Applications of economic theories of politics

In this section we consider three different economic approaches to politics: public choice, economic analysis of policy, and economic analysis of insti­tutions. These are not the only approaches that could have been examined.

Public choice theory

At the broadest level, public choice theory involves the application of eco­nomic methods to politics.

[It] takes the tools and methods that have been developed to quite sophisticated analytical levels in economic theory and applies these tools and methods to the political or governmental sector, to politics, to the public economy. (Buchanan, 1984:13)

While this definition may seem straightforward, the transference of eco­nomic methods from the economy to the polity involves some complications. These complications center on the aggregation of individual preferences into collective, or “public” outcomes, the problem of coordinating individual interests and choices to achieve collective outcomes, and the interdependence of individual decisions. These three problems have been dealt with in the literatures on voting rules, collective action theory, and strategic (or game) theory.

Public choice theory is relatively new as economic theories go, deriving in part from the literature on public finance during the 1950s (Musgrave and Peacock, 1958; Musgrave, 1959), and in part from the seminal contributions of Kenneth Arrow’s Social Choice and Individual Values (1951), Anthony Downs’s An Economic Theory of Democracy (1957), and James Buchanan and Gordon Tullock’s The Calculus of Consent (1962). Mancur Olson’s The Logic of Collective Action (1965) helped to place the public choice work of economists in front of political scientists and to facilitate the conceptualization of many standard preoccupations of political science (including interest group orga­nization, bureaucratic behavior, the organizational bases of influence, alli­ances) as problems of collective action.

The contributions of public choice are easily chronicled since there is a journal (Public Choice) and a professional society (The Public Choice Society). While the literature on public choice theory is not limited to the journal, the existence of a publication outlet, a society, and professional meetings has served to consolidate and give focus to what otherwise might be a fragmentary research program. The Public Choice Society took shape during the mid- 1960s, was headquartered at the University of Virginia, and was initially called the Committee on Non-Market Decision-Making (Tollison, 1984:3). In 1966 it published a series called Papers on Non-Market Decision-Making. In 1968 the group changed its name to Public Choice Society and the journal title to Public Choice. Still, the original labels are instructive since they imply that economics as subject matter is concerned with markets and decision processes that are individual, while politics is concerned with processes (or outcomes) that are collective.

Let us see if we can sharpen the previous definition of public choice as simply the application of economic methods to politics. It is after all possible to apply economic methods to political problems that are not public, at least not to any significant degree. For example, one could analyze the efficiency properties of a redistributional program.^[20] In most standard works, public choice is defined as the application of economic methods to politics (Mueller, 1979:1; Buchanan, 1984:13; Ekelund and Tollison, 1986:440) without linking these methods to outcomes that are in some sense public (for an overview, see Plott, 1976). In this regard the treatments by Stiglitz (1988:145ff.) and McLean (1987:9-11) are more helpful. Nevertheless, the public choice lit­erature makes it clear that the relationship between public choice and public goods is a close one. If outcomes were not characterized by externalities, private decisions would do. The focus remains that of the choosing individual - his or her preferences and maximizing behavior.

Public choice theory sees individual actors as central, whether operating as members of political parties, interest groups, or bureaucracies, whether elected or appointed, whether ordinary citizen or chief executive:

The fundamental premise of public choice is that political decision-makers (voters, politicians, bureaucrats) and private decision-makers (consumers, brokers, producers) behave in a similar way: they all follow the dictates of rational self-interest. In fact, political and economic decision-makers are often one and the same person - consumer and voter. The individual who buys the family groceries is the same individual who votes in an election. (Ekelund and Tollison, 1986:440)

Public choice theory is different from conventional economics not in its conception of the individual and the forces motivating action but rather in the different constraints and opportunities offered by the political as opposed to the market environment. In this new theory, economics (as market ex­change, production, consumption) and politics (as political exchange, power, authority relations) appear as special applications rather than as distinctive subject matters. Politics simply refers to those institutions and processes through which individuals pursue their preferences when these preferences refer to goods that are interdependent or public.

In the remainder of this section we will discuss two strands of public choice theory: normative and positive. The former often has to do with issues related to political design, basic political rules, in short, with the constitutional framework within which political processes take place. The work of Arrow (1951) and Sen (1970) exemplifies the normative strand. The positive strand has to do with attempts to explain observable political behavior in choice theoretic terms.

Normative public choice. Normative public choice deals with the analysis of desirable properties of the political system. What sorts of institutional ar­rangements are efficient, responsive, and fair? What types of voting rules most faithfully translate (aggregate) individual preferences into public deci­sions? What governmental structure is likely to prevent concentrations of power, stagnation, or instability? Will a federal or unitary system be better able to “contain” ethnic, class, and religious differences? These are repre­sentative questions consistent with a public choice viewpoint.

Buchanan and Arrow provide two examples of normative public choice. James Buchanan is interested in organizing society so as to increase the scope for free exchange, whether within economic (market) or political (state) set­tings. Within political settings he distinguishes between constitutional and postconstitutional politics. However, a strong contractarian position is main­tained in both. The political system is judged desirable to the extent that it facilitates voluntary exchange and proportionate relations between private costs and publicly provided (but privately consumed) benefits (Buchanan, 1987, 1988). Kenneth Arrow, in Social Choice and Individual Values (1951), deals with voting rules that consistently translate individual preference or­derings into group decisions. We will go through this example in more detail.

In 1951 Kenneth Arrow published his Social Choice and Individual Values, a book that was to stimulate wide interest in public choice. The basic problem was simple: In a representative democracy, where individuals vote or oth­erwise register their preferences in collective outcomes, how can those in­dividual preferences be consistently aggregated so as to produce collective decisions? The word “consistently” refers to the same transitivity condition necessary for rationality at the individual level.

What Arrow discovered (or really rediscovered, the results having been known to deBorda and Condorcet in the eighteenth century) was that indi­vidual preference orderings do not in general “add up” to a consistent social preference ordering (or social welfare function). The group literally cannot make up its collective mind (Buchanan, 1984:17). Yet, as Frolich and Op­penheimer point out, for the assumption of individual rationality to be of value in explaining political outcomes, which are perforce collective, there must be a consistent link between individual and group orderings (1978:15). This is precisely what Arrow found did not, given his assumptions, exist.

Let us run through a stylized example. Suppose we have three individuals and three choices of candidates or policies. See Figure 4, where the private preferences of voters 1, 2, 3 are given for candidates or issues A, B, and C.

Now let us try to aggregate these individual orderings by having a pairwise vote among alternatives. If A is compared to B, it dominates (beats) B. Voters

1 and 2 both prefer A to B. Comparing B against C, B dominates. If A dominates B and B dominates C, A should dominate C. But this is precisely what we don’t find. Instead, C is preferred to A by voters 2 and 3. Individually consistent preference orderings do not define a consistent ordering at the group level.

Arrow’s result, labeled a general impossibility theorem, has met with di­verse reactions. Some treat it as irrelevant (see Tullock’s “The General Ir­relevance of the General Impossibility Theorem,” 1967) while others argue that it attacks the very logical foundations of democracy.

What Arrow seems to be saying is not only that without restrictions on the form of individual preference orderings democracies are not democratic, but that they can not be, at least in the sense of having voting rules for consistently translating private wants into collective decisions. A democratic procedure that produces transitive, yet nonarbitrary group choices is impos­sible, given Arrow’s five reasonable assumptions.^[21]

Candidates/Issues

Figure 4.

Arrow’s own solution to the problem requires us to give up the assumption of full exogeneity and autonomy of individual preferences.

[I]t must be demanded that there be some sort of consensus on the ends of society, or no social welfare function can be formed. (1951:83)

But how can we secure such a consensus without assuming that preferences are in important ways formed so as to serve that consensus? If preferences are thus shaped, they are in this sense endogenous.

Since 1951 much research, normative-formal as well as empirical, has been inspired by Arrow’s problem. Arrow’s theorem implies that democracies must either be unstable and experience perpetual cycling of majorities or else be characterized by arbitrarily induced equilibrium. The choices are not attrac­tive. It is not surprising that much follow-up work was stimulated. This research attempts to find a way out of the paradox by exploring alternative voting arrangements. Some of this research only confirms and deepens Ar­rows’s proof, as in McKelvey’s “Chaos Theorem” (McKelvey, 1976, 1979), which demonstrates that unless preferences are perfectly symmetrical around

aggregation, the independence of irrelevant alternatives, citizen sovereignty, and nondicta­torship. They seem quite reasonable, indeed even weak, in that they only made explicit what most had always assumed in democratic theory. For a thorough discussion of these assump­tions, see Frolich and Oppenheimer (1978:19-23).

a median point, there is a global cycle involving all choices. This result has proved to be durable and has been expressed in game theoretical terms by Schofield (1978), who shows that multidimensional voting games have empty cores, where “core” refers to the set of possible equilibrium strategies (see McLean, 1987:186). Some of this work attempts to explore the properties of majority rule with the goal being the establishment of whether or not majority rule fulfills certain normative criteria (see Rae, 1969; Taylor, 1969).

Additional work has proceeded by attempting to relax Arrow’s already weak assumptions by allowing restrictions on preferences, varying intensities of preferences, and control over the agenda. While we will not discuss this research in detail here (see Mueller, 1979:ch. 3 for a discussion), a brief word is in order. Arrow’s assumption that any possible array of preferences is allowable is relaxed so as to consider particular distributions of preferences, such as those which are single peaked (a point from which all other points are lower). Single-peaked distributions can result in majority equilibrium, as Slutsky (1977) has shown. Allowing for varying intensities of preferences encourages logrolling, opening up possibilities for solutions not available with preferences ordinally defined. Finally, agenda control suggests a possible solution to the cycling problem although it is not clear how this solution is nonarbitrary.

Positive public choice theory. While normative public choice deals with de­sirable characteristics of the rules, procedures, and institutions through which collective choices are made, positive public choice attempts to devise expla­nations for these rules, choice processes, and their consequences. The fol­lowing questions are representative: Why and how do people establish laws, create political institutions, join groups, and vote? What factors figure in the formation and influence of groups? What conditions conduce to successful cooperation among members of cartels or classes. How (and when, under what conditions) do nation states make decisions about the provision of international public goods? What accounts for the behavior of bureaucrats, legislators, and lobbyists? Although not unrelated to normative issues, these questions are different from those concerning the equilibrium properties of majority voting procedures and whether it is possible to construct a consistent social welfare function from the ground up.

Thus, while normative public choice poses questions about how we might organize political life so that outcomes best express private (utilitarian) self­interest, positive public choice theory goes further. It assumes that citizens do in fact act on the basis of self-interest (in the economist’s sense) so that actual political outcomes can be explained on that basis.

Much of the research on positive public choice has been inspired by Olson’s Logic of Collective Action (1965). This research ranges from the experimental and simulation approaches of Axelrod (1981, 1984) to historical case studies on rational peasants by Popkin (1979), Taylor’s work (1988) on revolution as a collective action problem, and Bowman’s treatment (1989) of cooperation among capitalists from the same vantage point. Similarly, we can include much of the international relations literature, especially that addressed to the problem of public goods provision at the international level (Keohane, 1982, 1984; Gilpin, 1987; Gowa, 1989).

To illustrate this approach, we briefly discuss the phenomena of voting (Downs) and interest group organization (Olson). First let us turn to Anthony Downs’s treatment of voting in An Economic Theory of Democracy (1957). Voting leads to a single outcome (electoral victory or defeat) for all, a single party in power, and a single set of policies. Electoral outcomes have some of the characteristics of public goods, most importantly the difficulty of excluding anyone from benefits and costs associated with the winning coa­lition. If there are any costs associated with participation, there will be an incentive to free ride.

Downs’s rational analysis of voting recasts democratic political theory in economic terms. The politician is the supplier of policies and governmental services. The voter is the consumer, using his or her votes as dollars to express political demands. The politician trades services for political support. The voter trades votes for publicly provided services. It is easy to see how political advertising, campaign financing, and media consulting enter the picture. The resulting image is one of a political process where individuals, although occupying different political roles, are nevertheless motivated by self-interest, and are ready to engage in exchange to further that self-interest. The importance of ideology (for instance, belief in the New Deal) and tra­ditional forces (such as party identification) declines in favor of interest and perceptions of utility associated with one party or another, one policy or another.

How does the voting model work in Downs’s theory? There are two political parties in the model. Presumably this results from an institutional specifi­cation given outside the model. For example, a rule requiring single-member districts, with a plurality winner-take-all vote, would all but ensure no more than two parties. Voters in the electorate are arrayed along a unidimensional, liberal-conservative, continuum. Note that this avoids the Arrow problem, the distribution of voters being single peaked. Parties attempt to formulate packages of policies (programs) that appeal to a majority of voters. Parties do not offer their programs because they believe in them or have attachment to their contents. They offer them instrumentally, as a way of attaining political office.

With the preceding conditions in place, Downs is able to show that both political parties will move to the center of the continuum, toward the median

voter. Any other strategy would not be rational, for it would not be designed to win the election. Figure 5 illustrates this point.

In this figure the voters are arranged along a continuum from most con­servative to most liberal. The median voter lies at point X. Exactly half the voters are more conservative; half are more liberal. If party A gears its program for the more liberal voter, say at X + 2, all party B has to do is to aim for all voters to the left of X + 2 and it is assured of victory. The median voter identifies the reference point for victory. Voters X + 1 (and everyone else to the left of him or her) defines a minimal winning coalition.

Underlying this prediction that parties will move toward the median voter is a conception of the individual voter and the calculations underlying the vote. There are three components to this calculation. The first is the difference in utility deriving from one or the other party being in power. The second is the probability of affecting this result - that is, the probability of making a difference. And third are the rewards associated with participating in the democratic process. Downs reasoned that the voters care about elections to the extent that they care who wins or loses. Thus a critical variable is the difference to a voter in whether one party or another wins. This difference may be high without implying much motivation to vote. The individual may calculate that the probability of affecting the outcome is very low. In effect, this probability is equal to the chances of being the critical member (the decisive member) of a minimal winning coalition. If the electorate is large or the election is forecasted to be a landslide, the probability of being the decisive member will be very small. The conditions under which an individual will count (close elections) seem to be precisely those where differences be­tween parties reduce to zero. If parties behave rationally, they will adopt policies aimed at the median voter.

There is an unmistakable irony built into the way the utility of the party differential and probability of affecting the election result compete with each other. If the voter thinks it makes a difference who wins, the election cannot be close, since at least one of the parties must not be appealing to the median voter. On the other hand, if the parties are very close to each other in terms of the programs offered, they should also be close in electoral support, mean­ing that the probability of affecting the outcome is higher.

Another important area of positive public choice involves the analysis of interest groups. In The Logic of Collective Action (1965), Mancur Olson brought the ideas of public goods and the problem of collective action directly to the attention of political scientists. He did this by trying to show that phenomena central to politics - interest group organization and influence - were public goods.

In making his argument, Olson vigorously attacked both pluralism and Marxism for ignoring the collective action problem for groups and classes respectively. Pluralists believed, almost without question, that interest group organization was the natural expression of collectively held interests. Some argued that Marxists believed that the transition from objectively shared interests to class organization and mobilization was spontaneous. Olson saw things differently. He not only questioned the automatic nature of the tran­sition from private interest to group organization but seemed to turn the matter upside down by arguing that it is not rational for individuals to contribute toward pursuit of collective interests. Democracy had fallen on hard times. Arrow questioned the logical foundations of democracy through his analysis of voting rules. Olson attacked the rationality of interest group organization, the primary mechanism for translating preferences into policy between elections.

But it is not in fact true that the idea that groups will act in their self-interest follows logically from the premise of rational and self-interested behavior. It does not follow, because all of the individuals in a group would gain if they achieved their group objective, that they would act to achieve that objective, even if they were all rational and self-interested. Indeed, unless the number of individuals in a group is quite small, or unless there is coercion or some other special device to make individuals act in their common interest, rational self-interested individuals will not act to achieve their common or group interests. (1965:1-2, italics in original)

Olson laid down the challenge to pluralism in direct terms. Some interests are never organized, some groups forever “latent,” some classes forever “classes in themselves” rather than “classes for themselves.” And unlike the critics of pluralism such as Schattschneider in The Semisovereign People (1960), who chastised pluralists for overlooking the role of wealth and re­sources in fostering group organization, Olson seemed to be saying that the forces shaping group organization lay elsewhere.

Since 1965 there have been numerous studies of interest groups from an economic and public choice perspective. However, the basic logic of the argument remains the same. The benefits of group organization and influence are public; the costs of excluding anyone from those benefits are high. Once the benefits are present for anyone in a group, they are available for everyone. Thus, group organization and group activities take on the characteristic prop­erties of public goods: They are undersupplied and subject to free riding.

From a microeconomic standpoint it is the calculations of individuals that are important. An individual i in a large group would reason in the following way if he or she were rational, well-informed, and self-interested. Individual i is only one member of a large group. Any effort for the provision of public goods is privately borne, whereas benefits are jointly consumed. Furthermore, the public good (interest group organization) may be provided independently of i’s efforts. The larger the group, the greater the probability that i’s efforts will be irrelevant.

In explaining why interest groups do emerge and do provide collective benefits, Olson does not appeal to individual irrationality, collective purposes, or social norms encouraging cooperation. Instead, he tries to identify expla­nations consistent with his view of individual self-interest. Individuals may participate in groups because they are coerced, because they are provided with selective incentives, or because they constitute a “privileged group.”

The case of coercion is not of great interest for a theory based on voluntary choice. However, Olson stresses the importance of selective incentives - that is, payments or rewards to participants beyond the public goods benefits. Selective incentives “are benefits that can be conferred upon contributors and withheld from noncontributors - insurance policies, news publications, discounts on goods and services - anything at all that individuals may value and that can be selectively conferred or withheld” (Moe, 1980:4).

A group is said to be “privileged” if there is some subset of the whole group for whom benefits exceed costs regardless of what others in the group do. If it is to the advantage of the United States to cease emitting chloro­fluorocarbons into the atmosphere so as to protect the ozone layer, regardless of what all other countries do, we speak of the group as “privileged.” The existence of privileged groups is important for Olson’s theory. From this idea Olson argues that large groups are less likely to succeed than small ones and that concentrated groups - that is, groups with a few large members - are more likely to succeed than dispersed groups. If Olson is right, it would mean that hegemonic distribution of international power would be associated with international public goods (free trade, monetary stability) as Kindle- berger (1973, 1981) and Gilpin (1981, 1987) have argued. It also suggests that labor has more organizational difficulties than capital, consumers more than producers, small capital more than large, concentrated capital, and the numerous small countries in the United Nations Conference on Trade and Development (UNCTAD) more than the smaller organizations.

The Logic of Collective Action provides an economic theory of interest group organization and influence. As such, it is an economic theory of politics. It stresses the rational, self-interested bases of interest groups and tries to devise a theory of what group life would look like, including what groups would not exist except as aggregates of shared interest, if it conformed to the calculus of individual interest. Olson’s attempt to identify the conditions that are conducive to cooperation among individuals has stimulated much research among political scientists and sociologists.

The economic analysis of policy

In a sense the economic analysis of policy is an activity that begins after the tasks of public choice analysis are either accomplished or assumed. Public choice theory concerns itself with the method of combining individual pref­erences into social welfare functions and public choices. This section assumes that preferences have already been combined, communicated to decision makers, and the task is now for decision makers to choose among alternative policies in such a way as to maximize satisfaction (of those affected by the policy - not just the decision maker). Thus, while there is a close relationship between the economic analysis of policy and public choice, the focus of this section is different.

When approaching policy choices, what kinds of questions does the econ­omist ask? What orientations are assumed? What yardsticks are used? The neoclassical approach to economic policy should help answer three broad types of questions. These questions concern the proper sphere of government, the constitutive principles of government, and the best (the most efficient) way to achieve collective goals. While most of our attention will be directed toward the third item, efficient choice, a word about the first two may be in order.

What is the proper sphere of political activity or, put differently, what is the appropriate dividing line between state and market? This is a question asked long ago by Adam Smith (1776) and by Herbert Spencer (1843). Indeed, Spencer’s essay is titled “The Proper Sphere of Government.” To answer this question, economists start with the individual and the market. Inside the market, individuals engage in countless exchange activities, ac­quiring and giving up goods, services, and productive factors. It is a fundamental theorem of neoclassical welfare economics that if markets are perfectly competitive, an equilibrium set of prices will exist that will allow all welfare-improving exchanges to take place. In such an equilibrium no one can improve his or her position without damaging the position of someone else. In a market society, government will be limited to doing what the market cannot do, or at least cannot do well, in other words, to cases of market failure. Market power may include the specification and enforcement of property rights, the supply and organization of military power, public goods and externalities, and the control of concentrations of economic power (Rhoads, 1985:66).

The simplest statement of the demarcation is this: The market allocates whenever it can do so efficiently; when it fails, government enters and pro­vides through political action what could not be supplied through private initiative. This simple formula was for some time accepted, but a number of economists (McKean, 1958; McKenzie and Tullock, 1981) have pointed out that the theory of market failure must be accompanied by, and compared with, a theory of political (or governmental) failure. Now, the dividing line is not so clear. A demonstration that the market cannot provide certain goods, or cannot provide them efficiently, does not constitute proof that government can do better. Governmental action consumes resources that, presumably, have alternate uses. Interest group activities, lobbying, legislation, enforce­ment, regulation, and adjudication are all costly and must be compared against the resulting benefits.

The second matter concerns how government should be organized. There is an economic perspective on questions relating to basic political structure. The first principle is that government should be organized so as to maximize exchange - that is, to expand the sphere of voluntary private exchange even within government. While governmental activities must at some point involve indivisibilities, which infringe upon individual choices, the effort should be to disaggregate the political process, to “unpack” more comprehensive al­ternatives, and to write laws in such a way as to allow for coercion only when indivisible benefits and free riding are at issue. Thus, specific activities such as vote trading, logrolling, and tying tax payments to capturable political benefits, would be encouraged on the grounds that they free up Pareto- improving exchanges and allow intensities of preference to register in the political process.

The second principle is the Wicksellian unanimity (or near-unanimity) principle, which is the Pareto-optimal idea carried over to the political realm. In 1896 Knut Wicksell wrote a work that has come to be popular among advocates of the contractarian state and limited government. In it he at­tempted to devise ways in which democratic governments could organize themselves to implement policies in a consensual way. Wicksell tried to introduce concrete voting arrangements that would assure the minimum of coercion within government. Specifically, on issues related to spending and taxation, a simple majority in favor of new proposals was not adequate. Wicksell proposed 75 percent to 90 percent agreement among political rep­resentatives and urged representatives to withhold tax payments from their constituencies if the exchange was not deemed beneficial (Wicksell, 1896; Wagner, 1989:210).

Finally, once basic political structures are in place, economists can address particular policies. As Stokey and Zeckhauser point out (1978:22), economists are likely to approach policy issues with the same questions asked of private choice: What is it we want? What is it that we can get? The usual conditions are assumed: exogenous preferences, scarce resources, alternative paths of actions, beliefs about the relationships between alternatives and outcomes, and so on. The key differences are that the alternatives are policies and their outcomes affect many people; in the economist’s language, they bring non­excludable costs and benefits. The economic approach to policy emphasizes the continuity of individual decision making in market and governmental contexts.

The focus on the individual is retained in both an ontological and the­oretical sense.^[22] Within this framework, the individual constitutes the on­tological bedrock. The individual is a locus of preferences and an agent pursuing utility maximization. There are no needs or wants attached to groups, political parties, governmental agencies and bureaus, or legislatures, bureaucracies, and courts. Theoretically, political institutions and structures must be understood from the ground up, as resultants of self-interested actions.

Economic analysis of policy is grounded in individual utility. The market’s idea of consumer sovereignty carries over to the popular sovereignty of po­litically organized individuals. True, the unequal dollars of individuals in the market contrast with the equal votes of citizens and representatives, although a parallel is found with the unequal spending (and influence) of interest groups. The basic similarities are pervasive.

When addressing policy questions, economists are most comfortable in discussing efficiency. Given that certain goals are not individually attainable, the question is how best to attain these goals collectively. In other words, what is the best (most efficient) policy?

We saw earlier that efficiency meant getting the most out of a given set of resources. What does this mean in a policy context? The effectiveness of a policy has to do with whether it achieves its goal. A cost-effective policy may be one that, out of a number of policies that achieve the same goal, is the least costly. This is a truncated form of cost-benefit analysis (since benefits are assumed to be constant). A policy that is best from a cost-benefit stand­point is one that maximizes the difference between total benefits and total costs. Costs and benefits must both be explicitly included in the evaluative process and compared across the available alternative policies. This is the way the economist normally proceeds.

To illustrate the efficiency principle, let us develop an example. Assume that individuals within a political system desire both the capacity to deter external military aggression and a capacity for defense, should deterrence fail. In other words, they want to dissuade potential enemies from military incursions and to be able to fight such incursions if they take place. Military strategists generally concede a negative relationship between these two objectives. If a government wants to reduce the threat of military aggression as much as possible, it will put all of its resources into countervalue (including countercity and industrial targets) forces. This, however, leaves it with no ability to defend itself should attack occur. On the other hand, if it spends all its money on defense, it has nothing left with which to threaten an enemy (in his own homeland), which lowers the costs of a nuclear war.

While this may be a somber example, the economist approaches it in the same spirit as he or she would approach the question of how several alternative oil-drilling operations will affect oil yield and the environment. Let us assume that there are five different policies (force structures, strategies, battle plans, targeting, and so on) that attempt to satisfy the requirements of deterrence and defense. Further assume that these alternatives are equal in all respects, including costs, except for the different combinations of deterrence and de­fense yielded. Figure 6 illustrates the decision problem. In this figure are plotted our five policies and the amounts of deterrence and defense associated with each (using an arbitrary scale from 1 to 10).

The economist examines these alternatives in light of the desired goals. Recall that since both are desired, any policy that increases the government’s capacity for both deterrence and defense is preferred. If there is a single alternative that is superior to all others, this would be the “best” policy. It would be the Pareto-superior point, a single point lying to the north and east of all other points. No such point exists.

Given that there is no single Pareto-superior point, are there any that are Pareto-efficient - points from which we cannot deviate without doing worse with regard to at least one goal? The policymaker throws out points E and D. E is dominated by C (C is northeast) and D is dominated by both A and C. Thus any policymaker who values both deterrence and defense will prefer

Figure 6. Choice of best policy combining deterrence and defense

C to E and C and A to D. However, among points A, B, and C there is no clear preference. All three points are Pareto-efficient. All three points lie on a possibility frontier (a Pareto-possibility frontier) (see Stokey and Zeckhau- ser, 1978:24-5).

Figure 6 provides a graphical illustration of the policy possibilities. In a sense, these alternatives represent the familiar production possibility frontier of neoclassical theory. They provide us with an account of what is techno­logically possible, but they do not speak to what is desirable, except that “more is better.” However, the exact choice can only be determined from information that brings together the preferences of decision makers and avail­able policy alternatives. If we had wanted to illustrate this “best choice” solution, we would first have connected points A, B, and C by a line indicating the possibility frontier. Then we could have drawn indifference curves in the same space.

Indifference curves describe the rates at which one is willing to give up one good in return for another. In a sense they describe the individual sub­stitution ratios of different goods. For example, a policymaker might value both deterrence and defense, yet be willing to give up only a little (or a lot) deterrence for a lot (or a little) of defense.

The solution (the preferred point) lies where the slope of the indifference curve is just tangent to the slope of the possibility frontier, indicating that the marginal rates of transformation and substitution are equalized. The marginal rate of substitution is that rate at which one is willing to give up one good for another. The marginal rate of transformation is the rate at which one is able to acquire one good instead of another. When these rates are equalized, the best choice has been defined.

The illustration is based on highly simplified assumptions: two goals, unitary decision maker preferences, and fixed costs. Other cases call for a more complicated reshuffling of resources. In “Economic Reasoning and the Ethics of Policy” (1984), Thomas Schelling goes through numerous examples in which the economic logic is used to “solve” problems for which common sense often provides a different answer. Gas rationing provides but one example. If gas is in short supply, the price will rise, making it difficult or perhaps impossible for poor people to purchase it. One solution is to ration gasoline. This might seem to have desirable distributional properties since the burden would be supported by all. Schell­ing, invoking economic reasoning, suggests that policymakers let the price rise, tax the windfall profit, redistribute income to the poorer people, but give them the choice of spending their money on gas or some alternative. Schelling’s argument is based on the presumption that, if we are concerned about the impact of high-priced gasoline on the poor, we can deal with the impact more efficiently using the market in conjunction with tax in­struments than by rationing. The market can continue to do its work of reflecting relative scarcities through prices, hence encouraging substitution, while the distributional goal of providing purchasing power for the poor can be satisfied by other means.

If carried to the extreme, this approach might prove startling. Schelling’s example of the case for different traffic and safety measures surrounding rich and poor neighborhoods is provocative. The poor, to be sure, might prefer to spend their money in other ways than for airline safety measures, but is this relevant? One can also imagine the Occupational Safety and Health Administration (OSHA) asking workers if they would like to give up some personal safety in return for payments to spend as they like. If so, then why not apply the same reasoning to the right to trial by jury and so on? The problem, as Levine argues (1983), is that the efficiency-cum-choice orien­tation fails to provide a dividing line between individual choice guided by preferences and rights. The latter refer to a realm of entitlements based either on widely acknowledged perceptions of common needs or citizens’ rights, and are not allowed to exchange (that is, to be bought and sold). In discussing policy, the economist should distinguish between economic policy designed to improve choice and efficiency and policy designed to preserve or advance rights (Levine, 1983:84).

While institutions are not necessarily political, political institutions are central to politics. Hence we discuss here the way that economic reasoning is used to explain behavior within institutional settings as well as to explain the changing content of institutions themselves.

First, we need some definitions. North and Thomas define an institution quite broadly, as “an arrangement between economic units that defines and specifies the ways by which these units can co-operate or compete” (1973:5). In a different source, North specifies that “institutions consist of a set of constraints on behavior in the form of rules and regulations; a set of pro­cedures to detect deviations from the rules and regulations; and, finally, a set of moral, ethical behavioral norms which define the contours that constrain the way in which the rules and regulations are specified and enforcement is carried out” (North, 1984:8).

Note that these definitions treat institutions as constraints (or opportuni­ties) external to economic agents. This is important because it allows the preservation of both the individual focus and the maximization hypothesis, altering only the costs and benefits of various courses of action. Now the individual maximizes subject to the distribution of endowments, technology, preferences, and given institutional arrangements. The introduction of in­stitutions adds one more variable to the basic economic equation.

The economic analysis of institutions stresses the ways that institutions can foster instrumental cooperative behavior, reduce (or increase) transaction costs, and provide the organizational basis for production and exchange. The focus is on the relationship between institutions and efficiency - that is, the ways in which institutions facilitate or retard self-seeking.

Institutions, within this framework, are rules or procedures that prescribe, proscribe, or permit particular behavior. Political institutions, as applied to economics, may define appropriate objects of exchange, the rules guiding the exchange process, and property rights with respect to both benefits and liabilities. The political aspects of political institutions lie in their origins in the state and in the use of state power, authority, and sanctions to enforce prescribed behavior.

If the economic approach to institutions can best be understood in contrast with sociological conceptions,^[23] perhaps political institutions can best be understood in contrast with markets. The pure idea of market behavior has individuals pursuing their preferences in a world of freely exchanging agents and alienable goods. The value of these goods is determined by the interplay of relative scarcities and relative preferences. Political institutions also estab­lish opportunities and constraints, but they take the form of authoritative policies that alter costs. We return to this point later in the chapter.

Institutions and market behavior. The preceding discussion raises the issue of the connection between markets and institutions. There are at least three ways of conceiving the relationship. First, markets are themselves institu­tions. They are not simply unstructured collections of individuals freely engaging in buying and selling. The market economy is, as Polanyi argues, “an instituted process” (1957:243-70), not just in the sense that behavioral regularities are present in the market, but because rules are embedded in the market itself. These rules govern the terms of exchange and responsibility for external costs. Agreements about property rights and contract enforce­ment, and prohibitions against theft, coercion, and fraud are understood. These are part of the constitutive rules or conventions of the market without which exchange among individuals could hardly work (Field, 1984:684).

Second, institutions generally define the scope of market exchange. Some objects or characteristics may not exchange for personal or cultural reasons. A person may not sell a child even though it might be profitable to do so. Another person does not sell her labor for work that is considered objec­tionable. In addition to these individual “bans on exchange,” many such bans are politically proscribed. Governments usually ban trading votes for dollars (buying votes), though the rules governing the relations between interest groups and U.S. representatives leave this question somewhat open. Governments may or may not prohibit the manufacture and sale of alcohol, pornographic magazines, and insider trading on the stock market. They may or may not establish laws to govern the sale of knowledge gained through experience in “public service” - to control whether former public servants may publish memoirs before a certain period of time out of government or establish a consulting service to advise weapons industries on the basis of a brief appointment in the Defense Department.

Banning certain kinds of exchanges alters market allocations in important ways. If the rules work perfectly, a portion of the market is suppressed and potential “gains from trade” are lost. In addition, there is the opportunity cost of making and enforcing rules. The time, energy, and resources of legislators, bureaucrats, and lawyers have alternative uses that are forgone in making and enforcing rules. These are the social costs of politics empha­sized by some public choice theorists. ∙

Third, apart from their function as prohibitions, political institutions are typically used to alter the incentives underlying market exchange. In the former Federal Republic of Germany, special subsidies and depreciation allowances were given to industries specializing in exports. In Japan the high rate of personal savings has much to do with the nature of financial institutions and the politically supplied rewards for savings. In France special subsidies are given to cutting-edge technologies and in the United States, as everywhere else, tax policies and environmental regulations are not imposed in a flat (even) way, but discriminate among various sectors, penalizing some and rewarding others. Sometimes political rules “merely” channel economic ac­tivity from some sectors to others. However, rules can also affect the balance of wealth-producing versus transfer or rent-seeking activities (Krueger, 1974; Tollison, 1984).

Economic reasoning and political institutions. Neoclassical economic theory has traditionally ignored institutions. If it recognized them at all, it treated them as constants that did not function actively in explaining allocative activity. The usual justification was that while institutions do matter in the long run, for questions concerning short-term allocative activity, they can be ignored. In recent years there has been a revival of interest in institutions, spurred in part by increasing attention to economic history, in part by a recognition that comparative economic behavior cannot safely ignore institutions.

Among those economists who examine institutions, there are two schools of thought. One school allows that institutions are important, but takes them as provided by forces outside the economic model. The chief aim here is to understand the significance of institutions for allocative behavior. This, in turn, is accomplished by exploring the comparative incentive features of different institutions. Since institutions are exogenous (yet variable), we may refer to this as the “institutionally situated rational choice approach.” The second school takes institutions as endogenous; that is, institutions themselves become a variable to be explained by economic behavior. We return to this second approach shortly. Within the institutionally situated rational choice school, there is a distinction between those who do empirical and those who do analytic work. The empirical work conforms to the preceding discussion about the comparative incentive features of institutions. What difference do institutional variations make for the behavior of rational, self-interested ac­tors? This is the central question posed.

The work of Douglass North (1981, 1984), Margaret Levi (1988), Mancur Olson (1965, 1982), and Robert Bates (1981, 1983, 1988) falls within the empirical tradition. North examines the effect of changes in property rights on economic growth. He is interested in how institutions can change to bring private and social costs more closely together (for example, by reducing externalities). Bates has written extensively in the area of economic and political development and has argued that incorporating political institutions into models of development can help explain why political and economic actors seem to behave irrationally on purely economic grounds. His work on price setting of food goods by state agencies partly explains the attitudes of peasants toward production (Bates, 1981). The incentive to produce more food may not exist in a country where the terms of trade between farm and city are established by state institutions that depend on urban areas for political support.

The analytic branch of this school asks different kinds of questions, ques­tions that explore the logical properties of models that simultaneously in­corporate terms describing exchange behavior and institutions. Is it possible to design institutions incorporating democratic voting procedures that do not have the instability properties pointed out by Arrow? How, if at all, may the introduction of specific institutional arrangements affect the stability of var­ious voting systems (electorates, legislatures, and committees)? Do institu­tions “inherit” (see Riker, 1980; Shepsle, 1983) the same characteristics that preferences have? That is, are institutions unstable in exactly the same way that preferences for specific outcomes are unstable? These questions are representative of the analytic tradition.

Among those in the analytic tradition, the work of Riker (1980), Schofield (1980), Shepsle (1979a, b, 1983), and Shepsle and Weingast (1981) is par­ticularly relevant for our purposes. Much of the research in this tradition has been inspired by Arrow. Some have been anxious to confirm or generalize the basic instability theorem; others have taken Arrow’s book as a challenge and have asked a la Tullock (1981), “Why so much stability?” Riker (1980) starts from the theorem of voting instability and goes on to assess the im­plications for the study of political institutions. The central question he asks is whether institutions inherit the same problem of social intransitivity and cycling that characterize preferences over specific outcomes. If institutions are, as he argues, “congealed tastes,” this would seem to imply that insti­tutions are unstable too, even if these instabilities manifest themselves more slowly than specific policy instabilities.

Shepsle, a student of the U.S. legislature, starts with the point that the observable world is not as chaotic as implied by the work of Arrow and McKelvey. Why is this so? Shepsle claims that part of the answer lies in the way political institutions mediate preferences and social choices. Not just any preferences may be expressed in any particular fashion. There are rules about what preferences can be expressed, in which order, in combination with what particular other preferences - that is, which coalitions are possible (Shepsle, 1983:1-9). As Shepsle and Weingest put it:

In our view, real-world legislative practices constrain the instability of PMR [pure majority rule] by restricting the domain and the content of legislative exchange. The latter, in our view, is part of the problem (though by no means the only part) with PMR, not part of the solution. Throughout, then, we hope to convey what we believe is a compelling case for answering Tullock’s question, ‘Why so much stability?’ with “Institutional arrangements do it.” (1981:504, italics in original)

The work of Shepsle, of Riker, and of Fiorina (1982) in some ways returns the attention of the formal theorist to the more traditional preoccupation with the nuts and bolts concerns of political scientists. In his U.S. Senators and Their World (1960), Donald Matthews provides an account of the op­eration of the U.S. Senate that is rich in normative and institutional details. Formal theorists, starting from an abstract proposition about the instability of social choices in normatively and institutionally unconstrained settings, have (rediscovered, as Shepsle put it, “that tastes and th<ir expression are neither autonomous nor necessarily decisive” (Shepsle, 1983:1). According to Shepsle, the actual world of legislatures, committees, and electoral systems is characterized by complex divisions of labor (not atomistic exchange), re­strictions on preferences as in germaneness rules, and sequencing rules spec­ifying how items may come up, be amended, and voted upon. These conditions place some distance between operating institutions and the pure majority rule of Arrow’s world. Equilibrium is possible after all.

Finally, we come to rational choice theory with institutions as endogenous. This approach differs in important ways from the institutionally situated rational choice approach. The latter approach is a natural extension of the basic economic model simply activating a factor that almost everyone rec­ognized in the first place. By contrast, endogenous institutional theory alters the theoretical status of institutions, shifting them to the left-hand side of the basic explanatory equation. In this approach, institutions themselves become the objects of choice, arguments in utility functions, and “outputs” to be explained in the same manner as allocative activity in general. The revised institutional model now says that variations in the form and content of institutions are explicable by appealing to exogenous changes in endow­ments, preferences, and technology. For example, given an exogenous in­crease in the labor/land ratio, property rights assignments should shift in favor of holders of the scarce resource, that is, more favorable to landowners, less so to workers.

The attempt to create an endogenous theory of institutions is reflected in the work of North and Thomas (1973), North (1978, 1981, 1984), Levi (1988), and Basu, Jones, and Schlicht (1987). For North and many of his followers, the key forces acting on institutional change have to do with the potential gains associated with innovation, production, and exchange. Forms of prop­erty rights that cut down on external benefits (that is, positive externalities), discourage rent seeking, and reduce the costs of making and enforcing con­tracts are central to North’s analysis of institutional change. Thus the modern corporation (which limits personal liability), the stock market (which pools capital and reduces information costs), and patents (which protect income arising from innovations) are stock examples. Changes in relative prices of factors and products provide the stimulation for changes in institutions. Levi also relies to some extent on the role of institutions in reducing transaction costs, although her work is supplemented by a theory of power and bargaining (1988:17-23).

Finally, the work of Riker (1980) and Shepsle (1983), although primarily analytical, can also be interpreted as part of the endogenous institutional change movement. In commenting on Riker’s work, Shepsle reflects that Riker “treats institutions like ordinary policy alternatives in an important respect: they are chosen” (1983:27). Shepsle’s distinction between institu­tional equilibrium and equilibrium institutions attempts to capture the dif­ference between those institutions that can provide equilibrium of policy outcomes versus those that can themselves be institutionally stable.

The effort to create an endogenous account of institutions - that is, to provide a theory of how institutions are created and how they change, is not new, but it is important. It involves an effort to resurrect a project that was central to Marxian political economy. But while Marx saw the engine of institutional change in the dialectical tension between forces of production and relations of production, the new institutionalism of neoclassical econom­ics focuses on institutions as organizational, procedural, and rulelike re­sponses for economizing on transaction costs and capturing the gains from innovation in production and exchange. This literature is not limited to the theory of the firm. It extends well beyond this to include efforts to understand legal changes, changes in property rights, and changes in the form and content of political institutions over long historical periods.