May 3, 1997

Public Choice, Public Goods, and Constitutions


A. Introduction
B. Entrepreneurship and Private Goods
.....1. Subjectivism: Old and New
.....2. An Example
.....3. Entrepreneurship
.........a. Dampened Incentives
.........b. External Effects
C. Incentives to Supply Public Goods
.....1. The Appraisal and Production of Public Goods in the Market Economy
.....2. Two kinds of Incentives
.........a. Known But Not-Yet-Produced Public Goods
.........b. Unknown Public Goods
.....3. A Note on Subjectivist Methodology
D. Creating a Democratic Government: the Constitution
.....1. Preliminary Considerations
.....2. Rules for Hiring Politicians and Bureaucrats
.........a. A Representative Legislature
.........b. Election Procedures
.....3. Rules Governing the Actions of the Agents
.........a. Enforcement of Private Property Rights
.........b. Tax Rules
.........c. Rule of Law
.........d. Separation of Powers
.........e. Rules Encouraging the Supply of Information
.........f. Rules to Control the External Costs of Collective Decision-Making
.....4. Constitutional Change
E. Conclusion

Public Choice, Public Goods, and Constitutions

1. Introduction

James M. Buchanan and Gordon Tullock (B-T) defined the scope of Public Choice in theCalculus of Consent (1962). They said that the analysis of collective decision-making by means of government begins where the private goods, minimum government market economy ends. Whereas economics is concerned with the demand and supply of private goods, the economic theory of collective decision-making (later to become Public Choice) is concerned with the demand and supply of "public goods."(Buchanan and Tullock 1962: 34-35)

Ordinary logic suggests that their next step would be to define public goods and to go on to discuss the benefits and costs of using collective decision-making, as opposed to markets, as a means of causing them to be produced. Instead, B-T adopted an approach that focussed entirely on costs. In place of the notion that government institutions, rules, and policies could increase net benefits via the provision of public goods, they substituted the notion that these things could reduce costs.(ibid.: 44-45) They proceeded to compare what they called cost reductions due to the policies with the costs of making the policies. Included in the costs of making the policies was each individual's anticipated loss due to the prospect that collective decisions might harm him. After shifting the focus, B-T never returned to the public goods theme.

My aim in this paper is to follow through on B-T's stated goal. To do this I must first develop an analysis of the characteristics of public goods, namely, jointness and nonexclusion. If we assume that members of a collective wanted a government exclusively to overcome the incentive problem in the supply of goods with public goods characteristics, what kind of government would they want?

The goal is to develop a Public Choice along the following lines. First, assume a no-government, market society in which no public goods are currently being provided. Second, assume that the community members suspect that certain public goods can be more efficiently supplied by means of a government than through the market. Third, assume that the members of the community are predisposed toward peaceful cooperation in their choice of a means of forming a government.(1)What kind of collective institutions are they likely to contemplate forming? And how would the collective institutions measure up to the market provision of those same public goods?

To answer this question, we apparently must accomplish two preliminary tasks. First, we must choose a method of studying the market economy. Second, we must thoroughly understand the "public goods problem." Accordingly, this paper begins with a discussion of method and the public goods problem. After this, it proceeds to discuss possible collective solutions.

B-T regarded their book as most relevant to explaining and evaluating constitutional rules. Thus, a critic, or revisionist, ought to be able to provide an alternative, hopefully superior set of explanations for constitutional rules. This paper argues that the main problem faced by members of a collective in setting up a system to cause public goods to be supplied is that of determining how to hire and control agents. The constitution is defined as a possible solution to this problem.

The idea of describing government officials as agents can be thought of as an extension of Frank Knight's analysis of the market economy. Knight regarded the entrepreneur as the agent of consumers in the supply of private goods. (Gunning, 1993) Entrepreneurship is self-employed and, under certain conditions, motivated by business profit to act in the consumers' interests. A collective that wants to cause public goods to be supplied cannot rely solely on agents in the market. It must employ agents collectively in various capacities that aid in public goods supply. Of paramount importance is that the members of the collective must give the agents limited control of their presumed monopoly over force. The problems faced by the members are how to motivate the agents to carry out their assigned duties and how to control their use of coercion. We define the constitution as a means of employing and controlling two types of agents -- (1) agents to make decisions for the members of the collective (politicians) and (2) agents to administer the decisions made by the first type (bureaucrats).

In order to clarify the distinction between market supply and supply by government, part two of the paper describes the significance of entrepreneurship and entrepreneurial incentives in producing private goods. Part three defines a public good and shows why entrepreneurial incentives are absent or not as strong. Part four derives provisions of a constitution on the basis of the assumption that members of the collective want to control their agents' use of the monopoly over force. In doing so it provides a foundation for Public Choice that differs from that of B-T. Part five presents some brief remarks about the relationship between traditional Public Choice and the variation of Public Choice described in this paper.

2. Entrepreneurship and Private Goods

The only proper method for studying the market economy is methodological individualism. B-T used a form of methodological individualism that corresponds to that of some early Austrians, P. Wicksteed, and the London School of Economics.(Buchanan 1969; Buchanan and Thirlby 1981) This method, which is widely used in professional economics today, has become the standard for Public Choice. I have argued, however, that it is obsolete. Let me explain.

W. S. Jevons, C. Menger, and L. Walras each sought to develop a theory of value based on methodological individualism. They recognized that the fundamental problem is the complexity of a large market economy. They tried to deal with this complexity by constructing what we now call a general equilibrium model of an economy. While such a model helps to cut through the complexity, it distracts our attention away from the individual. The individual in the model is not really a human actor but a robot maximizer. Later economists removed this potential for confusion by developing the concept of the entrepreneur. J. B. Clark (1899), F. Hawley (1907), H. Davenport (1914), F. Knight (1921), and L. Mises (1966) each contributed to the development of a more sophisticated framework for helping us combine the general equilibrium model with the entrepreneur concept in order to apply methodological individualism. I have called the culmination of this line of thinking the "new subjectivism."(Gunning 1994; 1997c.)

Subjectivism: Old and New

Subjectivism, as most professional economists use the term, is the same as methodological individualism in its use by Buchanan and others. According to it, individuals' preferences are the source of all prices. And individuals' choices are the cause of all economic interaction and institutions. The subjectivist defines price, economic interaction and institutions in terms of individual preferences and choices. She uses no concepts that cannot ultimately be traced to individual preferences and choices. This idea has been most fully developed in the "theory of cost." Subjectivist cost always refers to the preferences and choices of a particular subject or subjects.(Buchanan 1969; Buchanan and Thirlby 1981)

As important as it is to take this view of cost and other phenomena, it must be admitted that it does not get one very far. If we want to talk about the market economy, about the effects of laws, or indeed about any economic event that is not totally isolated from the rest of economic interaction, we must develop a way of simplifying the complex network of interaction involving millions of specialists. This is precisely what the subjective value theorists aimed to do. They constructed an image of an economy based on functions and roles. This enabled them to show that, under certain specific conditions, individuals acting in the role of profit-maximizing producers have an incentive to allocate the factors in a way that generate what individuals acting in the consumer role regard as the highest-valued consumer goods. This was subjectivism at the turn of the century.

To avoid a protracted discussion of the history of economic thought, we can skip directly to Knight. Knight's image of the market economy was one in which individuals sort themselves into roles based on their personal beliefs about their specialties and about other variables in the profit equation.(Gunning 1993) Some of these individuals come to act as producing entrepreneurs. They become the leaders, the employers of others, who risk their fortunes in production and sales ventures and thereby become the most important cause of the satisfaction of consumer wants. Knight conceived of entrepreneurs as agents of the consumers. Each agent bets that his knowledge of others' preferences, abilities, and knowledge is more correct than that of other individuals who might employ a given set of factors of production. Each bets that he can earn money by producing what ordinary people regard as wealth. Today, agency ordinarily implies either an explicit or implicit contract. The agent is hired by the principal. Since the entrepreneur function does not require a contract, Knight used a broader definition of "agent."

Knight stated the conclusion we can draw from this approach. He said that the only legitimate basis for supporting the market economy is our intuitive belief that the self-sorting and voluntary "risk-bearing" will cause what consumers regard as wealth to be produced.

The authors referred to here were aware of the fact that their image of the market economy did not include many important phenomena of the real economies in which they lived. Besides politics and culture, it disregarded externalities and certain goods that yield general public benefits or harms. The latter phenomena were usually described in an addendum. Public goods are not important for our purposes, since we are concerned at this stage with constructing an image of the supply of private goods through the market economy. We discuss external effects below.

An Example

We can show how the market economy provides an incentive to act entrepreneurially with an example. We begin with a situation in which there are private property, specialization, and the use of money as a means of acquiring private goods for consumption. A prospective producer suspects that he can earn a profit by producing an item and then selling it. He believes that the item is superior to some existing good in satisfying consumers' wants. He believes that it will have either a lower price or superior want-satisfying capacity in the eyes of paying consumers than its alternatives. At first, he lacks both the funds to finance production and the knowledge needed to convince financiers (partners, stockholders, lenders) that enough consumers will pay a high enough price to cover his costs plus a reasonable financial return on their investments. He judges that the best way to find out whether there is a sufficient effective demand is to personally finance the production of a few units so that consumers can sample them. If he confirms his suspicions, he can then present his findings to the prospective financiers.(2)

He must also give the financiers an estimate of production costs. To do this, he may have to conduct a survey of the current prices of various factors. To be convincing, he may have to secure the testimony of a trustworthy business consultant. He may also need to discuss his business with an insurance agent and obtain an estimate of premiums and terms of coverage.(3) His aim is to provide the financiers with information to make their own separate, individual, subjective, estimates of the prospect for making a profit or loss. He knows that the financiers will not agree to finance until they feel that their appraisals are adequate.

Consider the producer's actions more closely. He begins with a preliminary appraisal of factors -- a vague notion. This notion is enough to convince him that it is profitable to invest in more information to make more accurate appraisals. At some point he decides that it is profitable to invest in trying to convince financiers to help him carry out the project.

For the production-consumption sequence to occur, the producer, financiers, factor-suppliers, and consumers must reach a point where they are willing to bear the uncertainty connected with their actions. Each of them must make sufficiently high projections of their respective gains and/or find someone else to bear the uncertainty, typically at the expense of lower gains to themselves.

The height of each one's projection of gain depends on the existence of rights to obtain rewards for one's actions. The producer must be able to obtain the right to the residual that results from his employing factors and selling the goods. Financiers must be able to acquire the right to future repayment of loans or possession of any guaranty pledged by the producer. The consumers must be able to acquire the right to consume a promised or actual good. And employees must be able to acquire the right to be paid for the factors they supply. Thus private property rights enable the otherwise unrealizable and, indeed, unknown gains to occur.


So far, we have discussed a single production-consumption sequence. The market economy contains uncountable sequences of this type. The new subjective value theorists simplified this complex interaction by creating the personage of the pure entrepreneur and the image of the pure entrepreneurial economy. The imaginary pure entrepreneur does all the appraising,(4) makes all the production decisions, and bears all uncertainty. In this image, the financiers are merely savers, who receive uncertainty-free interest, the consumers receive consumption benefits and pay money, and the factor-suppliers supply factors and receive incomes. The non-entrepreneurs are like robots who behave according to algorithms. This image of entrepreneurship helps us isolate the incentive to participate in production-consumption sequences by placing all of the distinctly human decision-making in the mind of a distinct personage. Thus, modern subjectivist economics speaks of the incentive to produce wealth as the incentive to act entrepreneurially.(5)

In this image, the entrepreneur must anticipate sinking some of his own wealth into the project. For one thing, he must assure himself that his initial suspicion is correct by acquiring more information. For another thing, he must provide guaranty to the savers and possibly to the factor-suppliers and consumers.(6)

Now imagine many pure entrepreneurs in various lines of business bidding for loan funds, factors, and consumer spending. Competing entrepreneurs base their bids on their respective beliefs about the worth of the factors. In this "pure entrepreneurial economy," the prices of the factors and interest rates reflect the highest bids of competing entrepreneurs.

No single entrepreneur will anticipate a profit unless factor prices and interest rates are low enough. He will not take ultimate control over a factor unless he believes that he has a more profitable use for it than the other entrepreneurs believe they have. The others, by deciding not to bid still higher for the factors, show their beliefs that consumers of their products will not pay a high enough price.

This isolation of the role of the entrepreneur enables us to see clearly three aspects of entrepreneurship that give us great confidence that individuals in the market economy will produce what individuals acting in the consumer role regard as wealth. The first is the profit incentive. Individuals have an incentive to learn about opportunities to profit from satisfying others' demands for goods. The second is cooperation. Individuals share their knowledge to the extent that is necessary in order to induce the cooperation of others in joint production projects. They inform each other, within limits, about the potential gains from production and exchange. In addition, they offer others opportunities to earn more income from their savings or investments, opportunities for employment, and opportunities for consumption. Third, it helps us understand competition. Each pure entrepreneur must bid against other entrepreneurs for loan money, for factors, and for consumer spending.

A private property system and a regime of free trade give individuals an incentive to discover opportunities and to bet that their appraisals of factors are more correct than those of other individuals. In addition, with the two exceptions described below, the bet of any single appraiser is a bet that the consumers to whom he plans to sell value the resources more than the consumers of other products. If we add the reasonable assumption that the bettors are likely to be right more than they are wrong, then we have shown that the system of private property rights gives individuals the incentive to discover and use opportunities to produce what the consumers of their products regard as "wealth."(7) We can now consider two qualifications.

Dampened Incentives

The incentive to discover and take advantage of such opportunities is dampened by two factors. First, there is a prospect for being copied. Anyone who discovers another entrepreneur's prospective profit-making opportunity may copy it. Patents, copyrights, a common respect for "intellectual property," and other arrangements may help overcome this problem; but they cannot fully overcome it. Besides, these legal sanctions are costly to enforce. Second, perfect price discrimination is hardly ever possible. Some of the gain must be shared with consumers (and factor-suppliers). Thus, the profit incentive is not as strong as we can imagine it being. Due to copying and the necessity of pricing below consumer valuation, the true value to consumers of an entrepreneurial action can be substantially higher than what the consumers actually pay for the performance of the action. Anticipated profit may fall substantially below the gain to consumers.

External Effects

In addition, when we speak of the wealth created by actions, we must consider the prospect that the actions will have effects for which actors do not account -- external effects, such as pollution and congestion. External effects could not exist if there was a complete set of private property rights, one that would lead each individual to account for all of the harm and benefit due to his action. However, a compete set of private property rights cannot exist in reality. On the other hand, we must keep in mind that individuals have an incentive to create property rights when they consider the external effects to be "large." To assess wealth-production in any specific case, we must account for external effects and the creation of property rights.(Gunning 1994: chapter 7)

3. Incentives to Supply Public Goods

The predominate approach of professional economics to the public goods problem has been to associate it with market failure. Market success is defined in terms of efficiency which, in turn, is defined in terms of general equilibrium models of the market. General equilibrium models involving only private goods are compared with general equilibrium models involving both private and public goods. Typically, this approach amounts to nothing more than comparing mechanical models of the Walrasian type, in which the individuals are assumed to be robots. In the more revealing analyses done by "property rights theorists" and "new institutionalists," the individuals are endowed with innovative and creative abilities that an economist may be unable to specify. Although these analyses are far superior to the comparisons of mechanical models, they do not take advantage of the increase in explanatory power achievable by using the modern subjectivist concept of the entrepreneur. To extend this approach to the study of public goods, we must define the public goods problem as entrepreneurship's failure to discover and produce public goods for consumers. A public goods-based Public Choice would focus on how members of a collective disposed toward democracy could either (a) improve the performance of entrepreneurship in a market economy that contains goods with public goods characteristics or (b) provide economical substitutes for entrepreneurship, presumably through government bureaucracies. The first step would be to tell why and how public goods characteristics interfere with the incentive of entrepreneurship to discover and produce public goods. This is our next step.

The Appraisal and Production of Public Goods in the Market Economy

Economists ordinarily define public goods as having two characteristics that distinguish them from private goods: (1) jointness (non-rivalry) and (2) nonexclusion (non-payers cannot be excluded). Jointness means simply that many individuals can enjoy the good simultaneously; there is no extra cost of adding a consumer. Non-exclusion means that consumers can avail themselves of a produced good without having to obtain permission. A producer-entrepreneur who produces a good cannot exclude non-payers from partaking in the consumption. The hypothetical pure public good yields simultaneous benefits to all consumers.

In everyday life, there are probably no goods that resemble the pure public goods of economic theory. Accordingly, the public goods problem in everyday life is mainly concerned with situations in which public goods characteristics seem to be important. Since our main concern liis with setting up the theoretical problem, we shall merely follow the standard theoretical practice of assuming that pure public goods exist.

The "public goods problem" mainly concerns goods that have not yet been produced. For such goods, a prospective producer can always exclude all consumers by simply deciding not to produce. He can demand payments of various amounts from different consumers before he begins to produce. The problem arises because we assume that self-interest leads at least some prospective consumers to conceal their preferences from each other, to drive hard bargains withother consumers, or to reject offers to contribute because they anticipate that the benefits of free riding exceed the benefits of contributing to supply.(8) Because of the prospect for concealment, hard bargaining, and free riding among consumers, entrepreneurs have insufficient incentives to produce pure public goods even though they believe that the sum of the consumers' gains is greater than the marginal cost. This implies, in turn, that entrepreneurship would have insufficient incentive to appraise the items that could become factors in the production of the pure public-goods.

The prospect for concealment and hard bargaining also reduces the incentive to produce private goods. The difference is that in the case of a private good, a producer typically faces many consumers. As a result, he has an incentive to search for ways to reduce concealment and hard bargaining -- for example, by offering nondiscriminatory take-it-or-leave-it prices. The prospective producer of a public good also has an incentive to promote organization among consumers in order to help them find a solution to their concealment, hard bargaining and free rider problems. However, the consumers themselves must solve their problems multilaterally. The producer's means are limited. The requirement of a multilateral solution is what differentiates the two cases.

A public goods' based Public Choice begins by assuming that members of a collective believe that some public goods and factors will not be discovered and some discovered public goods will not be produced in the collectively desired quantities unless appraisers and producers are provided with special incentives.(9) It then explores what members of the collective could do to provide the special incentives.

Two kinds of Incentives

In describing how members of the collective might view the incentive problem regarding the supply of public goods, it is useful to distinguish between two extreme, or ideal, classes of not-yet-produced public goods. The first class consists of goods that everyone knows to be public goods. The second consists of goods that no one currently knows are public goods. Consider each in turn.

Known But Not-Yet-Produced Public Goods

At first glance, it seems that if members already know that an item is a public good and the costs of producing it, they can hold a meeting, agree to a tax plan, and set aside the money needed to pay the producer. However, they face many types of uncertainty. First, individuals would be uncertain about each others' preferences for the public good. Second, they would face uncertainty about the will of each individual to hold out for a better deal than that offered in a particular collective agreement. Third, individuals would be uncertain about the costs of production. Thus, they would not know whether one or another production method is cheaper to use. Fourth, if they reach a collective agreement to pay for the good, they would face uncertainty about whether each member would keep his promise to pay his specified share of the price. Finally, all would face uncertainty about whether the supplier will actually supply the public good according to the agreed terms.

Members can deal with some of these problems at one level but only by introducing uncertainty at a different level. For example, to solve the problem of a lack of information about costs, the collective could employ an agent to solicit bids from prospective contractors. They would be uncertain, however, whether the agent was doing his job according to the directions given him and whether the bidders had colluded. For another example, the collective could employ an agent to collect the money from the members at the time of the meeting and then hold it in escrow until the public good is produced. They would be uncertain, however, whether the agent would take due care to protect the money and earn the highest market interest.

Unknown Public Goods

Now consider items that are not known to be public goods. Assume that members of the collective believe that the benefits they can collectively derive from some as yet unknown public goods would be greater than the opportunity costs. One way for the members to provide an incentive to discover and produce such goods is to promise to pay compensation to a discoverer after the discovery is made. Such a promise is unlikely to be convincing, since a prospective public goods entrepreneur would have to rely on the good faith not of a single person but of the individuals who are designated to represent the collective. Another solution is to hire an agent to determine the reward, but this introduces additional uncertainty.

An example might be helpful. We can imagine a future in which the technology will exist to alter the path and change the intensity of hurricanes and typhoons. Many individuals would benefit from the discovery of such a technology. At some future time, a single individual may achieve a position where he finds it worthwhile to invest the time and money needed to appraise the investment in such research. Sooner than that, a consortium may form to raise funds for the appraisal. In the near future, term, however, the member of the collective may feel that they cannot rely on private initiative. So they may agree to consider ways to reward the research after it is done.

It seems likely that no public good fits neatly into one of these classes. The value of distinguishing them is that it helps us recognize that members of the collective demand agents not only to supply goods but also to make judgments in their stead.

A Note on Subjectivist Methodology

Subjectivism requires that we endow members of a collective with analytical abilities that are at least no less than we assume ourselves to possess.(10) This implies that we should assume for the sake of realism that members can recognize that if entrepreneurship had the same kinds of incentives to discover and produce public goods as it has for private goods, the members would be better off. The question is whether the members will believe that they can cause these incentives to be provided at a sufficiently low cost to be worthwhile. Thus, the fundamental problems we face in Public Choice, as defined here, are those of (1) identifying the means of providing incentives to produce public goods that members of the collective in a market economy would think up and (2) determining how the members would view the benefits and costs of employing these means. Thus we must make judgments about their perceptions and understandings.

Experience seems to suggest that members of a large collective would think that the benefits of supplying some local public goods are greater than the costs. Whether they would think the same about less local public goods is difficult to determine. One thing seems certain. The ideal pure public good does not exist, although some members of the collective may think that it does. Whatever opinion we might have about this matter, if we want to get on with the project of understanding democracy from the public goods point of view, we must assume at a theoretical level that members share a belief that "national" public goods exist and that the benefits of having them supplied are greater than the costs. Then we can ask how the members might go about causing the public goods to be produced. If we later judge that national public goods are not important enough to be worth considering, then our analysis will be irrelevant to everyday life. Nevertheless, it will still be relevant to the task of evaluating the logic of a popular ideology. This ideology states that if national public goods are sufficiently important, it is worthwhile to use democratic institutions to remedy the market failure that is deducible from the theory of public goods. By trying to deduce the constitution that individuals who hold this ideology would construct, we can better judge whether the ideology is reasonable.(11)

4. Creating a Democratic Government: the Constitution

Preliminary Considerations

Up to this point, the discussion has focused on a single public good. This approach has helped us form a preliminary understanding of the basic concepts of (a) incentives and (b) agency in the supply of public goods. We now must expand the discussion to cover many public goods. Imagine that members of a collective want to cause incentives (1) to supply many (national) public goods and (2) to appraise many possible investments in public goods research. The supply of some public goods would presumably be a continuing activity, as would the appraisal of new investment opportunities. As a result, we assume, the members would want to employ agents to be part of a standing government with specialized departments, or ministries.

Under these conditions, supplying more efficient quantities of public goods appears feasible only if the members employ some agents to hire other agents. We call the former politicians and the latter bureaucrats. The collective would presumably aim to devise some means of providing the politicians with incentives to hire, supervise, and monitor bureaucrats in the interest of the collective. A contingent contract -- containing renewal procedures and possible impeachment -- seems optimal. But the members of the collective would be uncertain whether the politicians would truly act in their interest. Politicians could shirk, abuse their power to enrich themselves, or be swayed to systematically act in the interests of some members but against the interests of other members (rent-seeking).(Buchanan 1980: 7-8)

The problems of shirking and power abuse are present in all relationship that involve agents. However, these problems, combined with rent-seeking, may have especially serious consequences for members of the collective because the politicians must be given control over the presumed monopoly of force. The politicians must possess this monopoly because they must have the capacity (1) to enforce the collective agreement that establishes the government, including the collection of taxes, (2) to supply national defense, and (3) to supply the deterrent of last resort against violation of private property rights.(12)

The potential damage due to shirking, power abuse, and rent-seeking is compounded by the fact that actions aimed at reducing these activities or otherwise dealing with them are themselves public goods. If a member of the collective expends effort to discover shirking etc., many members of the collective are likely to benefit jointly. Also, the producer of such information does not want exclude other members since, once the information is produced, its value depends on the cooperation of other members in devising means of correcting the problem. Although some market mechanisms exist to help overcome the free rider problem in supplying information -- e.g., advertising that supports investigative media reporting -- there is still a problem of enforcement. Even if everyone knows about an abuse or bias, someone must take the initiative (bear the cost) to reform the system or fire the compact-breaking politicians. These actions also have public goods characteristics.

Thus, we arrive at a position where we conceive of a market economy in which each individual is a member of a collective looking ahead to a future that contains a standing government. The members envision hiring the top officials of that government to make decisions on national defense, to enforce property rights, to cause the supply of other national public goods, and also to hire and monitor bureaucrats.

We can define a constitution as the terms of the agreement, or contract, among the members of the collective under these circumstances. It would apparently contain two general classes of specifications. The first are rules concerning which politicians will he hired and how they will be hired. The second are rules governing the actions of the politicians and bureaucrats. We consider each in turn.

Rules for Hiring Politicians

A Representative Legislature

Before hiring politicians, a decision would first have to be made about the hierarchical structure of the agents who are elected. It would be possible to hire a single politician and make him responsible for hiring all of the bureaucrats. Doing this would entail the least cost of conducting the election. Moreover, if there was only one elected politician, the information needed by the community members to evaluate alternative candidates would be minimized. But members of the collective would presumably not want to entrust any single individual with so much power. Instead, it seems reasonable to assume that they would envision dividing the power to make decisions by selecting a legislature.

A representative legislature is a group of politicians elected by the collective to represent them. Assuming that the members of the collective agree to have a representative legislature, they would have to decide how many legislators they want to elect, how long they should serve, and how they would be elected. Logically, community members would want enough legislators to make the costs of collusion among them high. Yet the more legislators there are, the higher the cost of maintaining them. In between the two extremes is a wide range of possibilities. Regarding length of service, the length should be long enough that representatives would have sufficient time to learn the job. But it should not be so long that it substantially increases the risk of collusion. Term limits of one or two terms seem a logical means of limiting the collusion risk. Another way to limit the risk is to stagger the election times. For example, if the usual term is three years, one-third of the legislators might be elected during each year. Yet another way is to make the formation of political parties less beneficial or more costly.

Regarding the method of choosing the representatives, there seem to be two extremes: a single, nationwide constituency and voting districts. Consider the nationwide constituency. A 90-seat legislature might be filled my means of a staggered national election for 30 candidates each year for a term of three years. One hundred fifty candidates might stand for the national election out of which thirty must be chosen. As opposed to this, the nation might be divided into thirty voting districts, each of which has the task of electing one representative. In this case, five candidates might stand for election in each district out of which one must be chosen.

The latter method would seem to economize on information and monitoring costs. Voters who live in the same district as the legislator are likely to have more information about how a candidate will perform as a politician. Thus, the vote for or against a given candidate is less likely to be polluted by outsiders who know little about the candidate. Moreover, since the electee will be expected to represent the constituents, they have a greater incentive to produce information about candidates' potential for service. And, given the power of impeachment, they have a greater incentive to monitor a politician's activities once she is in office. On the negative side, the representative of a district is not likely to be concerned with providing benefits for all of the nation's members. Her prospect for election depends on the judgments of her electorate. It follows that dividing the nation into voting districts gives electees an incentive to favor laws that benefit their constituents regardless of the harm such laws may cause to the constituents in other voting districts.

There are only two obvious solutions to this conflict. The first is to elect some legislators from a national constituency and some from voting districts. The second is to use the voting district method but to constrain the types of legislation that can be introduced. Members might require all legislation to pass a test to wit: benefits must be of a general nature and/or only beneficiaries can be taxed. Such a law might be supported by creating a relatively independent auditing bureau, whose task would be to make a determination of the identity of the beneficiaries and payers and to estimate of the amount of benefits and costs to each.

Election Procedures

How would politicians be hired? Each member of the collective would presumably have an incentive to persuade others to hire a politician who he prefers. To avoid potentially endless bargaining, members would want to make a prior agreement to abide by the outcome of a vote -- an election. To do this, they would first have to establish some rule for deciding how to choose candidates and which candidate would be declared the winner.

Establishing a voting rule means that each member agrees to hire those candidates who satisfy the voting rule criterion. For example, if they agree to hire candidates who win a plurality in a popular election, someone who voted for losers must still pay his designated share of the tax cost that is decided by the winning candidates.

There is no need here to go into detail regarding the possible methods for picking candidates and deciding elections. Experience suggests that a variety of selection methods may be used. Candidates can be picked by political parties, through independent nomination coupled with endorsement by a given percent of the population, or in some other way. Candidates can also be required to meet certain minimum qualifications such as residency, age, and education. There are also several methods of election. For example, a candidate may have to win a majority of votes in order to be elected; or she may only have to win a plurality of votes. Eligibility to vote in elections may be attenuated by restrictions on age, mental competence, and the nature of one's previous actions (e.g., criminals and subversives can be excluded). Finally, members would presumably want an impeachment procedure and a means of calling for a new election when there are irregularities. There are many possibilities. The only point we want to make here is that members of the collective would want to specify these rules for hiring politicians in the constitution. The main reason for this is to avoid costly meetings of the collective. Such meetings would be required in order to establish rules each time that the members wanted to have a new election.

Rules Governing the Actions of the Politicians and Bureaucrats

Enforcement of Private Property Rights

Members would want strong protection of the private property rights that existed at the time the constitution was made. Not only would each member of the collective would want to protect his own property against arbitrary seizure, he would also want to protect the market economy. For the same reason, he would want to strictly limit the power to tax. Especially significant is the fact that taxes reduce the incentive of individuals to engage in the taxed activities. Of course, the collective could not limit the taxing and subsidizing powers so much that they hampered the politicians' abilities to provide incentives to cause the supply of desired public goods. There is a tradeoff.

In addition, members of a collective could provide encouragement for the market economy by agreeing to use the monopoly over force to be used to enforce contracts. This would allow individuals to voluntarily subject themselves to jointly-determined penalties for actions that are in breach of a contract. Of course, members would want the parties to the contract to pay for any expenses incurred by agents.

Members may want specific provisions to correct for the reduced incentives due to copying (patents) and for externalities if they have reason to believe that the benefits of such activities are likely to exceed the costs. Or they may wish to turn the job of making laws regarding these things over to legislators.

Tax Rules

This is not the place to discuss alternative tax systems. It is perhaps enough to say the following. First, under normal circumstances, the collective would want a system that entails minimum market uncertainty. This could be partly accomplished by a rule that prohibits the sudden imposition of special taxes on some goods. In the interdependent market economy, changes in taxes always have an impact that extends beyond the subjects who pay the taxes. A constitutional rule could be made requiring all tax bills to be accompanied by a study that specifies the probable incidence of the tax so that discriminatory impact of taxes on any particular subset of the collective could be identified and avoided. A rule against taxes that are too discriminatory might be passed. If such rules were practical, individuals' uncertainty due to prospective taxes would be less than otherwise. The same principle would apply to the inflation tax, if members understood it. Second, because different individuals benefit differently from public goods, the collective would presumably want them to be taxed differently -- according to the benefit principle. Third, because tax collection is costly, members of the collective would presumably want the least expensive system. These objectives conflict, of course, and members of the collective would presumably recognize this fact. How they would deal with the conflicting objectives is difficult to say.

Rule of Law

Democracy has two essential characteristics: representation and the rule of law. The rule of law assures that the exercise of force does not depend upon the identity of the people who control it. Precisely what its exercise does depend on is not so easy to specify. The rule of law appears to have developed in England as the King sent out jurists to give citizens their first right to appeal the sanctions imposed by local lords on the citizens under their charge. As time passed, the principles that were thought to be contained in early decisions were applied to later decisions. Eventually a body of teachable principles evolved which helped to influence future decisions and gave appellants a means of predicting the outcome of an appeal to those principles. A most important characteristic, as the system developed later, was the right to appeal an administrative action or lower court decision to a higher court. In the U.S., the right of appeal was extended to rights that were specified in a written constitution.

In practice, the rule of law in the U.S. has meant that the exercise of powers by the representative legislature and the elected chief executive were subject to judicial review according to a set of principles derived from statements in the constitution. The ultimate judicial decisions are in large measure predictable, implying a tradition and continuity.

Members of a collective who are contemplating forming a government would surely want the rule of law and the independent judiciary that it implies. But they may also want to give some guidance to the judges. They could do this by establishing a written constitution. Specific rights to protect members against discriminatory and discretionary law making and enforcement could be included in a list of rights of the people. Among other things, they would want to include respect for private property rights and limits on the taxing power. In addition, they would want such things as equal treatment under the law, the right of habeas corpus, trial by jury, and a speedy trial. Additional elements are discussed below.

Separation of Powers

Another means of controlling abuse is a set of checks and balances. This can be achieved by separating the functions of making laws, enforcing laws, judging guilt, and administering penalties. Thus, an executive, legislative, and judicial branch of government would seem to make sense. It might also be useful to have a separate branch whose duty is to search out abuses in other branches.

The rule of law, enforced by an independent judiciary, implies some separation of powers. Moreover, a representative legislature in which legislators are elected from voting districts also implies a kind of separation of powers -- or, more correctly, a dilution of power. Thus the additional division of the government into branches may be unnecessary. This is suggested by the durability of the British system, in which the representative legislature chooses the chief executive from among its members and which has no constitution or independent judicial review of laws.

One of the functions of the separation of powers in the U.S. and other presidential systems is to require a greater than simple majority of voters in the legislature to pass a bill. The president can veto a bill. When the president and legislative majority are of different parties, this ordinarily amounts to requiring a supra-majority to pass a bill.(13) We discuss the desirability of supra-majorities below.

Rules Encouraging the Supply of Information

The employment of politicians and bureaucrats introduces new levels of collective uncertainty. It seems to follow that members of the collective would want to encourage the supply of information about public goods and about the actions of agents hired to cause them to be provided. They would want to assure that the monopoly over force was hardly ever used to inhibit freedom of communication both between members of the collective and between any member of the collective and an agent. Also, they would want agents to keep the collective informed about their activities. Accordingly, they are likely to want politicians and bureaucrats to make periodic reports and to keep records of their activities that are available to all. They would strongly oppose secrecy among politicians and bureaucrats.(14)

Searching out and concealing information is like a game. An agent who acts improperly tries to conceal his improper actions from the members of the collective. To do so, he makes judgments about the methods that members of the collective might use to discover the information he aims to conceal. Then he develops strategies based on his judgments. At the same time, the member who wants to discover the information develops strategies based on her judgments about the means an agent might use to conceal information. To help her keep accounts, she may record her judgments, strategies, and previous actions. When this occurs, she wants to conceal her records from the agents. Since members of the collective want to encourage members to find out information about agents, they would ordinarily want to prohibit agents from seizing members' private papers and communications. The exception is when there is "due cause" for believing that such a seizure would help the agents discover a violation of some provision of the collective agreement. Thus, a constitutional provision that protects a member from unreasonable searches and seizures in a constitution seems reasonable.

Rules to Control the External Costs of Collective Decision-Making

Perhaps the centerpiece of Buchanan and Tullock's theory of a constitution is the idea that each member of a collective would anticipate that unless he had veto power himself, there is some probability that the decisions made by other members of the collective would go against his interest. They called the damage associated with this possibility the "external costs of collective decision-making." According to B-T, individuals would willingly incur some of these costs in order to avoid the still higher costs that are associated with obtaining the consent of everyone in order to pass a bill. So each person would accept a majority decision-making rule even though she knew that to do so would raise the probability that collective decisions against her interest will be made. She may not agree to a 50+% majority, but she would agree to some majority rule under 100%.

B-T used the same principle to explain why members of a collective would opt for a representative legislature over a general assembly of all citizens. The costs of collective decision-making (i.e., the costs of assembling all the members at one place and time) would be too high relative to the expected savings of external costs of collective decision-making.

In the necessarily simple models that B-T used, they imagined that there would be an optimal majority rule -- i.e., some percent of the majority required for making a law. This rule would presumably be put into practice in the legislature. Consider a rule that requires a seventy-five per cent majority of legislators to pass a bills. Under this rule, the costs of collective decision-making to the members of the collective would be higher, but the external costs of collective decision-making would be lower. By selecting the optimal majority rule, according to B-T's simple models, it would be theoretically possible to achieve the optimal point in the tradeoff for a given legislature.

I have used the term supra-majority to refer to more than a simple majority. Continuing in this use, we can say that members of the collective would want to control the size of the majority required to pass a bill. They could do this for a single-branched, single-legislature government by making a constitutional rule that determines the supra-majority rule required for the passage of a bill in the legislature. For a system that has both a legislature and executive branch, they could give the president veto power. Or they could create a second legislative branch and demand that bills receive a majority of votes in both houses. These latter forms are less certain to achieve the stated objective than a simple legislative house with a single supra-majority, since all elected branches may come to be controlled by a single party.(15)

Constitutional Change

The members of a collective may want their constitution to contain provisions for constitutional change. They may look ahead to a time when they die and are replaced by their offspring. Or they may consider uncertainty and the corresponding possibility that they will regret having included or not having included some existing provisions. Given the costs to members of the collective of having meetings, they may wish to avoid considering (or making) such changes unless they are sufficiently "important" to a large enough group. They may even want to delegate the authority to make constitutional change to agents elected by various geographical (or other) sub-divisions of the polity. However, since a free market economy, the free flow of information, and rights to property and person are so fundamental, the members are unlikely to allow changes to occur easily. They may wish to retain ultimate control by requiring the approval of two-thirds, three-fourths, or some greater majority of original founders or heirs. This could be achieved by requiring all constitutional change to be submitted for approval by a supra-majority referendum. Alternatively, members of the collective may delegate constitution-changing power to elected legislators but require a large supra-majority in the legislature.(16) An aid to the constitution-changing process could be a special body elected for the sole purpose of exploring the effects of constitutional amendments and recommending changes.

5. Conclusion

The aim in this paper has been to follow the path of development for Public Choice that Buchanan and Tullock suggested over thirty years ago but did not take. We explored in rough fashion the problem that individuals in a fully functioning market economy might face in providing incentives to produce public goods. We argued that because of the jointness characteristic of public goods, the incentive problem must be handled collectively. Modeling our inquiry after that of Knight, we found that the fundamental problems faced by the members of the collective are (1) hiring the agents needed to appraise public goods investment opportunities and (2) controlling the hired agents. We pointed out that members of a collective who wanted to provide incentives to produce any "important" public goods would find it sensible to institute voting and elections in order to economize on collective decision-making costs. Also, the members would want to have strict "constitutional" rules to constrain actions that interfere with the market economy, to encourage the supply of information, and to limit agents' shirking, abuse of the monopoly over force, and rent-seeking.

It is, of course, uncertain that beginning with a market economy, anyone (or small group) would think that there is an incentive problem concerning public goods or, if they did, that they would believe that it is worth solving. In other words, it is uncertain whether individuals would think that it is profitable to create a democracy or a constitution for reasons concerning public goods. For one thing, the pure public good that is so important to economists for illustrating the jointness and exclusion problems does not exist. There are many club goods, local public goods, and spillovers from the local supply of public goods. But we have not in this paper discussed these kinds of situations. It is surely not true that today's democratic nations and their constitutions were created under conditions of peace and solely for the purpose of producing pure, or even "national" (whatever one takes that to mean) public goods.

The idea of supplying public goods was probably quite a distance from the minds of the founders of most of the modern democratic constitutions. In the U.S., it seems likely that the dominant considerations were the desire by particular individuals (though not all) to avoid being the victim of future aggression, to avoid a war between colonies, to aggress against the natives (or to protect against aggressions and reprisals), and to receive money promised as repayment for revolutionary war spending. With the possible exception of national defense, these all fall into the category of property disputes, not public goods problems. Accordingly, it would be quite incorrect to think that the Public Choice theory presented in this paper can provide the basis for a complete historical explanation of modern democracies and constitutions. In some cases -- e.g., Germany, India, Japan, The Philippines, South Korea, and Taiwan -- the theory provides no explanation at all. Each of these cases is different but the theory that democracy is a logical extension of an ungoverned, free market economy because of a public goods problem is not very helpful.(17)

Buchanan and Tullock avoided the difficulty of confronting reality by not following up their initial definition of the new "Public Choice." Instead of a theory of collective decision-making based on the theory of public goods, they constructed a theory based on the actors' perceivedoutcomes of a "collective activity."(Buchanan and Tullock 1962: 33-5) They assumed that each actor ranks prospective "collective activities" in relation (a) to private action and (b) to market interaction. Thinking about having to make many future decisions on collective activities, individuals contemplate the effects of binding themselves contractually to a "decision-making rule."(ibid.: Chapter 5) This rule has what B-T regarded as one dominant characteristic: the number of people required to reach agreement. Anticipating that many decisions may be made against them, individuals come to favor a rule of greater than simple majority for making decisions on collective activities. The problem with this approach, from the standpoint of this paper, is that B-T gave actors no insight into the "public goods problem" and what they might do to solve it.

By avoiding this issue, B-T changed what they started as a theory of the demand and supply of public goods into a more general theory of constitutional democracy based on potential harmful negative externalities associated not with public goods but with collective decisions. In so doing, they deflected their attention away from the agency problem. As a result, the relevance of their discussion was limited to institutions concerned with making collective decisions.

We have seen in this paper that by focusing on the agency problem, we can construct a different explanation of constitutional democracy. The one suggested here explains not only a greater-than-simple-majority rule for constitutional change but also specific provisions of a constitution. These include voting and elections, respect for private property rights, respect for privacy, limitations on government powers, checks and balances, the rule of law, freedom of information, contract enforcement by the judiciary, rights of defendants in criminal cases, and special rules for constitutional change.

There have been many constitutional democracies but only a few qualify as having stood the test of time. The survivors seem to have had a large measure of the characteristics described here, while the dead democracies did not. On the other hand, it would be folly to suggest that the purely theoretical economic explanation can by itself explain the characteristics of surviving constitutional democracies. To explain these, we would have to know in greater detail the particular circumstances that prevailed in each case. Most importantly, we would have to describe the relative ability of a democratic government to win a war and suppress a rebellion, since these appear to be the most immanent threats to a democracy. We would also have to consider, of course, the prospect for maintaining the market economy in the face of challenges from special interests. Finally, we would have to consider psycho-cultural factors.

Due to space limitations, this paper has focused entirely on one class of agents -- elected politicians. A more thorough analysis would explore the making of rules relating to bureaucrats also.


1. One might argue that no society has ever developed in this way. How could a society have a market economy without private property rights, one might ask. And how can a society have private property rights without a government? One way to justify the approach is to assert that the tasks involved in defining and enforcing property rights are logically separable from the tasks involved in causing public goods to be supplied. Although they may be logically separable, the two activities have some tasks in common, such as tax collection and disbursement, the employment of agents, and political overseeing. Thus, some of the propositions that will be derived in this paper can be applied to the job of building a theory of the minimal state. For the theory developed in this paper to be completed, however, it would have to be joined with the theory of the minimal state.

2. The particular procedure he uses to reduce his own and his financiers' uncertainty about demand is immaterial. It is only important to assume that he finances the procedure himself, since the ultimate idea we want to convey is that he must make a bet that his appraisal of some factor (in this case the factor used to produce the samples) is superior to that of others.

3. Again the particular method he uses is immaterial.

4. Appraising in the most general sense refers to identifying means of satisfying wants and attaching a monetary value to them. The "means" may be factors of production, a method of combining factors, or a final good (before it is consumed).

5. This more abstract view of entrepreneurship is mainly due to Ludwig von Mises, 1966. For a more complete discussion of the economist's procedure, see Gunning 1994 and 1997b.

6. Guaranty to factor-suppliers can be in the form advance payments. Guaranty to consumers can be in the form of rental contracts that guarantee consumer satisfaction with a promise to make compensation in the event the consumer is not satisfied.

7. It is worth pointing out the predominant difference between this analysis and one that neglects entrepreneurship. In the latter, we assume that the factors and methods of production are already known. The incentive problem refers to the incentive to employ the factors and methods efficiently. It is an allocation, or rationing, incentive. In a proper entrepreneurial analysis, the incentive problem includes identifying and appraising factors, which of course includes identifying preferences. Another difference is the assumption of intersubjective uncertainty.

8. The hard bargaining need not be direct. It is sufficient that a consumer chooses not to pay a higher price because she believes that if she does not agree, other consumers will pay more. The consumers need not meet face-to-face. They may simply respond to a producer's tentative offers.

9. This statement does not imply that members of a prospective collective could easily determine what kind of situations fit into this class or that, if a situation fits into this class today, it will also fit tomorrow. It is a statement in pure theory. The term "collectively desired" refers to the aggregated independent desires of the members of the collective.

10. This statement seems to require no explanation. Nevertheless, in professional economics assumptions to the contrary are often made without explanation. For the reasoning involved to reach this conclusion, see Gunning, 1994, chapter 1.

11. The idea that evaluating ideologies is the ultimate purpose of economic thinking was described and defended by Mises, 1966, p. 178-187.

12. We could imagine a case where the use of force was competitive and members' principal purpose in forming a government is to collectivize the use of force. If we further assume that the members appreciate the importance of using the monopoly to enforce the property rights needed to promote an integrated market economy, the reasoning about a government formed under this assumption would proceed no differently than it does here, except that at the outset, the agreement would have to contain a specific designation of property ownership. This is because without the collectivization of force, no clear inter-regional property rights would exist. Given our assumption that Public Choice begins where the market economy ends, we proceed from the assumption that private property rights and a monopoly over force already exist.

13. A second house of Congress adds to this, since not only may the majorities in the two houses disagree, different political parties may have a majority in each house. It is also true, of course, that some congresspersons are not affiliated with the major parties.

14. There are three exceptions. The first is war or the threat of war. Members of the collective would not want agents to reveal war and anti-espionage strategies The second is detection of crimes (for example, where the robbery detection devices will be employed on a given day). The third concerns cases where government activities require knowledge of a business's trade strategies. Members would not ordinarily want rivals to have access to information gathered by government agencies about trade secrets, since this would reduce the incentive to develop them. The government's gathering of such information should be restricted, however, as indicated below.

15. Political parties are not the only organizations that may try to gain control. Pressure groups, especially large ones, may have opportunities to control more than one branch. A discussion of pressure groups is beyond the scope of this preliminary analysis.

16. It may also be reasonable to require different majorities for different types of proposed changes.  

17. One could develop a public goods theory of democratic change. Such a theory would begin with a specific constitution that was imposed by an authoritarian regime or a victor in war. Then it would suggest ways that eligible voters could exercise their rights under the constitution to recommend changes in laws and in the constitution. Such a theory would in many respects be similar to the theory described here, except that changes would be constrained by the initial constitution.


Buchanan, James M., Cost and Choice: An Inquiry in Economic Theory, Chicago: Markham, 1969.

Buchanan, James M., "Rent-Seeking and Profit-Seeking," chapter 1 in James M. Buchanan, Robert D. Tollison, and Gordon Tullock (eds.), Toward a Theory of the Rent-Seeking Society, College Station, Texas: Texas A&M University Press, 1980.

Buchanan, J.M. and G.F. Thirlby, L.S.E. Essays on Cost, New York University Press, 1981.

Buchanan, James M. and Gordon Tullock, The Calculus of Consent, Ann Arbor: University of Michigan Press, 1962.

Clark, John Bates, The Distribution of Wealth: A Theory of Wages, Interest and Profits, New York: Macmillan, 1899

Davenport, H. (1911) "Cost and Its Significance," American Economic Review 1 (December): 724-752

Davenport, Herbert J., Economics of Enterprise, New York: Macmillan, 1914.

Foldvary, Fred, Public Goods and Private Communities: the Market Provision of Social Services, Brookfield, Vermont: Edward Elgar, 1994.

Gunning, J. Patrick, "Entrepreneurists and Firmists: Knight vs. the Modern Theory of the Firm,"Journal of the History of Economic Thought, March, 1993.

Gunning, The Failure of the New Subjectivist Revolution: A Constructive Evaluation and Partial Reconstruction of the Economic Theory of Ludwig von Mises, P.O. Box 60002, Ft. Myers, Florida: Nomad Press, 1994.

Gunning, J. Patrick, "The Theory of Entrepreneurship in Austrian Economics," in Keizer, W., Tieben B. and R. Van Zijp (eds.), Austrians in Debate, London: Routledge, forthcoming 1997a.

Gunning, J. Patrick, "Ludwig von Mises's Transformation of the Austrian Theory of Value and Cost," History of Economics Review, Summer, 1997b.

Gunning, J. Patrick, "The Idea of the Entrepreneur Role as Distinctly Human Action: A History of Progress," 1997c (manuscript)

Hawley, F.B., Enterprise and the Productive Process, New York: Putnam, 1907.

Knight, F. H., Risk, Uncertainty, and Profit, New York: Houghton Mifflin, 1921.

Menger, Carl, Principles of Economics, Translated by James Dingwall and Bert Hoselitz, New York: New York University Press, 1981. (Originally published in German in 1871).

Von Mises, Ludwig, Human Action: A Treatise on Economics, Yale University Press, 1949 (third revised edition published by Henry Regnery Company, 1966).

Copyright © 1997 by James Patrick Gunning

Send comments to:

Gunning’s Address
J. Patrick Gunning

Professor of Economics/ College of Business
Feng Chia University
100 Wenhwa Rd, Taichung

Taiwan, R.O.C.
Please send feedback  


Go to Pat Gunning's Pages

Mirror Site of Pat Gunning's Pages

Popular Pages