2004; Pollitt & Talbot, 2004; OECD, 2002). Table 1 offers a number of examples. The final selection of cases will depend on the preferences of the researchers who participate in this study and the presence of potential cases in their home country. 2 Table 1. Possible cases for comparative study Trustee bargains Delegated Agency Directed agency bargains bargains Regulation of the financial Research Child support system Competition Policy Meteorology Prisons Telecommunications Road maintenance Social benefits Police authorities Hospitals National heritage management Railway (or airport) Broadcasting companies Immigration Theoretical explanations Based on existing literature we can offer some preliminary explanations for differences in the role of agencies in policy-making processes. A more comprehensive framework needs to be developed however. Also, we need to keep an open mind about the possibilities of theory development after the empirical research (cf. Mahoney, 2001:580). There are so many potentially important variables that it will be very difficult to decide beforehand which variables have the highest explanatory power or the strongest causal effects. Comparative case studies will allow researchers to determine which variables have which effect, its intensity and through which mechanism, because case study data offers rich informative accounts and detailed contextual information which can help to interpret how and when particular causal mechanisms are taking place. By discovering patterns in the data, hypotheses can be generated to build new theory. Or as George and Bennett (2005:206) put it: Process-tracing forces the investigator to take equifinality (multiple causality, SP/SVT) into account, that is, to consider the alternatives paths through which the outcome could have occurred, and it offers the possibility of mapping out one or more potential causal 2 Additional data on the presence of agencies will be made available from the 2008 survey among the CRIPO team by S. van Thiel. Results to be presented at the 2009 meeting in Brussels.
paths that are consistent with the outcome and the process-tracing evidence in a single case. With more cases, the investigator can begin to chart the repertoire of causal paths that lead to a given outcome and the conditions under which they occur- that is, to develop a typological theory. Some examples of possible explanations of differences in the role of agencies in policymaking processes are: Type of task: this (independent) variable is related to the aforementioned public service bargains (see table 1). Tasks in trustee bargains are of regulatory type, while agency bargains are of executive nature. As such one could argue that regulatory tasks require more expertise and independence, hence agencies will have more influence on policy contents and development than agencies with a more executive task, especially in directed agency bargains. Policy sector logic: policy sectors may have their own logic or style when it comes to involving agencies in policy-making and/or concerning the influence of the sector (interest groups) on the agency (or ministry) in role in policy making. High degrees of corporatism might facilitate agency influence but could also lead to capture of the agency (Wilson, 1989). Trajectories of reform (state tradition): theories on state tradition and policy styles (Richardson, 1982) suggest that different countries may have different preferences for agencification but also for the interaction between government and agencies. In some countries for example agencies are by design involved in policy-making (Scandinavia), while in other countries the policy-administration divide is very strict (Anglo-Saxon). Another potential explanation related to state tradition points to the effects of consensual policy-making styles or adversarial styles; in the first agencies will be involved, in the latter probably not. The presence of numerous countries in the CRIPO project will allow for a good test of such differences (cf. Pollitt & Bouckaert, 2004). The level of Europeanization of the sector in which the agency operates may vary, for instance between high in policy competition and low in child support or national heritage management. Such variations might be expected to contribute to the influence of agencies in policy-making processes particularly if agencies are involved in EU committees and networks. Single versus multipurpose agencies: agencies often perform more than one task at the same time. One could argue that agencies with multiple tasks can exert more influence because they have more expertise and resources, or at least more opportunities for influence in multiple policy-making processes. However, for the sake of simplicity, it might be better to leave this variation out of the current proposal.
Agency leadership: the role of particular actors (i.c. the heads of organization) vis-à-vis the principal could be important to explain differences in agency influence on policymaking. In this regards, the distinction of Hood and Lodge (2006:110) could be helpful; they differentiate between personal loyalists (who have been appointed by and follow the bidding of the principal) or serial loyalists (who are loyal to the government of the day, no matter what principal is in charge of the appointee in any given time.) Following this terminology, we can expect jester, judge, executive and partner bargains. It would be expected that trustee functions are filled with judge bargains by which players are semi-autonomous and have loyalty to some higher entity like the State, the Constitution. They are likely to have an independent role in policy-making. In delegated agency functions, we should find executive bargains of those who maintain loyalty within specified terms and act to pursue some set of defined goals in a limited period of time and space. In direct agency bargains, jester bargains are expected to be found in which agencies act as reality checkers for rulers. Finally, partner bargains are expected to found in any of the three categories. This involves acting as collaborator of principals, with the right to be heard. Kind of policy: conditional zweck-program Each of these explanations could be elaborated further, either before, during or after the empirical research depending on the choices of the researchers who are going to carry it out. Also, several combinations of explanations might be tried out. Research methods When the final case selection has been made, we propose that for each agency one major policy is selected and studied in detail to trace the role of the agency in its development (ergo, the policy has to have been developed already although not necessarily implemented). A combination of methods will be used to ascertain this role or influence: Content analysis of documents such as the selected policy, minutes from project meetings about the development of the policy, reports of discussions in parliament about the policy, reports from advisory bodies about the policy, and so on; (Semi-structured) interviews with key players in the policy development process, both from the ministry and the agency, and if applicable from other interested parties; Optional: questionnaire (scale) to rate the level of influence or determine the importance of involved actors in the policy network. Such an instrument could help to substantiate conclusions from the more qualitative methods;
Optional; expert assessments of specific variables such as policy style and state tradition provided these variables are part of the research. A case study protocol has to be drawn up to ensure that similar questions and operationalizations/codes are used by the researchers in the different countries and exchange of data is possible for comparative analyses (Gerring, 2007; Yin, 1994). Instruments like the questionnaire (scale) could later be included in the Cobra survey methodology that many CRIPO countries have carried out or will replicate. LITERATURE Barzelay, M. (2007) 'Learning from Second-Hand Experience: Methodology for Extrapolation-Oriented Case Research', Governance, 20 (3), 521. Christensen, T. & P. Laegreid. (eds.). (2003). New Public Management: the transformation of ideas and practice. Aldershot, Ashgate Publishing Ltd. Osbourne, D. & T. Gaebler. (1992). Reinventing government: how the entrepreunerial spirit is transforming the public sector. Penguin Books Inc. USA. Gehring, T. (2004) 'The consequences of delegation to independent agencies: Separation of powers, discursive governance and the regulation of telecommunications in Germany', European Journal of Political Research, 43 (4), 677-698. George, A. L. and Bennett, A. (2005) Case studies and theory development in the social sciences. Cambridge, Mass.: MIT Press. Gerring, J. (2007) Case study research : principles and practices. Cambridge ; New York: Cambridge University Press. Hood, C. (2002). Control, bargains and cheating: the politics of public-service reform. Journal of Public Administration Research and Theory, 12(3), 309-332. Hood, C. and Lodge, M. (2006) The politics of public service bargains: reward, competency, loyalty - and blame. Oxford ; New York: Oxford University Press. Jacobsen, D.I. (2006). The relationship between politics and administration: the importance of contingency factors, formal structure, demography and time. Governance, 19(2), 303-323. Lijphart, A. (1971) 'Comparative Politics and the Comparative Method', The American Political Science Review, 65 (3), 682-693. Lowi, Theodore (1964) American Business, Public Policy, Case Studies and Political Theory. World Politics 16, 677 715. Mahoney, J. (2001) 'Beyond Correlational Analysis Recent Innovations in Theory and Method', Sociological Forum, 16 (3), 575.
Moynihan, D.P. & S.K. Pandey. (2006). Creating desirable organizational characteristics: how organizations create a focus on results and managerial authority. Public Management Review, 8(1), 119-140. OECD. (2002). Distributed public governance: agencies, authorities and other government bodies. Paris: OECD Press. Pollitt, C. & C. Talbot. (eds.). (2004). Unbundled government: a critical analysis of the global trend to agencies, quangos and contractualisation. London, Routledge. Pollitt, C., & Bouckaert, G. (2004, 2nd ed.). Public management reform: a comparative analysis. Oxford: Oxford University Press. Pollitt, C., C. Talbot, J. Caulfield & A. Smullen. (2004) Agencies : how governments do things through semi-autonomous organizations. Basingstoke, Palgrave Macmillan. Richardson, J. J. (1982) Policy styles in Western Europe. London, Boston: Allen & Unwin. Seawright, J. and Gerring, J. (2008) 'Case Selection Techniques in Case Study Research: A Menu of Qualitative and Quantitative Options', Political Research Quarterly, 61 (2), 294. Spence, D.B. (1997). Agency Policy Making and Political Control: Modeling Away the Delegation Problem. Journal of Public Administration Research and Theory, 7(2), 199-219. Svara, J. (2001). The myth of the dichotomy: complementarity of politics and administration in the past and future of public administration. Public Administration Review, 61(2), 176183. t Hart, P. & Wille, A. (2006). Ministers and top officials in the Dutch core executive: living together, growing apart? Public Administration, 84, 121-146. Van Thiel, S. (2008). The empty nest syndrome: Dutch ministries after the separation of policy and administration. Paper presented at IRSPM conference, 25-27 March 2008, Brisbane Australian. Veenswijk, M., & Hakvoort, J. L. M. (2002). Public-private transformations. Institutional shifts, cultural changes and altering identities: two case studies. Public Administration, 80 (3): pp. 543-555. Verschuere, B. (2009). The role of public agencies in the policy making process: rhetoric versus reality. Public Policy and Administration, 24(1), 22-46. Vibert, F. (2007). The rise of the unelected. Cambridge University Press. Wilson, J. Q. (1989) Bureaucracy : what government agencies do and why they do it. New York, Basic Books. Yin, R.K. (1994, 2 nd ed.). Case study research: design and methods. Thousand Oaks, SAGE..