Those who analyze political institutions often employ broad labels. They refer to democracy or authoritarianism; presidential or parliamentary governments; or to systems of plurality or proportional representation. In this paper, I refrain from addressing politics at this level and adopt an approach that is far more micro in perspective. The approach I take can be referred to as "neo-classical," with the attributes that the name implies: rationality in choice, an emphasis on the individual decision maker, and an explicit treatment of the process by which individual actions aggregate into collective--in this case, policy--outcomes.

Several considerations underlie the adoption of this approach to the analysis of institutions. It is often difficult to reason from the macro-level attributes of political institutions to the content of public policies. Are, for example, democracies more likely to adopt liberal economic policies than are authoritarian regimes? There are few ways to achieve purchase on this issue, with the result that multiple lines of argument can--and have been--advanced on both sides.¹ Where such arguments are persuasive, it is often because they appeal to micro-level properties of the political structures. By micro-level features, I refer to the identity of the agenda setter, gatekeepers, and veto players--players whose powers are created by the structure of institutions and the rules that they embody

It should be noted that an implication of this approach is that there is no reason to expect a particular kind of government to be "best," when rated in terms of expected choice of policies. Rather than the institution itself, it is the structure that it imposes upon the process of decision-making and the incentives that it generates for decision makers that shape policy outcomes. Two parliamentary, while formally similar, can differ in terms of these properties, and so generate different policies. By the same token, a presidential system and a parliamentary system can generate similar economic policies, in so far, say, as those who demand low rates of inflation occupy veto points in each system.

The contribution of this paper, then, will not be prescriptive; it will not call for the adoption of one kind of institution as opposed to another. Rather, it will be methodological. The forms of analysis that it advances can be applied widely and employed to discern the policy bias that inheres in a particular institutional setting.

Policy analysts often view governments as unitary actors with coherent sets of objectives, which then choose policies in order to facilitate the attainment of those objectives. Some states are considered "developmental" and others "predatory," for example, and their policies explained accordingly.² Or policy choices are attributed to the preferences of a collective group: the president and his inner circle, "rent-seekers," or "the technocrats."³ Either for good or ill, policymaking is thus viewed as the rational selection of means for attaining outcomes.

But policies rarely result from the deliberations of a single entity; rather, they emerge rather from competitive political processes, in which diverse interests advance competing alternatives, resulting either in the victory of one or a compromise among several. Even totalitarian states, we have learned, possess internal political conflicts; they contain internal "empires" and competing bureaucracies, both economic and political, each with its own special interests.⁴ And where governments are subject to the dictates of a single party, we often find, upon inspection, that the party is rife with internal conflict among competing political factions.⁵

A neo-classical perspective encourages us to abandon the unitary actor mode of explication. It drives us to the micro-level. It encourages us to attempt to account for collective outcomes as the result of political competition. It therefore draws our attention to the internal political processes that underlie the formation of policy. And it gives us the tools by which to account for the way in which policy choices emerge as outcomes from a competitive political process.

To illuminate this framework, I apply it to the decision to form capital and, in particular, the decision to make investments that are irreversible at great cost.

The best way in which to introduce the approach is through an example. In the illustration below, Figure 1, there are two dimensions of public policy. One focuses on the provision of social services; the other on military preparedness. A point in the surface defined by the two policy dimensions represents a combination of the two policy outputs. We assume that there are three decision makers, whom we label A, B, and C, and that they possess policy preferences. Using a spatial framework, we can order the possible policy combinations in accord with the actors' preferences. The "ideal" or "bliss" point for each appears as a point bearing the actor's identifying label; this point represents the actor's most preferred combination of social services and military preparedness. The further a combination of the to policies from an actor's ideal point, the lower the utility she derives. Points of equal loss appear as indifference curves; these lie equidistant from the ideal point. Points lying between a given indifference curve and ideal point lie closer to that actor's bliss point, and are therefore preferred to any point on the curve.

A moment's consideration reveals that the triangle formed by the three ideal points constitutes the set of policies that are efficient. Within the triangle (and including its borders), no change in policy can make any actor better off (i.e. move the policy choice closer to one actor's ideal point) without making another actor worse off (i.e. without moving the policy outcome further away from the ideal point of another). The triangle thus defines the Pareto set. Socially rational policy choices should therefore fall within this triangle. It should contain the outcomes produced by "good policy."

Should collective decisions be made in the context of market institutions, then we know where the outcome would lie. By the fundamental theorems of welfare economics, were the three individual's preferences aggregated into collective outcomes by competitive markets, then the policy choices would fall within the triangle, or Pareto set.

The same is not true for the outcomes of decisions made in the context of political institutions, however. Consider the point marked SQ, for status quo. When the indifference curves for the actors are drawn through it, they form three "preferred to" sets, one for each actor; each consists of all the points lying closer to the ideal point than does the status quo. And now introduce a political rule for choice making. By majority rule, when two out of three of the actors prefer a policy combination to the status quo, then that choice would replace the status quo. In Figure 1, the intersection of the "preferred to" sets form sets that are shaped like petals. By majority rule, any point in these petals would over turn the status quo by majority rule.

Several implications follow. Note that elements in the sets of points majority preferred to the status quo include points that fall outside of the triangle, i.e. outside the Pareto set. The political outcome is therefore likely to be inefficient. Note also that not only is the set of outcomes in the petals large, but also, the elements in the set cannot be ordered. Political outcomes, then, are likely to be indeterminate as well.

To illustrate, consider points 1,2, and 3, in the diagram above (Figure 1). Each is majority preferred to the status quo. Ordering them in terms of their distance from his ideal point, the actors would rate them in the order recorded in Table 1.

In an effort to choose a policy, the three decision makers would vote. But the results of the vote would be indeterminate.

To see this, note that:

2 is socially preferred to 1:
A and B prefer 2 to 1.
Only C prefers 1 to 2.
By majority rule, then, 2>1.

3 is socially preferred to 2:
B and C prefer 3 to 2.
Only A prefers 2 to 3.
By majority rule, then, 3>2.

But 1 is preferred to 3:
A and C prefer 1 to 3.
Only B prefers 3 to 1.
By majority rule, then, 1>3.

The polity is thus unable to define its best interests. Instead, it achieves an ordering that cycles, with alternatives that are least preferred rising to the top (as in the instance of alternative 1 in Table 2) and those defeated in earlier votes being selected at later points in the deliberations.

Given the large number of points that lie in the "petals," there are thus a large number of possible alternatives that will be preferred to the status quo. And the polity is liable to cycle indeterminately among such points; it is unable to rank them in an orderly fashion. Political outcomes are thus highly indeterminate and the policy environment unstable. We have assumed that the individuals are rational; but, choosing as rational individuals, they would behave in ways that are socially irrational.

Consider, then, the problem faced by an investor. In particular, consider the problem faced by an investor who seeks to form fixed and specific capital, i.e. to invest in a project, a plant, or a facility that is of value in one site or in one use only. Being specialized, the investment lacks much by way of value in its next best use. The investor is therefore vulnerable.

To sharpen the point further, assume as well that the returns to this investment depend upon the choices of policy. They may depend upon the rules defining costs and what therefore the revenue department will define as taxable profits; policies toward the exchange rate, the convertibility of the currency, and the remission of profits; and the regulations governing the hiring, the termination, and the remuneration of labor. The investor therefore incurs risks, some arising from the nature of the investment and others from the impact of public policies.

When confronted with the indeterminacy of policy choices, as suggested by this example, the investor's best response may be to choose to refrain from investing. She may instead chose to keep her options open and to postpone launching projects that would be costly to abandon.

The illustration thus highlights the economic dangers of political life. It also suggests possible corrections. One source of stability is the imposition of further restrictions on the process of choice. Put another way, it is the creation of structures and procedures, which can sometimes be called institutions.

To illustrate, return once again to the points labeled 1, 2, and 3. As shown in Table 2, left solely to majority rule, these alternatives cannot be coherently ordered, resulting in political instability. But the illustration also suggests that in such situations the outcome that is achieved depends upon the procedures that are adopted in the making of collective decisions. It therefore suggests ways of achieving stability, and offers insights into the origins of institutional bias.

Table 3 depicts three possible agendas for choosing among the three alternatives. In the first, policy 1 is offered as an alternative to policy 2, and the winner then paired against 3. Point 2 lies closer to the ideal points of A and B than does point 1, and so 2 wins in the first round; point 3 lies closer to the ideal points of B and C than does point 2, and therefore defeats 2 in the second. The agenda - 1 vs. 2, and the winner vs. 3 - thus yields policy 3 as the choice for society.

The procedure thus reduces the multitude of possible outcomes to one that is unique. Outcomes that were unpredictable thus become determinate. In an environment that hitherto has been unstable, the existence of an equilibrium can thus be structurally induced.⁶ Structures thus make possible predictable outcomes. Note moreover, rows two and three, which depict alternative agendas. They illustrate that even in the presence of the same structure of preferences, different agendas yield different outcomes. They therefore illustrate as well that political procedures are not neutral; they privilege the fate of some policy options over others. In so doing, they instill systematic patterns of bias in polities.⁷

Structures and procedures therefore produce determinacy in political life. Insofar as they determine the agenda by which outcomes are chosen, they create equilibria in situations, such as those governed by majority rule, where they may previously have not existed. They help us as well to institutionalize a pattern of bias in the polity, yielding a systematic preference for particular kinds of outcomes.

The structure and procedures that govern political life thus affect the content of policies. Insofar as this is true, however, then we should expect sophisticated actors to seek t alter the political structures to their advantage. Insofar as there is stability in policy, them, it is because people do not opportunistically modify the procedures that govern the policy process. What accounts for such stability? Put another way, when do structures become institutions?⁸

The answer most commonly given is: When actors fear to deviate from the rules. When they fear doing so, then the rules become self-enforcing and the structures will remain in place; they will become institutionalized. That answer, however, only gives rise to further questions, the most obvious of which is: under what conditions will deviations be penalized?

To answer this question, we need to anchor our response in a particular setting. By way of illustration, I choose the case of the coffee industry in Colombia, focusing in particular on policy toward exports. The industrial structure of Colombia's coffee sector resembles that of many nations in Africa; it contains a plurality of small-scale farmers. Yet, since the mid-1930's, the coffee sector has been, by and large, subject to policies that would be the envy of growers anywhere. The government of Colombia has tended to provide coffee exporters with a highly favorable set of economic policies: an exchange rate that closely approximates the market rate, low levels of inflation, and taxes that, when imposed, are then placed in funds controlled by the industry. In the making of export policy, the government of Colombia thus exhibits a policy bias in favor of the coffee industry. The policies have tended to be stable and predictable. The commitment to the coffee industry appears institutionalized. Why? And why, in particular, given the contrast in the way peasant producers have been treated elsewhere in the world?

The origins of the commitment appear to lie in the structure of Colombia's institutions. These institutions confer power upon coffee growers, insuring that policy makers will take their interests into account. To deviate from pledges to safeguard the interests of coffee growers would be to incur political losses, thus rendering pledges credible to the industry.

This analysis is based upon an understanding of the structure of political competition in Colombia. The issues that defined the rivalry between the two great parties--the sanctity of property rights, the power of organized labor, and the position of the church--sufficiently correlate that they form a single policy dimension. We can therefore apply the form of spatial reasoning developed above, but in this instance, to a one dimensional issue space. Within the structure of politics defined by these issues, the coffee-growing regions lay at the political midpoint. To secure a political majority, i.e. to secure more than a half of the votes of the electorate, those ambitious for power at the national level then had to bid for the support of the coffee growers; should they fail to capture political support from the mid-point of the electorate, they would then fail to secure the majority needed to prevail in the competition for office. Both the Conservatives on the right and the Liberals from the left therefore propounded "pro-coffee" policies. No aspirant for national office could afford to deviate form this position, thus rendering the favorable treatment of Colombia's small farmers an institutionalized component of the policy regime.⁹ The incentives created by political structures therefore rendered the government's policy pledges credible. It was in the interests of those who aspired to office to implement them.

Figure 2 illustrates the structure of political competition in the 1930s. Similar diagrams could be composed for later periods.

Once policies are credible, additional forces come into play that further enhance their stability. People become willing to commit to investment programs. In particular, they become more willing to make fixed and specific investments, the value of which may be contingent upon the maintenance in place of the policy regime. And once they have done so, they then acquire a vested interest in the policy. Should the policy regime change, then those who have invested would face a loss of value. Because the investments are of value in this use but of little value in any other, those in the industry cannot readily employ the market to safeguard themselves. They cannot exit; rather, they must give voice (Hirschman 1970). They thus become a lobby in favor of the policies.

When predictable and credible, the policy choices made by governments thus elicit the formation of capital whose value is conditional upon the continuation of these policies. The owners of the assets thus formed constitute a lobby on behalf of the of the policy regime. Should political agents seek to deviate from their promises, they would them trigger "alarm bells" (McCubbins and Schwartz 1987), or protests by organized interests. And those with a vested interest in the industry would seek to punish them politically for their defection.

We can thus appreciate how political structures can generate a pattern of bias in policy making, and thereby introduce political stability. We can see how patterns of choice become self-enforcing, thereby institutionalizing policy choices.

It may be useful to apply this framework to a relevant problem: the role and the power of technocrats. Qualitative evidence, at least, suggests that the reform of economic policies--the reduction of budget deficits, the promotion of market forces, and so forth-- is accompanied by the empowerment of technocrats and the depoliticization of economic policy making. By the latter is often meant the making of economic policies by independent agencies, such as a central bank, within which economists debate and choose policies, free from interference b politicians, private interests, or partisan political forces. The inference is therefore sometimes drawn that "good policy" results from powerful technocrats, who are disinterested and a-political and positioned above the political fray.

In engaging such arguments, it is useful to recall that technocrats possess neither wealth nor power. They merely posses PhD's. Noting that, we may assume that if technocrats possess political clout, it is because they have been rendered powerful. That is, they have been found to be politically useful and placed in positions that impart influence to their policy preferences. The depoliticization of economic policy making is itself a political act.

In keeping with the methodology outlines in this paper, we may best begin by considering political preferences. This implies the initial recognition of the distributional component of economic policy. In accord with their endowments and preferences, people will favor or oppose policies that call for the stabilization of the economy, the liberalization of markets, and the promotion of openness and trade. The adoption of such policies favors the interests of some groups and, at least in the short run, hurts the interests of others. Insofar as economic technocrats are able to implement such policies, then, it is because those who possess power find it not only normatively desirable but also politically advantageous to empower them, so as better to ensure the selection of such policies.

As we have seen, one way in which power can be created is by structuring the process by which proposals become policies. Politicians can privilege the position of any group by injecting its proposals into the appropriate place in the flow of decision-making. The chief executive is the agenda setter; he controls the staff that structures the flow of policy papers and the deliberations of cabinet committees. He can therefore privilege the position of technocrats by assuring that their viewpoints enters sequence of deliberations at the stage in which their viewpoints are most likely to prevail. Under the oversight of a sophisticated agenda setter, the technocrats will appear powerful because their policies so frequently prevail. But that success reflects the structuring of the political process by their political patrons, rather then the power of the technocrats themselves.

Another source of power is the capacity to veto. In the single dimensional issue space, discussed above, the median voter occupied such a position; no coalition lacking that player could command a majority. If the chief executive organizes the flow of policy making such that the ministry of finance or the central bank must sign off on proposal for it to become policy, then he empowers those agencies. Nothing can go forward that they oppose; and those who seek to advance their own proposals must incorporate the viewpoints of those agencies.

The question then arises: Why would the executive give powerful positions to technocrats or chose to empower their agencies? The answer must be: because he finds the championing of their policies politically rewarding. This condition will hold so long as a constituency that benefits from the policies advocated by the technocrats mediates the political fortunes of the agenda setter. If those who make, or break, his political prospects benefit from price stability, market-clearing prices for capital and foreign exchange, and openness to international markets, then the agenda-setter will structure the policy process so as to privilege the position of technocrats, empower their agencies, and so institutionalize a pattern of bias in favor of "good" economic policy.

Economic technocrats, we have noted, are merely PhD's. While they may lack wealth or power, they do possess expertise. While unable to evaluate the technical abilities that they bring to bear upon the formulation of policies, a politician can monitor the impact of their policies upon his constituents. He will seek to augment, or diminish, the power of the technocrats, depending upon the impact of their advice upon the welfare of his constituents, and thus upon his political fortunes. Taking into account the visible impact of policies on his constituency, politicians therefore chose whether or not to trust technocrats. And employing the same information, private agents chose whether or not to lend credence to the pledges of political elites to adhere top the technocrats prescriptions. Political accountability to a constituency than can penalize or reward the political leadership thus underlies both the willingness to delegate power to the technocrats and the effectiveness of economic polices.

References

Bates, Robert H., 1997, Open Economy Politics. Princeton: Princeton University Press.

Bates, Robert H., and Paul Collier, 1991, The Case of Zambia. In Political and Economic Interactions in Economic Policy Reform, edited by R. H. Bates and A. O. Krueger, Oxford: Blackwell.

Bates, Robert H., Avner Greif, Margaret Levi, Jean-Laurent Rosenthal, and Barry Weingast, 1998, Analytic Narratives. Princeton: Princeton University Press.

Haggard, Stephen , and Steven B. Webb, eds., 1994, Voting for Reform, New York: Oxford University Press for the World Bank.

Hirschman, Albert O. 1970, Exit, Voice and Loyalty, Cambridge MA: Harvard University Press.

Hough, Jerry and Merle Fainsod, 1979, How the Soviet Union is Governed, Cambridge, MA: Harvard University Press.

McCubbins, Mathew D., and Thomas Schwartz. 1987, A Theory of Congressional Delegation. In Congress: Structure and Policy, edited by M. D. McCubbins and T. Sullivan, New York: Cambridge University Press.

Meier, Gerald M., ed. 1991, Politics and Policy Making in Developing Countries, San Francisco: International Center for Economic Growth.

Przeworski, Adam and Fernando Limongi, Modernization: Theory and Facts, World Politics, 49:155-83.

Roeder, Philip, 1993, Red Sunset, Princeton: Princeton University Press.

Schattschneider, Elmer Eric, 1965, The Semisovereign People, New York: Holt, Rinehart and Winston.

Shepsle, Kenneth A., 1979, Institutional Arrangements and Equilibrium in Multidimensional Voting Models, American Journal of Political Science 23:27-60.

Shepsle, Kenneth A., and Bary R. Weingast, 1981, Structure-Induced Equilibrium and Legislative Choice, Public Choice 37:503-519.

Williamson, John, ed. 1997, The Political Economy of Economic Policy Reform, Washinton D.C: The Institute for International Economics.