The modern theory of economic policy emphasizes the importance of incentive constraints in the policy-making process, but most economic research, until recently, has neglected the political nature of some of these constraints. On the other hand, there is a large literature on the theory of public choice devoted to the analysis of collective decisions and political institutions, which has so far largely neglected the economic consequences of such policy choices.
Recent attempts to integrate these two fields of research, by analysing the interplay between political institutions and the formulation of economic policy formation, were the subject of a conference on `Economic Policy in Political Equilibrium', held by the Centre for Economic Policy Research and the Institute for International Economic Studies (IIES) in Stockholm on 14-16 June 1990. The conference was organized by Torsten Persson and Guido Tabellini, both Research Fellows in CEPR's International Macroeconomics programme. Financial support was provided by the Ford and Alfred P Sloan Foundations, as part of their support for CEPR's International Macroeconomics programme, and also by the Jan Wallander Foundation.
The conference was divided into two parts: the first was devoted primarily to theoretical issues of institution design and collective choice, and the second to specific policy issues in macroeconomics and public finance.
Political Institutions
Recent research on the theory of economic policy has stressed the value of the ability to make policy commitments. A main theme of the conference was the specification of the kind of institutions that enable society to commit itself to a particular future course of action. John Ferejohn (Hoover Institution) addressed this question in a paper entitled `Agency Policymaking with Presidential Veto', in which he first discussed whether and how the US Congress is able to increase its `commitment capacity' by delegating authority to an agency. When Congress delegates authority to an agency, it also delegates the task of monitoring that agency to a Congressional committee, which is usually granted both the right to make proposals that affect the agency (`proposal power') and also the right to prevent other members from making such proposals (`gatekeeping power'). If these two powers cannot be overruled by Congress, then the agency will be able to implement a policy that conflicts with the ex post preferences of the majority in Congress. In this sense delegation provides some commitment capacity, which may be strengthened further by the possibility of a Presidential veto.
Ferejohn then focused on the case when the agency has private information, and he noted that the possession of such information surprisingly may reduce the agency's leeway in enacting non-majoritarian policies. In Ferejohn's model, however, the commitment capacity `created' by the committees is entirely due to the pre-existing commitment by Congress not to overrule the gatekeeping authority of the committee without which the committee would be irrelevant. Thus in the end the problem remains: the delegation of authority to a committee cannot `create' a capacity to commit beyond that already in existence.
In a paper on `The Merchant Guild as a Nexus of Contracts', Avner Greif (Stanford University), Paul Milgrom (Stanford University) and Barry Weingast (Hoover Institution) addressed related issues concerning the design of institutions. They focused on the evolution of institutions to protect the property rights of merchants in medieval Europe, by analysing the relationship between a major trading centre and foreign merchants. A city aspiring to serve as a `pivot of trade' had to commit itself to a set of arrangements concerning security and trading privileges.
The authors argued that the value of reputation in a bilateral exchange between a merchant and a city was not sufficient to support the efficient level of trade, so that good behaviour could only be enforced if transgressions by the city were punished by a multilateral response of a large number of traders, which required coordination among the traders. They suggested that the merchant guild evolved in order to provide merchants with leadership to coordinate their reactions. The guild acted as a nexus of contracts a fictional entity with no interests of its own that made agreements and coordinated the behaviour of the merchants.
Methods of Collective Choice
Mats Persson (IIES) addressed the questions of how institutions can create commitment capacity and enforce property rights in his paper on `Political Democracy as a Solution to the Time Inconsistency Problem'. He used an overlapping generations model with capital and distortionary taxes, in which individuals live for three periods. A benevolent dictator faces the classic capital levy problem: that the optimal (second-best) tax policy is not credible, since at each point in time there will be the temptation to tax fully the capital already accumulated. If the tax policy is chosen under majority rule by all generations currently alive, on the other hand, then the capital levy problem disappears (or at least becomes less relevant). Under appropriate assumptions, the majority rule equilibrium will coincide with the policy preferred by the middle generation, which will never want to tax capital fully, since its capital holdings are positive. Hence welfare will in general be higher under majority voting than under a benevolent dictatorship.
In the next paper, `A Reconsideration of the Theory of the Firm: Property Rights and Dynamic Coalitions', Varadarajan Chari (Federal Reserve Bank of Minneapolis) and Edward Prescott (Federal Reserve Bank of Minneapolis and University of Minnesota) considered the enforcement of property rights within a firm, viewed as a coalition of agents. In their model, each such coalition has access to the same production technology, but it may choose to accumulate capital in different amounts. Capital accumulates in young coalition members, who will face the temptation when they are older to leave in order to start up new firms with their embodied capital. The behaviour of this economy will depend on which of the contracts between members of the coalition are enforceable.
Chari and Prescott found that the equilibrium outcome of consumption, labour and investment corresponding to the case in which all contracts are strictly enforced may be achieved even when all contracts are unenforceable. Such a `full-enforcement allocation' may be achieved if the old coalition members sell their right to control the firms to outside financiers for ever, but the efficient allocation can not be achieved if the workers control the firm.
Assar Lindbeck (IIES) and Jörgen Weibull (IIES and Princeton University) presented a paper on `A Model of Political Equilibrium in Representative Democracy'. They assumed that voters have preferences over policies as well as candidates, and the candidates have policy preferences as well as preferences for winning per se. As in the median voter model, in equilibrium the winning candidate is that preferred by a pivotal voter. Their model differs from the median voter model, however, since the pivotal voter is not the median in the space of policy actions, but rather the median in the more complicated space that also takes into account voters' preferences for candidates. Their model also differs from the conventional median voter model in that the two candidates do not converge to the same electoral platform. The equilibrium policy will lie between the preferred policy of the most popular candidate and the utilitarian optimum.
Fiscal Policy Implications
The remaining conference papers focused on specific problems of economic policy. Alex Cukierman (Princeton University), in a paper on `Social Security and the Deficit in the US: A Political Economy Approach', considered how the continued accumulation of surpluses in the US social security trust fund will affect future decisions on fiscal policy. He considered a model of `logrolling' in Congress, in which a larger social security trust fund changes the status quo, which in turn has two effects. It leads on the one hand to higher future social security benefits and lower social security taxes, and on the other hand to a reduction in the non-social security budget deficit. Cukierman argued that the first effect will prevail over the second, and hence that the continued growth of the trust fund will lead to increases in the overall budget deficits. In addition, increasing the size of the trust fund will induce a substitution away from payroll taxes into other taxes.
The creation of an internal European market in 1992 will increase the international mobility of resources and commodities. Torsten Persson (IIES and CEPR) and Guido Tabellini (University of California, Los Angeles, and CEPR) considered the effects of these developments on fiscal policy in different countries in a paper entitled `The Politics of 1992: Fiscal Policy and European Integration'. They first considered the taxation of capital in a two-country model, in which the tax policy of each country is chosen by a democratically elected government and the two countries behave non-cooperatively. In this case, increasing the mobility of capital will have two effects. First, it will lead to increased tax competition between the two countries, which in a symmetric equilibrium will in turn bring about a reduction in their tax rates. Second, increased capital mobility will also change voters' preferences and lead to the election of a different government. This purely `political' effect will offset the increase in tax competition, so that equilibrium tax rates will drop, but by less than they would in the absence of this political effect.
In the second part of their paper they considered the effects of increased mobility of final commodities on a labour income tax, under the assumption that labour remains internationally immobile. Again, increased mobility involves both economic and political effects, which lead in opposite directions. In this case, however, the political effect may prevail: thus the tendency for the change in the political equilibrium to preserve the status quo emerges as a general result.
Mervyn King (LSE and CEPR) and Mark Robson (LSE) presented a paper, `On Stochastic Tax Policy', in which they examined data on tax rates for the UK and the US, which go back to 1842 for the UK and to 1916 for the US. Their main finding was that tax rates tend to regress towards the mode, which is difficult to reconcile with theories of optimal `tax smoothing'. King and Robson concluded that the evidence supports the view that policy is made stochastically, i.e. that tax rates are inherently unpredictable and that tax policy reflects changing political preferences and ideologies. They also examined a number of econometric issues in the time-series modelling of variables bounded between zero and unity.
Growth Theories
The next two papers at the conference both focused on the modern theory of growth. Robert Barro (Harvard University) and Xavier Sala i Martin (Yale University), in a paper entitled `Public Finance in Models of Economic Growth', studied a class of growth models incorporating public services. They found that optimal tax policy hinges on the characteristics of these public services. If they are publicly-provided private goods, which are rival and excludable, or publicly-provided public goods, which are non- rival and non-excludable, then lump-sum taxation is optimal. On the other hand, if they are rival but non-excludable, then income taxation works approximately as a user fee and may therefore be superior to lump-sum taxation.
In principle, low-income countries should be able to achieve very rapid growth simply by copying what has been done in developed countries, but this rarely happens. In a paper on `Trade, Politics and Growth in a Small, Less Developed Economy', Paul Romer (University of Chicago) sought to explain this fundamental puzzle in the theory of economic development, on the premise that the different economic outcomes of different countries reflect different policy choices, especially with respect to integration with the rest of the world, which have significant redistributive consequences for different economic groups within the countries. His main result was that the trade policies most conducive to growth will harm the existing holders of domestic assets, and conversely, that the policies preferred by existing wealth holders (free intertemporal trade in consumption goods and restricted trade in capital goods) entail particularly bad consequences in terms of efficiency and income distribution.
Political Cycles
The last two papers of the conference focused on the `political business cycle'. In a paper entitled `How Often Should Elections be Held?', Kenneth Rogoff (University of California, Berkeley) applied a principal-agent model to the optimal timing of elections and the means by which they should be called. The voters are unanimous and act as the principal: the incumbent government acts as the agent, which has superior information regarding its own competence and is driven by its desire to win the elections. The political business cycle results from the efforts of the incumbent to signal its competence. The size of the cycle is determined by the electoral laws.
Rogoff noted that if elections are frequent, then the incumbent is more likely to be skilful and to act in accordance with the preferences of the voters. Frequent elections entail the cost, however, that political business cycles will occur more often, while allowing the incumbent to call early elections has the benefit of reducing the size of the electoral cycle. A result of the paper is that introducing randomized votes of confidence, in which there is a small probability that the government will be subjected to a vote of no confidence, will lead to a welfare improvement over the policy outcome that would obtain under any system of term elections.
Finally, in a paper on `Political Cycles: Evidence from OECD Countries', Alberto Alesina (Harvard University and CEPR) and Nouriel Roubini (Yale University and CEPR) reported the results from testing recent theories of the political business cycle on time-series data for 18 OECD countries for the period 1960-87. This is the first paper to carry out such tests systematically for such a large number of countries. Alesina and Roubini found that these recent models of the political business cycle in general significantly outperform their predecessors, and specifically that there is no evidence in favour of the early Nordhaus model in which voters have adaptive expectations. There is instead some weak evidence that an electoral cycle affects the inflation rate, which is consistent with a rational expectations version of that model. More importantly, there is strong evidence in favour of a rational expectations version of the `partisan theory' of business cycles, according to which policy-makers are elected by rational voters on the basis of their ideological preferences, which they pursue once they are in office. This theory is more strongly supported by the data for those countries with a bi-partisan political system.
A selection of the papers presented at this conference will appear in a forthcoming issue of the Review of Economic Studies.