Political economy models and agricultural policy formation : empirical applicability and relevance for the CAP

This study explores the relevance and applicability of political economy models for the explanation of agricultural policies. Part I (chapters 4-7) takes a general perspective and evaluates the empirical applicability of voting models and interest group models to agricultural policy formation in industrialised market economics. Part II (chapters 8-11) focuses on the empirical applicability of political economy models to agricultural policy formation and agricultural policy developments in the European Union. Part I is preceded by an overview of arguments and motives for government intervention in the agricultural economy in 'conventional' mainstream (agricultural) economics (chapter 2). A further delineation of research aims based on a state-of-the-art survey of the political economy of agricultural policies is given in chapter 3.Why a political economy approach?The economics literature is dominated by two - partly complementary - views on government intervention in agriculture (chapter 2). On the one hand, there is the farm problem view which focuses on the singularities of the agricultural sector and the specific characteristics of agricultural markets in particular. According to this view, the government should alleviate, and, if possible, solve market and income problems associated with this 'farm problem'. On the other hand, there is the market failure view. Market failure arguments not only refer to public goods such as food security, (guarantee of) food quality, agricultural R&D Or 'state insurance' against production failures and natural disasters, but also to external effects of agricultural production, such as erosion and pollution of air, soil and water.Both the farm problem and the market failure view implicitly regard the government as an omniscient benevolent dictator who has the capacity to take corrective action in an adequate and timely manner wherever necessary, with the overall objective of maximising social welfare. Both views, however, neglect the fact that government consists of multiple agents with partly parallel, partly conflicting motives, which often do not coincide with 'the public interest'. Furthermore, they ignore that these agents - which make up the policy supply side - are subject to lobbying and voting influences.Conventional econom(etr)ic models treat government behaviour in an ad hoc and usually exogenous fashion. This practice contrasts sharply with the meticulous attention for the micro-foundations of consumer and producer behaviour. In the (new) political economy or 'public choice' approach, however, government behaviour is the explicit object of analysis. The political economy approach regards policy as the endogenous outcome of a political supply and demand process in which farmers, industrial interests, consumers and other politico-economic actors exert policy demands, for instance by lobbying or going to the ballot box, and in which politicians and bureaucrats supply policies in exchange for political and budgetary support. The political economy perspective takes, like 'mainstream' neo- classical economics theory, methodological individualism as its point of departure. It builds on three crucial assumptions: actors behave rationally and are driven by self-interest, have stable preferences, and their interaction tends towards equilibrium states. The political economy approach provides the methodology for explaining why agricultural policies in industrialised market economies have almost universally taken the form of farm(ers') support and why policy instruments usually have an indirect and intransparent nature which is contrary to what one would expect on the basis of conventional welfare economics theory. Agricultural policies often are frequently inefficient, tend to under- or overshoot predefined targets, and diverge both with respect to the choice and the setting of policy instruments from the implicit recommendations of 'mainstream' neo-classical economics theory.Empirical applicability of political economy modelsThe political economy approach supplies us with the theoretical ammunition for explaining agricultural policy. First generation 'public choice' theories have provided us with plausible arguments for why agricultural policies have taken the form and content they have, but have largely ignored the issue of empirical testing. The first generation analyses build on the pathbreaking studies of the founding fathers of public choice, such as Olson, Downs, Buchanan and Niskanen, and 'explain' agricultural policy either on the basis of qualitative arguments derived from 'theoretical' public choice, or by means of statistical regression analysis. In the latter case, the relationship between politico-economic theory and the choice of the empirical model is usually a loose one, featuring ad hoc-specified models without proper politico-economic micro-foundations. The second generation public choice models are - more than their predecessors - characterised by a strong theoretical- deductive and mathematical basis. Furthermore, these models more explicitly try to establish a link between the politico-economic activities of (groups of) actors and the final policy outcome. The label empirical applicability may refer to the ability to empirically estimate politico-economic relationships, with policy as the endogenous variable. Empirical applicability can also refer to empirical testing of hypotheses derived from a consistent political economy model. It does not apply, however, to the ad hoc manner in which some model studies refer to 'real world' examples to support the outcomes of theoretical-deductive models. Part I evaluates the assumptions, methodology and limitations of existing political economy models from the perspective of empirical applicability.Voting modelsThe class of voting models - which take the interaction between politicians and voters as their focus - includes spatial voting models, proportional voting models and political business cycle models (chapter 4). At first glance proportional voting models appear to offer the greatest explanatory power. The central thesis of these models is that politicians strive for political support and exploit the differences in welfare impact of income redistributional programmes supplied to specific voter groups. In most industrialised market economies the agricultural population consists of a small and declining minority of the total population which does not yield sufficient votes to win the elections. By using the more inclusive notion 'political support' the proportional voting approach tries to circumvent this problem. An important drawback is that it does not explicitly incorporate the activities that make up this political support. Moreover, empirical operationalisation of the political support concept appears difficult. Finally, by its focus on a single policy domain the effect of other policy issues and their impact on the behaviour of political-support-maximising politicians remains unexplained. The spatial voting model - also known as the median voter model - can be regarded as second-best alternative. Yet the median voter model is too restrictive for a useful empirical application on agricultural policy formation. This is not only due the fact that agricultural policy cannot be ranked on a unidimensional left-right scale, but also because more 'realistic' theoretical extensions of the standard model, such as the inclusion of multiple issues or extension of the number of parties or candidates to three or more, incessibly lead to instable equilibria or do not yield any equilibrium at all. In a more- than-two party system strategic behaviour will tend to dominate, which causes that the central assumption of sincere voting behaviour - which holds that only policy preferences determine voters' choices - cannot be upheld. Furthermore, theoretical arguments suggest that the spatial voting model is not suitable for explaining redistributional issues. It forms another major objection against the use of the spatial voting model, since a substantive part of agricultural policies is explicitly aimed at redistribution. Political business cycle models may be illuminating in explaining macroeconomic developments and more specifically the political trade-off between inflation and employment. They do not, however, yield a workable alternative for the explanation of sector policies.A fundamental problem that jeopardizes the empirical application of voting models concerns the operationalisation of the voter group concept. Potential pitfalls involved in the use of analytical distinctions, for instance between the urban and rural population, can be illustrated with the notion of the 'farm vote' (chapter 5). In both the interpretativedescriptive political science and the public choice literature, the farm vote is used to indicate the political strength of farmers. Some equate the concept with the labour force engaged in primary - first-stage - agriculture or the 'agricultural population'. Others use an even broader farm vote capturing the rural population or even those people who "appreciate the values and attributes of farm life and sympathize with the agricultural profession", including pensioners with a professional agricultural background, people borne on farms, and people with a (vague) idealistic notion of agriculture and farm life (see, e.g., Moyer and Josling, 1990). However, it is doubtful whether the preferences of these voters for agriculture/ agricultural policy are decisive in their electoral behaviour. Furthermore, the implicit assumption that the farm vote is homogeneous can be judged as weak. Large differences in initial income and wealth positions between farmers, which include differences in production factor ownership such as land, buildings, livestock and machinery, imply that policy preferences may differ strongly between farmers. The systematic evidence for a relative strong party attachment of farmers as an occupational group and their preference for right-wing parties frequently found in public opinion surveys, does not help us in establishing a clear empirical notion of the farm vote, since in these studies the underlying reason for these party preferences usually remains unexplained.Interest group modelsThe 'demasque' of the farm vote concept underlines the importance of organisation in attempts to influence government and policy formation. One of the results of the political economy of (the organisation of) collective action is that large groups do not easily engage in collective action. Individualised selective incentives can help to overcome existing dominant free-rider behaviour and can encourage individuals to act in a group-oriented way, as Olson (1965) already pointed out. In smaller, so-called privileged, groups organisation is usually not a necessary condition to forestall the provision of public goods (chapter 6). Olson did not explicitly address the question how the public (group) good was to be secured, and which role could be assigned to the government in this respect. Studies in the rent-seeking and DUP, but also the Chicago political economy tradition have more explicitly focused on the attempts of interest groups to manipulate government, with 'government capture' as reflected by the creation of (quasi)monopoly positions and/or indirect income transfers as a result. With respect to the form of policy instruments Chicago political economists assume that political competition ensures that the most efficient form of redistribution is chosen ('what is, is efficient'). Virginian political economists, on the other hand, perceive policy instrument choices as being determined by information characteristics, with politicians having a preference for indirect, inefficient instruments to secure voter groups.The way in which groups attempt to establish political influence can serve as a basis for classifying interest group models (chapter 7). This transmission mechanism is most explicit in models in which the transmission of information, respectively political campaign contributions play a central role. In both types of models a clear traceable relationship can be distinguished between the political activities of groups and the policy decision and/or policy outcome, assuming that other influences are absent. Although these models are preferable from a theoretical perspective to the two other alternative model classes - which take an influence function, respectively a composite utility function as their pivot-, the reverse applies if we look at the issue of empirical applicability. Interest group models which centre on campaign contributions obviously are relevant only for those countries in which politicians/political parties can actually be supported by interest groups. For useful empirical applications of such models adequate registration of interest groups and their political expenses is a prerequisite. For most European countries as well as the European Union such data are not available. Models which take the transmission of information as their focus are also difficult to apply empirically. Many, if not all, policy influence can be traced back to the transmission of information. However, the registration of these informational flows and the ex post determination whether this information could have played a crucial role in 'real-world' policy decisions/outcomes has to be judged infeasible. Models that take an influence function, respectively a composite utility function as their point of departure offer better opportunities for incorporating empirics. There clearly is, however, a trade-off between the more explicit inclusion of political activities and the impact on the eventual policy outcome. Influence function models provide clearer microfoundations for the competition-for-influence process among interest groups and incorporate the resource costs needed for influence activities. Yet, the question which influence activities are involved and how these are related is not addressed. Models with a composite utility function focus on the determination of politico-economic (power) weights implicitly assigned to socio-economic groups such as consumers, producers and taxpayers (in the political preference function [PPF] approach) or social classes (in the interest function approach) and measured by means of revealed preference methodology. The PPF approach is by far the most popular approach in agricultural economics. Its application to the political feasibility of (discrete) changes in the policy instrument mix is, however, contestable (see chapter 7). Where the <em>development</em> of agricultural policies is concerned the combined PPF - influence function approach in the tradition of Gardner (1987) is most apt from an empirical point of view. Yet, most of these models focus on price developments or the degree of nominal (effective) protection policy issues in specific agricultural markets, and leave the overall explanation of agricultural policies and, most importantly, the issue of instrument choice over time untouched.Political economy models and the explanation of EU agricultural policyPart II focuses on the empirical applicability of political economy models to agricultural policies within the European Union. This applicability depends, apart from the intrinsic characteristics of these models as described and analysed in part I on the politicalinstitutional context and the decision- making structure of the European Union. Understanding the relevant decision-making characteristics would enable a well-founded choice among the available political economy models. The sui <em>generis</em> characteristics of supranational decision-making and its institutions are described in chapters 9 and 10. An overview of stylised facts and figures of 'real-world' agricultural policy developments is given in chapter 8. An evaluation of the applicability of political economy models on agricultural policy formation of the European Union is provided in chapter 11.The descriptive analysis of the empirics of agricultural policy developments in the European Union since the late 1950s predominantly focuses on the changes in policy instrument choice. The period between the beginning of the 1970s until the mid-1980s shows a tendency toward renationalisation of the Common Agricultural Policy. Not explicitly stated as a goal, renationalisation manifested itself through the use of Monetary Compensatory Amounts and non-tariff barriers. Moreover, since the CAP's inception member states have kept some national policy competence in the field of agriculture. These 'national' agricultural policies are to a large extent complementary to Community policy, and predominantly aim at the provision of (quasi)public goods such as agricultural infrastructure, education, research and extension. Since 1962 the CAP has been subject to pressures for change, from within - often initiated by the European Commission - as well as from outside, for example as the result of international trade negotiations. The high degree of border protection, the relatively high common price level and the increases in agricultural production have led to increasing surplus problems, a high budget burden and a distortion of trade relations with third countries. It is from this perspective that Most CAP adjustments have to be judged: the introduction of co-responsibility levies, guarantee thresholds, quotas, budget stabilisers, set-aside and extensification measures, as well as the later MaeSharry reform.The identification of politico-economic actors, their behaviour (as optimizing actors under constraints) and the way in which they interact are fundamental requirements for an endogenous policy model. Chapter 9 addresses the formal decision-making procedures as laid down in the Treaties of Rome and Maastricht and highlights the role of the Commission, the Council and the European Parliament. The annual price review is taken as an example. Although the MacSharry reform has changed the nature of the annual price review, with a relative shift in attention for institutional prices, direct intervention and export refunds toward direct income transfers, it can still be regarded as the most important yearly recurring 'package deal' process within the CAP.What strikes the eye is the marginal influence of the European Parliament in EC agricultural decision-making and the lack of parliamentary scrutiny and review of ECdecisions at the national level. The 'democratic deficit' at the EU-level combined with the relatively modest size of the EU's administrative apparatus provide ample opportunities for policy influence by interest groups (chapter 10). For the Commission the importance of interest groups lies first and foremost in the possibility of the ex ante sounding out of (the acceptability of) new policies and in transmitting the information needed in the design of complex regulatory policies. Since the start of the CAP "Brussels" has stimulated the formation of many Euro-groups with an EC-wide member base. Apart from informal, ad hoc contacts with Community bodies, a large number of these Euro-groups also has a formal role in Ec agricultural decision-making as part of the advisory committee structures set up by the Commission. Interest groups organised at the national level try to influence EC-agricultural policy-making as well. Contrary to Euro-groups, however, these 'national' groups mostly concentrate their lobbying efforts on member state governments and ministries.Political economy models and the explanation of EU agricultural policiesThe extent to which political economy models can be empirically applied to and offer an explanation for the formation and developments in agricultural policies in the European Union is the central theme of chapter 11. One important conclusion is that the CAP should not be analysed in isolation, but in connection with the 'national' agricultural policies of the individual member states. Ignoring this strategically important national policy dimension in policy analyses could result in biased and flawed interpretations of (the causes of) policy developments. Integrating both policy levels, which can yield meaningful results on the interaction between Community and member state policies and which can establish a traceable relationship between policy outcomes on the one hand, and the motives, ends and politico-economic activities of the actors involved in EU agricultural policy formation on the other, seems too high an aim. A flexible, all-inclusive and testable model which can explain agricultural policy developments at both levels empirically is not available and could not be developed in the context of this research project.If we restrict the focus to the applicability of political economy models to the CAP, and take the relatively large (potential) influence of interest groups and the limited influence powers of the European Parliament on EC agricultural decision-making into account, interest group models appear the most obvious choice of the models analysed in part 1.However, in establishing an explicit and traceable relationship between the politicoeconomic activities of policy agents and the eventual policy result, as well as in the explanation of changes in the policy instrument set, even the better empirically applicable models still fell short. This does not detract from the value of political economy models as such. However, their strength has more to sought in signalling and in their ability to trace down <em>possible</em> determinants behind policy changes rather than in 'hard' empirical testing and unequivocal explanation of policy developments.