Property, Power and Public Choice

by A. Allan Schmid. copyright, 1987, 1997.



Freedom for the pike is death for the minnow.

--Isaiah Berlin

Plainly the sheep and the wolf are not agreed upon a definition of liberty.

--Abraham Lincoln

Freedom's just another word for nothin' left to lose.

--Janice Joplin


The focus shifts from the positive analysis of the previous chapters to an exploration of the possibilities of a normative analysis.1 The varieties of interdependence have been explored, and the prediction of substantive performance related to alternative property rules has been noted. What can the analyst now say about the best public choice of alternative property rights? Where people conflict, can the economist offer any prescription of what should be done? These are the questions of welfare economics which Schotter (1981, p. 6) defines as the study "that ranks the system of rules which dictate social behavior." This chapter will contain an argument that the answer to these questions is in the negative, although the economist can provide useful consumer and voter information.2 But, regardless of the answer, its exploration should reveal more about predicting the substantive results of alternative institutions.

Like all scientists, economists are careful not to argue that the authority of science gives them the ability to tell others what they must do. One seldom sees an economist making a disciplinary argument for why A should have a right and B an exposure. Still, it is common to find arguments in the literature that purport to prove that private property and the market are the most efficient of all institutions and produce a maximum of social welfare (Alchian and Demsetz 1972). Some writers make a counter-argument that efficiency demands planning, government regulation, and socialist ownership (Krutilla 1967 and Galbraith 1967). How do economists who eschew authoritative value judgments that favor A over B come to decide on which institutions are better? The route is through the doctrines of Pareto better efficiency and consumer sovereighty, which are implemented by the competitive market (Buchanan and Tullock 1962, p. 92). What follows is an exposition and critique of this doctrine. It will be argued that competition and income distribution as traditionally defined do not speak to all of the varieties of interdependence discussed above. Because of this, many theoretical propositions about institutional choice are found wanting, leaving citizens to reach their own political compromises.

Pareto-Better Efficiency and Consumer Sovereignty

Initial income distribution is exogenously determined in most economic models. The determination of factor ownership is somehow made prior to economic analysis. There is some assumed political process that allocates ownership. The economist proceeds to ask how each property owner can maximize welfare without infringing on the property of others. The conventional answer uses the doctrine of Pareto-better efficiency as a guide to institutional choice. That institution is better that makes it possible, when starting from some given factor (or goods) ownership, for some person to be made better off without anyone else being made worse off. This maxim apparently enables the economist to keep out of the argument between A and B and still have some expert advice on the choice of institutions. If no one is going to be worse off, all ought to embrace the economist's suggestions.

The essence of the Paretian approach is the separation of income distribution and resource (factor) allocation. The distribution of ownership of factors is acknowledged to be a political question, but, once it is settled, the economist asserts that his advice cannot be rejected by rational men who hold that people with income should be able to use it as they wish. In other words, the consumer should be sovereign, and institutions should protect and further this sovereignty.

The institution that apparently fills the requirements is the market, which alone can ensure that only Pareto-better exchanges are made.3 Starting with their original factor entitlements, each person voluntarily enters into trades to obtain factors and goods that have more utility to the individual than her original set. These are referred to as Pareto-better trades. Any opportunity of this sort that has not yet been consummated is referred to in the literature as a "Pareto-relevant externality" (Buchanan and Stubblebine 1962). If people have knowledge of them, and they are not overwhelmed with transaction cost, the market will eliminate (internalize) all Pareto-relevant externalities, when it is in equilibrium. People will then have made all of the mutually advantageous trades that are possible. Any further movement of factors and goods would be objected to by one of the parties and is termed "Pareto-worse." A Pareto-worse goods movement would either involve theft or a new political distribution of property rights (income).4 After voluntary market trade stops, only Pareto-irrelevant externalities are left.5 These terms describe something about market trade and as such are positive. However, the terms also carry value-laden implications.

In order to understand the implications of the Paretian approach, the discussion must become more technical. The role of income specification in terms of goods or factors and the role of competition will be made explicit--but not in a detailed manner, since there are already numerous texts on the subject (Burkhead and Miner 1971 and Ochs 1974, for example). The essence of the model can be seen by positing a two-person and two-good world in the context of an Edgeworth-Bowley box diagram shown in Figure 2. (This is a static model where the total amount of goods is given. It does not indicate effect of distribution on production.) The horizontal axis shows the total amount of good X available and the vertical axis shows the amount of good Y. The amount held by A is measured from zero in the lower left-hand corner and the amount held by B from zero in the upper right. The original ownership of factors (or in this case of two goods) is assumed to be point W on the diagram. Thus, A has 0 X2 of good X and 0 Y1 of good Y, with the balance owned by B. There is some utility level of each person associated with this original position; this level is shown by the highest utility indifference curve of A and B that intersects this original goods endowment. What is the Pareto-optimal point, given this beginning? Can it be achieved through market bargaining?

As we move left along A's highest indifference curve I A1, we find A as well off as before, but each such movement puts B on a higher indifference curve until reaching point C2. Any point along this range makes B better off without making A worse off. A similar situation exists along B's highest indifference curve intersecting point W. Point C1 makes A much better off with no worsening for B. The shaded area in the figure that is in the shape of a lens contains many points that are Pareto-superior to the starting place W and better than any point outside this area. Prices can be made explicit in the model by drawing a line from point W to point C1 and also point C2--thus illustrating the ratio in which good Y is traded for good X at those points. Given the original goods ownership (W) where A has much X and little Y and given the shape of the indifference curves, he would prefer price line P1 to line P2 since Y is thereby cheaper and A can increase his welfare the most. B would prefer P2. At price line P1, B is no better off than at W, and there is no incentive for trade. B could buy X1X2 and sell Y2Y1 and stay on the same indifference curve I B1. A would thereby be overjoyed to sell X1X2 and buy Y2Y1 and move to a higher level of utility. A would prefer an even lower price for Y, but, beyond that point, B would refuse to trade and would just use the goods already owned. That is to say, any point outside of the lens is Pareto inferior to any point inside.

The trading locus that A would prefer is Pareto-optimal once it is reached. Once point C1 is reached, no departure can meet the Pareto-criterion. This is the point and implied price that would be reached if A is the strongest bargainer. It represents the best price ratio that A can achieve if he can dictate prices.

Person B would prefer price line P2, in which case A stays on the same indifference curve that he started with (IA1) and B improves his welfare. A can sell X3X2 and buy Y1Y3 and stay as well off as before. B will be pleased to buy that X3X2 from A and sell Y1Y3 and move to a higher indifference curve. B would like to reach point C2, which is Pareto-superior to point W. Point C2 once reached is Pareto-optimal, and no departures can meet the Pareto-test. We now have two Pareto-optimal points, and in fact there are an infinite number along the line between C1 and C2, which is called the contract curve (or conflict curve). This line defines the points of conflict between the parties, given the original goods ownership. Any point off the line is inferior, and Pareto-better trades can get to the line (ignoring transaction and information costs).

Where the parties will finish up along the line depends on their relative bargaining strength, and this depends on their total opportunity set. If one of the parties has much property and many other alternatives, while the other has little, the stronger party will make the best deal and force the price to the end of the conflict curve that is most advantageous. It is not enough then to know only the original goods ownership (or factor endowment). If competition is lacking, the relative wealth of the two parties may mean that one of the parties can afford to outwait the other and do without the particular product. The seller with a few alternatives may have to sell now or starve. This situation cannot be understood from knowledge of factor ownership alone. We need to know about all of the rights that each party has that give them bargaining power and also something of the personality of the people (Raiffa 1982). One may be a gambler, and the other a risk averter. The parties may employ strategies to such an extent that they enrage the other person and prevent even Pareto-better trades from occurring. This is the formal model behind the earlier discussion of game-theoretic indeterminancy in Chapter 5 involving small numbers of bargainers.

This model clarifies a point made earlier. There are many Pareto-better solutions rather than a unique one even within the lens, and there are many outside it if rights change. While we cannot say which point along the conflict curve will be reached, we do know that if transaction costs do not interfere, and if people do not get mad in the process, some point on the conflict curve will be reached. (This is not necessarily true under game-theoretic conditions.) This is to say that one of the many Pareto-optimal points will be reached. Pareto-optimality is just as consistent with monopoly rights as with pure competition. Bargaining power and skill influence outcomes. Most texts show that monopoly reduces welfare by restricting production. However, this theory is seldom explicated along with Pareto-better strictures. The abolition of existing monopoly advocated by most economists is not Pareto-better. This situation points to a selective application of the Pareto-criterion to policy.

The model also makes clear the role of the original rights distribution. If we started at point W1 on Figure 2, we would reach a different set of points on the contract curve. They too would be Pareto-optimal. Pareto-optimality is only relative to a given starting place and does not instruct what that initial distribution must be. Alan Randall (1972, p. 25), states, "Efficiency is an inadequate criterion because what is efficient changes as property rights, the distribution of wealth, and the distribution of income change. Just to demand efficiency leaves open the question of which of the infinite number of theoretically possible efficient solutions is preferable."

We have been examining transactions between two persons (or by extension among a few people). The situation changes significantly when there are large numbers of buyers and sellers. While the market has been referred to as a bargained exchange system, with pure competition, there is literally no bargaining. No one person's actions can affect price. Everyone is a price taker. There are no explicit negotiations between buyers and sellers. As far as each person visualizes it, price is given, and the only choice open is to adjust quantity taken. Each person adjusts purchases so that his marginal rate of substitution between goods is equal to the ratio of their prices. As all buyers do this, price may change, but it is not the result of explicit bargaining among the parties. Each person makes her buying and selling decisions independently. The conflict curve is not visible to anyone.

The relationships can be portrayed if we return to Figure 2; only now our two parties are not alone but are just two of many. Assume that, in some previous time period, the price ratio of the two goods was given by the line P3. With the new indifference curves in the current time period as shown in the diagram, A is in a position to improve her welfare and move from her initial goods ownership at W through trade. She would like to move to point A, which is the intersection of the price line with her highest possible indifference curve. A wants to sell quantity X5X2 of good X and buy Y1Y5 of good Y. However, A's offers will not be fully accepted. There will be an excess of supply for X. B and all of the other like traders would prefer to move to point B, which is the intersection of the same price line with his highest possible indifference curve. B only wants to buy X2X6 and sell Y1Y3. This is a disequilibrium. In the face of excess supply, the price of X will fall. It will fall because of the quantities independently offered and purchased and without any explicit negotiating. The price will eventually fall to price line P . The equilibrium is at point C. At this point, when the price line intersects the highest indifference curve of A, it is at the same quantity where the price line intersects the highest possible curve of B. The two parties have adjusted their own purchases and sales so that they reach the highest possible welfare, given their preferences, original goods ownership, and prices. Point C happens to be on the conflict curve, which is Pareto-optimal.

Individual adjustments of quantities offered and taken in the competitive market indicate the preferences of the parties as constrained by their original goods ownership. The relative wealth and related bargaining power of the parties do not enter into the calculations. The rich man can put no pressure on a poor seller since the seller has many other possible buyers to turn to. Whether rich or poor, each person can only adjust his own actions to the given price. This situation is in contrast to the bilateral-monopoly or small-numbers case discussed previously where the trades are influenced by the process of bargaining and the distribution of rights affecting bargaining power. This is the logic by which the competitive market is seen to implement the preferences of consumers made sovereign by the original distribution of income. It appears that performance is determined by competition and the original distribution of factor ownership.

Pure competition is like having an additional property right that prevents unequal wealth from being turned into bargaining power. It is like a law (for example, an antitrust law) that makes part of some people's total and large opportunity sets inoperative.

Space does not permit a complete general equilibrium model that incorporates such elements as production costs. For our purposes here, it is sufficient to note that total demand for the incompatible-use goods that we have been analyzing is obtained by summing horizontally the demand of individuals. This is shown in Figure 3. Person A is poor and buys less of the good at the given market price than does the rich B. Person A buys Q1 and B buys Q2, which sums to Q3. In equilibrium, price equals marginal cost and the sum of the individuals' marginal values just equals the marginal cost of production. Supply equals demand, and no one can get more without someone getting less. The key point is that for incompatible-use goods in pure competition, each consumer is made sovereign by the ability to adjust quantity taken to the given price, with the only constraint being the original ownership of income. No one person can affect price, and thus power is limited by competitive rules.

The usual question raised about the Paretian model of pure competition is whether such competition actually exists in the real world. Economists are divided on the subject. Some say that the only approximation that exists is in some agricultural products and individual providers of some services. Others say that, while competition is not pure, it is close enough for effective competition. Settling this argument is beyond the scope of this book. But regardless of the degree of competition (number of buyers and sellers), there are many other ways for people to affect each other. Power is not just a matter of competitive rules.

In developing the following argument, several points will be examined. First, the character of the Paretian criterion itself will be criticized, and its implications for the debate over regulation or reliance on the market will be drawn. Then the traditional specification of income distribution in terms of goods or factors will be discussed.

The Paretian model labels certain changes in goods holdings as Pareto-better or relevant and others as Pareto-worse or irrelevant. It is easy to slip from Pareto-irrelevant to policy irrelevancy. Income distribution may fall off the research agenda. The terms may suggest that not only is distribution exogenously determined but also it is not a very important question. Note the following from Dean Worcester (1972, p. 58), who states that "making the better choice between two sets of property rights is a minor matter as compared to the need to establish some set of rights." Peter Steiner (1970, p. 40) suggests that we agree "some way, anyway" on social values and then use them to guide choice of institutions. It is one thing to be neutral between people as they struggle to be declared the owner of a resource and quite another to argue that it is a minor matter. The urgency to get on solid ground where one's nominally scientific craft can be practiced is often blinding.

What about the Pareto-criterion itself? Is it ethically neutral? (The position taken here is similar to that of Samuels 1972, pp. 68-93.) At first it sounds like a criterion that should have wide acceptance. It allows for improvements but protects from harm. Still, it implicitly gives stature to the status quo distribution of rights. To accept that only Pareto-better trades are legitimate is to accept the original distribution of rights as legitimate. The Pareto-criterion neither suspends judgment nor frees the analyst from choosing between A and B but rather accepts the choice between them already implicit in the existing distribution of rights and income. To accept the distribution of rights as a prior given and to employ the Pareto-criterion is to do the same thing. All statements of Pareto-optimality should be read as the conditions where welfare is maximized, given the original property endowments. One can be charitable to all writers who make Pareto-optimum-policy conclusions by adding the phrase "given the original property rights," even where it is inadvertently omitted. Of course, when the phrase is made explicit, the policy advice on alternative institutions loses its punch.

Regulation Versus the Market

Perhaps the subtle inferences from the hidden endorsements of existing rights distributions contained in Pareto-better conclusions can be illustrated with an example. With growing industrialization and use of chemicals, the ability of a manufacturer to affect downstream water quality was increased. Those harmed began to scream. Pigou (1946, pt. 2) looked at the situation and said pollution created a difference between marginal private cost to the manufacturer and marginal social cost. Pigou's conclusion led some to suggest government regulation, which would prohibit certain discharges or require that a certain stream quality standard not be violated or levy a tax on all waste discharge at rates that would maintain a certain stream quality (Kneese and Schultze 1975).

These solutions involve a direct role for government--that is, a movement of goods that not all parties agree to. It is not Pareto-better. Can the analyst then criticize these government actions? R. H. Coase (1960) argues that regulation is not usually desirable. J. A. Seagraves (1973, p. 620) argues that "governmentally levied stream charges would not necessarily encourage a market test regarding the optimum quality of the river .... and would imply a redistribution of property rights which is larger than necessary." The writer goes on to recommend a proposal that "would allow a downstream city to bargain with an upstream city so that the former might pay the latter to adopt higher standards than those dictated by their discharge rights." Nowhere does the writer say that the private market institution is superior only given original rights distribution or even why he presumed the upstream city has the right to use the waste disposal capacity of the stream. Perhaps this author agrees with the one who said that the distribution of rights is a minor matter, in which case it is of no importance whether the resource is owned by the upstream or downstream city, or by the man in the moon.

The partial Pareto-logic can lead to an issue classification with heavy value implications. For example, in the case of pollution, it leads to the policy recommendation that a market be established for water quality rights that have been vested in one party or the other. Some of the pollution then may be a Pareto-relevant externality that will be subject to a voluntary market bid by the downstream city in the case cited. If the bid is worth more than continued pollution by the upstream city, it will sell, and pollution will cease or be reduced, and the result is Pareto-optimal. If the upstream city rejects the bid, the externality continues, and while internalized, it is termed Pareto-irrelevant often with a strong implication that it is policy irrelevant as well. In that case, any government regulation to reduce pollution would have prevented Pareto-optimality and the most valuable use of the resource.

It is a heavy burden for a policy to bear the appellation "Non-Pareto-Optimal." The writer who would never want to be in the explicit position of arguing why A should own a resource and B should not, nevertheless will label government regulation that protects A as "non-Pareto-better" and slip into an assumption that B owns the resource and the market solution is optimal. In effect, a writer who would never claim the capacity to define scientifically a social welfare function that weighted conflicting interests nevertheless endorses the interests of the one over the other. Paul Samuelson (1969, p. 105) has warned against the analyst who "tacitly lapses into the cardinal sin of the narrow 'new welfare' economists: 'If you can't get (or even define!) the maximum of a social welfare function, settle for Pareto-optimality,' as if that were second-best or even 99th best."

Regulation is conceptually indistinguishable from any factor ownership. Does government regulate private water courses or does it grant the use of public streams to private parties under certain limits? If it is accepted that there is some role for public ownership, this ownership does not have to be complete. The government may only be interested in some aspects of a resource for public use and may wish to grant or charge for private use of other aspects. Neither a partial-use sale, nor tax, nor grant can be challenged on Pareto-efficiency grounds.

The Pareto-criterion can be used against the Pigouvian argument that the existence of externalities was a prima-facie case for direct government action via regulation or punitive taxation. But we must be careful of overkill. In putting Pigou down, some have tried to put down all regulatory action. Some new regulation was never intended to improve the efficiency of existing rights but only to change these rights. Since Pareto-optimality just carries out given rights, it cannot be used to select these rights. Pareto-efficiency is derived from given rights, not instructive of them. The Pareto-criterion carries out the implications of an original distribution of rights. There are as many Pareto-optimal solutions as there are ways to define and distribute rights, as shown above. One must not avoid the necessity of making a moral choice between A and B by adopting the Pareto-criterion, which only confirms the one already made by implying it should continue.

It is a tautology to say that Pareto-optimal solutions depend on carrying out the original distribution of income and that this distribution should not be frustrated by regulation. If you agree with the original income rights, regulation that would change these rights is illogical. But people agree neither on the original rights distribution nor on whether it should be continued. If economists are silent as to the distribution between A and B, they must also be silent as to the regulation that implements A's rights as opposed to private ownership by B. If B owns, then regulation violates that ownership. But if A owns, regulation can be the vehicle of that ownership. The efficiency of regulation depends on who has the right and the performance objective one has in mind.

The above line of reasoning puts the ethical choice (value judgment) between conflicting parties back on the center stage of public choice. The Pareto-criterion does not avoid the necessity of this choice.


It seems intuitively obvious that the distribution of ownership of incompatible-use goods affects income distribution, but what can be said of resource allocation? There is a prominent line of reasoning that begins by asking the resource allocation consequences of whether a resource is owned by A or B. The question is put in terms of a rule of liability for the effects of use by A on the welfare of B and can be illustrated by reference to the pollution example above. (This example was first developed by Kneese 1964, pp. 43-46.) Assume first that A owns the stream and can use it at will without any liability for damage to the fishing pleasure of B. A is a manufacturer and enjoys a profit of $10 because his riverside location saves him the cost of an alternative waste disposal system. Pigou was wrong in implying that A would completely ignore his effect on B. Anytime B wants to make A an offer to stop polluting, A will listen (unless he is malevolent). Suppose B can only make an offer of $5. The offer will be rejected because it is not as valuable as its use for waste disposal, which will be continued. Now, suppose the property right is shifted to B, and A is now liable for any damage created. The direction of the bidding will now be reversed. B will be better off by accepting any offer over her reservation price of $5, and A will be better off to offer any price up to $10. If they do not get mad at each other in the process, the resource will be sold to A and the pollution continued, as when A owned the resource originally.

This result has been worked out by Ronald Coase (1960) and deserves to be called the "Coase rule." The rule states that the distribution of ownership does not make any difference for resource allocation if transaction costs are zero. (Transactions costs were explored in Chapter 6 of this book.) Suppose that A is a large group instead of a single individual. Its total valuation is $10. But suppose the cost is $6 to get the group organized to make a bid if it does not already own. In that case, if it already owns, its reservation price is $10 and it would not sell to B for $5, but if it must buy, its bid to B is only $4 ($10-$6 transaction cost) and the bid will be rejected. The use would differ depending on who originally owns. The Coase rule can then be stated as follows: "The ultimate result (which maximizes the value of production) is independent of the legal position if the pricing system is assumed to work without cost." Coase emphasizes that the market may be too costly to organize where a large number of people demand the same good. This problem will be explored in the next section. Some writers nominally following Coase have noted that income distribution is different with different property rules, but ask us to keep our eye on the resource allocation. After all, is it not use that matters, and is not income distribution a minor side effect? The normative implications of this have been critiqued by Samuels (1974a). A bastardized version of the Coase rule is often shortened to conclude that property rights do not matter.6 Coase (1994) himself in his Nobel Prize acceptance speech emphasized that since there are many cases of transaction cost, property rights do matter. (Also see Medema, 1995.)

The neoclassical literature has come full circle. First, it argues against regulation because regulation gets in the way of Pareto-better trades. This argument puts the status-quo income distribution on a pedestal. Then it implies that income distribution (factor ownership) is not very consequential since, in the process of making Pareto-better trades, resource allocation is unaffected. This implication knocks the status-quo income distribution off its pedestal. You cannot have it both ways.

The Coase rule is itself a positive proposition . But it must be kept in mind that, regardless of an author's intent, models are used for policy implications. Many readers of this literature seem to remember two key messages: (1) regulation is bad and (2) there is no case for changing status-quo rights since resource allocation is unaffected anyway. With cold logic, it seems that one might conclude that since rights do not matter, we might as well change them and equalize income distribution. But that would be a radical conclusion and the economic profession mainstream is not radical.

No one makes an explicit argument that rights do not matter for income distribution. But, by emphasizing the consequences for resource allocation, attention is drawn away from income distribution. The Coase logic calls for another more important research item. This is to look for ways to reduce transaction costs since these are the costs that keep the theory from working perfectly. Transaction costs become the enemy of the people. The most obvious place where they are infinite is where goods are held in common and trade is prohibited. The Coasian analysis suggests that common property must go.7 A new rule for institutional choice then becomes that all property should be private and individual except where transaction costs rule it out. Common property rules exclude no one even though high exclusion costs do not exist inherently in the good. When people legally cannot be kept off, resource destruction may result. Also, where property is held in common, it may be used by a property holder when someone else would pay more to use it. This violates Pareto-optimality, but only if third-party interests are not possessing of rights. The fact that A cannot sell is sometimes the device by which C enjoys some use. Some people's use rights cannot be ignored when making Pareto-calculations of exchange rights.

Whether the reader thinks that income distribution is an important topic of research or whether we should accept existing rights that are direct and explicit and make them more valuable to existing owners by reducing transaction costs is up to his or her own value judgment. Before moving on, however, let us examine the Coase rule on its own ground of effect on resource allocation. Several complications need discussion.

(1) The first involves the impact of rights on the marginal utility of money. This creates a difference between bid price and reservation price (compensating and equivalent variation, Mishan 1961, 1967 and Krutilla and Fisher 1975). This difference can be related to the Coase rule discussion above. In the Bowley box of Figure 2, there are only two goods, and it is obvious that if you lose possession of one, both real income and the demand for the other good are affected, depending on the income elasticity of demand. If Coase had used the Bowley box, he would never have assumed that he could switch ownership from A to B and not affect resource use. If a change in original ownership of good X (e.g. from W to W1) is not offset by new rights in Y, a change in X will affect the range of new Pareto-better points that can be reached on the contract curve.)

The examples used by the Coasians implicitly or explicitly assume that the alternative rights distributions under consideration do not affect the marginal utility of money and thus do not feed back on the relevant demand curves. An assumption is made that whether A has the right or not, his real income is not affected, and therefore he buys the same mix of products with or without the right in question. The assumption is that income does not affect the demand curve; this restricts the theory to a set of minor and uninteresting cases. Many major debates over ownership are of the type that affect real income and would upset current consumption patterns. The point can be illustrated with the pollution case again. Suppose that the fisherman affected by pollution is poor. If he does not have the right to be free of the pollution, his bid price may be zero. He cannot offer the polluter anything to stop. But, if the fisherman is the owner of the stream, his real income is significantly different. He would never sell the stream for zero. His bid and reservation prices differ because the right affects his real income and thus his demand curve for the stream. Similar situations exist in many of the current major policy debates such as ownership of off-shore oil rights, copyright in photocopied material, and so on.

(2) There is another possible effect of changing ownership between A and B. People do not have the same preferences; they use their incomes for different products. This affects demand for other goods and resource use and thus incomes of third parties (producer and consumers). For example, if a product has economies of scale, consumer A is affected if as a result of income redistribution between B and C the demand for the good declines and its per-unit production cost rises as less is produced (see Chapter 4). In public investment analysis, it is common to note the indirect or secondary effects of an expansion in output on the income of input suppliers and subsequent processors of the output.

(3) Another characteristic of Coasian examples is that they presume pure competition in the rest of the economy, but less than pure competition in the firms under consideration (or in existence of rents), Recall the example of the polluting manufacturer earning a profit of $10. This noncompetitive return (or rent) allowed him to buy out the fisherman. What happens if the firm is in competitive equilibrium and earning no profit or rent while utilizing the stream for waste disposal? If the industrial firm owns the resource and the fisherman makes a $5 bid, the theory would suggest that the firm would accept the bid and go out of business, moving his resources somewhere else if earning the same or better competitive return as before (no rent or profit). But, the mobile resource alone may not earn as much elsewhere as it did originally in combination with the second resource whose ownership is in question. In that case, A keeps if it is owned, but could not afford to buy. If the industrial firm does not own and the fisherman insists on using his property, the firm could not afford to buy out the fisherman since the firm has no excess returns to pay for it. The firm moves. The theory produces its conclusion again that the right does not affect resource use. But it does so under an assumption that assets are perfectly mobile (as discussed in Chapter 6).

In the real world it seems highly unlikely that a competitive firm can pick up all of its assets and move to another location or line of production and earn the same returns as before. Industries understand this and oppose being denied ownership of streams because they will not be able to buy out other interests and still stay in business and thus will lose the value of their immobile plants. If the firm owns, it will sell only if bid exceeds value of immobile assets. If fishermen own, the firm cannot afford to buy. Resource allocation is affected by the distribution of rights where only competitive returns are being earned and some assets are immobile in contrast to the noncompetitive, perfectly mobile case (Posner 1972b, pp. 37-38). If the loss of waste disposal rights were applied to all firms in an industry, it would shift the industry supply function to the left. Price would rise, production would drop, and some higher-cost firm would likely leave the industry (with accompanying loss of immobile assets). The effect on firm output depends on whether the increased waste disposal costs change the output point where marginal cost equals marginal revenue (Davis and Whinston 1962).

(4) The Coasian examples also make an assumption as to legal remedies available to owners and the ease of estimating damages. In Chapter 6, it was seen that the character of remedies for interference with an owner's use is an important detail of factor ownership. Assume B owns an apartment building with one side facing south. A is about to build a taller building that will cut off the sun to B's apartments. What is the effect on resource use of who owns the "sun rights?" Begin with the rights owned by B, who has a reservation price of $15,000. A's bid is only $10,000 and would be rejected by B. However, it is easy for A to interfere with B's use if the building is built anyway. B's effective right depends on the available remedies if A builds and blocks the sun. If B can get an injunction, then A will have no choice but to refrain from building. But, if no injunction is issued, A can force sale of the sun rights. The consequence depends on the court-determined damage relative to B's reservation price. If B places some unique value on the sun that is not reflected in the market price of similar apartments, the court-determined damage is less than B's reservation price with a loss in the value of the resource. The use of the sun goes to A whenever A's bid price of $10,000 is equal to or less than what A expects to pay for damages. (If damage award is more than $10, 000, A has made a mistake but still uses the resource.)

Now, assume A owns. B will make a bid of up to $15,000, which exceeds A's reservation price of $10, 000. Use of the sun rights shifts to B. In this situation, the location of ownership does affect who uses the resource for what. This is another exception to the Coase rule, which asserts the neutrality of resource ownership with respect to resource allocation. Also see Schap (1986).

In producer-producer conflicts, the court-set damages reflecting market values are likely to approximate the owner's reservation price. But in conflicts with or between consumers over final consumption goods, the objective damage award may depart from the subjective reservation price that would have prevailed if an injunction were available to the owner or if interdependence with the owner's use were further protected by criminal sanctions.8 If the parties have differential ability to interfere with each other's use of a resource, any court award at less than the owner's reservation price (no injunction) means that a superior reservation price may be frustrated. These nitty-gritty details of what it means to be a factor owner cannot be ignored.

(5) There is one further important contradiction to the Coase rule implication that property rights alternatives do not affect resource allocation. Firms and individuals who want to increase their income have two broad alternatives: (1) accept the rules of the game and utilize available opportunities such as investment or (2) utilize available resources to change the rules. In the latter case, wealth in one period of time can be used politically to change property rights and then enhance wealth in the next period. Resources can be given to political candidates and bureaucrats and to influence public opinion.9 When income is used to change rights and income in the next period, it will affect resource allocation. To those who have not read the Coasian literature, this will seem obvious, yet this literature persists.

(6) The Coase rule applies only under certainty. While if transaction costs are zero, there might be a market for risk sharing, it is further necessary that risk attitudes be independent of wealth (Greenwood and Ingene, 1978). If one party is more risk averse than the other, the original placement of the right can affect resource use.

On its own ground of impact on resource allocation, the Coase rule is severely limited in its application. For most of the interesting cases of public debate over rights, transaction costs are not zero, rights distribution would affect the marginal utility of money and risk aversion and thus cause a divergence between bid and reservation prices, exchange affects use rights of third parties, and many assets are immobile. Even if you only care about resource allocation, it is hardly true that the distribution of property rights does not matter.10 And, if you do care about income distribution, property rights distribution matters even more. The lack of data on the size of the above factors prevents resolution of debate. Those who like the policy implications of the Coase rule that existing property rights should be left alone deny that the above factors are empirically significant (Coase 1974). Others believe they see enough examples of their significance to question this policy thrust (Goldberg 1974). The theory is relatively well developed, and it is time for more empirical work (Hoffman and Spitzer 1981).


Joint-impact goods create an interdependence that raises some additional issues for property rights. When more than one person utilizes the joint good, the issue is who pays and who gets to choose the physical level of the good. This question arises even if the goods have low exclusion costs, but the problems are compounded when exclusion costs are high. Does the Paretian logic give any rules for appropriate property rights with these joint-impact goods? Is it adequate to describe income distribution in terms of money or factor ownership?

Suppose we are analyzing the demand for a good like broadcast television. (This case is parallel to what is often referred to as positive externality, such as the classic illustration of bees that produce honey and also pollinate fruit trees.) While many aspects of television present policy issues, quantity of the good is interpreted here as the number of available channels and the marginal cost of another channel is positive. The good is a high exclusion cost and joint-impact good with the characteristics that whatever number of channels are available to one viewer in the broadcast area is potentially available to all at no extra cost. Use of the good is avoidable, so the person who places negative utility on the good merely abstains and there is no interdependence with those who use it. There is pre-emptive conflict, however, if it is not possible for different individuals each to adjust their quantity purchased to a given price. If one person has three channels to choose from, then all have three (regardless of how many they actually use or the hours of use). For a joint-impact good the demand of individuals is summed vertically, as shown in Figure 4 (in contrast to horizontal summation for incompatible-use goods shown previously in Figure 3). First, assume that in spite of high exclusion costs we have some magic meter to obtain individual demand curves and that these are as shown in the figure, which depicts what people will pay for a given quantity of channels.11 At a quantity of three channels, the sum of individual marginal values just equals the marginal cost or producing the third channel. However, this Pareto-optimal equilibrium (Johnson 1971, Appendix C) is obtained by subtly introducing a new property right into the system.

A close examination of the demand curves for the individuals shows that, at the Pareto-equilibrium, the parties are paying different prices. One person pays $5, and another $6. Since the quantity (number of channels) available cannot be adjusted to accommodate different preferences, the price has been adjusted as suggested by Lindahl (1958). This model contains an inexplicit value judgment. We might justify the fact that the rich take more of an incompatible-use good than do the poor in terms of their ownership of income (if we accept the legitimacy of the distribution of that income). But it takes an additional value judgment and right to insist that the rich (or persons with strong preferences for television) must pay more for the same good than another person. In settling conflicting interests, it is not sufficient to accept only the original distribution of money income even where demand is known.

More than one pricing scheme is possible. The marginal cost of another channel is $11. If the price were the same to all customers, a price of $5.50 would be sufficient to cover costs. This is not an equilibrium price, but whether it is a Pareto-optimal price depends on whose interests are backed up by a property right. At a price of $5.50, person A would prefer a lesser amount, but such a state cannot be achieved for two reasons. Person A cannot have a different quantity than that available to all consumers because it is a pre-emptive good. Further, quantity of channels is not infinitely divisible. The consequence of moving from two to three channels is quite different than moving from two to three apples. Person A does not find a three channel system worth $5. 50 but will probably take it rather than stop watching.12 Person B on the other hand would prefer to have a larger quantity at the price of $5.50. But it is not possible for the two parties to have different quantities available. At least, person B finds the three channels worth $5. 50, even if more are preferred. Person B has a consumer surplus and does not have to pay as much for the three as she would be willing to pay. If we begin with a price of $5. 50, person B will never agree to pay $6 for the same quantity. Such a move would not be Pareto-better for B, even if A were pleased to see his price drop from $5.50 to $5.

If there is no pre-existing take-it-or-leave-it price, the parties might agree to the differential prices. However, even if they only have to pay what the marginal unit is worth to them, it is conceivable that some people may be offended to pay more for the same good than their neighbor (Tresch 1981, p. 118). Some may not want their consumer surplus captured by producers (see Chapter 7). A detached, objective observer might regard such an attitude as irrational, but it is hard to argue that it does not exist. If it does exist, some additional rule is needed that says whether this taste is to dominate over other conflicting tastes.

People with different values on the intramarginal units of incompatible goods do not like to be exposed to a discriminating monopolist and to be forced to pay more than their neighbors, just because they have different elasticities of demand. Can it then be assumed that people are not concerned when other people with different marginal valuations for the commonly available joint-impact good are charged different prices? Samuelson (1969, p. 122) derisively calls this "interpersonal 'Robin-Hood' pricings." It is the same problem as differential pricing of goods with economies of scale discussed in Chapter 4.

Economists like to postulate a set of exogenously generated preferences independent of the economic decision system. The system should make the prior tastes effective, not modify them (consumer sovereignty). However, the pricing policy discussed here needs to be related to the Chapter on psychology above. Preferences are learned and the decision system can be an ingredient in this learning. Suppose I am a member of the group that values three channels at $6 while others value them at $5. I note that I pay more than my neighbor as a result. If the preferences before the multiple-level tax were unequal, are they likely to remain so after the tax? In the context of incompatible-use goods, I am aware that people who are poorer buy less of certain goods than I. But that does not cause me to question my own values since my income is greater. But, if I become aware of different preferences for the same quantity of a good, I must begin to question my own values.13 Why is this good so valuable to me? When the magic value-measuring meter is applied next time, I may have changed.

If the parties can bargain, they will negotiate over the amount of the good and the price (cost or tax share). This can be shown in Figure 5. Here, individual demand curves are added vertically as in previous Figure 4 but are shown on a single graph. Person A, acting alone, would buy a one channel system at its marginal cost of Pt and B would aim to buy two at that price. If A discovers their interdependence first and sees the effects of B's purchases, A will cut back and buy none hoping to enjoy a free ride. In the absence of bargaining, this move would be stable if B does not get mad. If bargaining can occur, the outcome becomes indeterminate in small-group situations as strategic bargaining is employed (including chicken games). (See Chapter 5.) There is no reason to think that the price rule of each person paying his marginal valuation will be the one chosen. Even if exclusion were possible (avoidance), it would not be demand revealing with pre-emptive goods. A high demand person may pretend (bluff) a low marginal value and then offer to pay a low price knowing that if she is allowed any use at a price, the physical quantity made available to her can't be varied independently of others. Such a person can't be distinguished from the person of genuine low demand.

Given B's purchase of two channels, A would buy more along the curve GH, and B in turn would buy more along ZK. It is possible for three channels to be purchased if A contributes the difference between the marginal cost schedule and B's demand schedule ZK. For all three channels, B pays P2 and A pays P1 , which meets the pricing rule of everyone paying their marginal valuation, and output is optimal with the sum of these marginal values equal to marginal cost (production efficiency).

But several pricing schemes are possible. A could drive a hard bargain and insist that B pay Pt for two channels and P2 for the third, while A paid only P1 for the third (but of course can enjoy all channels because it is a nonexclusive joint good). B may not regard this scheme as fair even if both pay their marginal values. With the hard bargain at three channels, A enjoys a huge consumer surplus Da, O, Q2, J, H. Person B, on the other hand, has a smaller consumer surplus of Db, Z, Pt plus triangle MZK. If B suspects this, he is going to try to get a better bargain or be mad if he can't. If B drives a hard bargain he may insist that A pay P1 for all three channels, or, if really tough, he may require A to pay P1 for the first and a price along ZX for the second and third (which again is P1 for the marginal unit). In this case B pays nothing for one and pays P2 for the marginal unit. A pays P1 and B pays P2 for the marginal unit, even though they pay various combinations of prices for the intra-marginal units.

In each of these quite different total-cost shares, the price of the marginal unit to each party stays the same, and if we ignore real income feedbacks, the quantity they want stays the same. As Buchanan notes, "Equilibrium may be consistent with almost an infinite number of sharing schemes for the costs over infra-marginal units" (1968, p. 37).

The problem is actually much more complicated, as Buchanan (1968, chap. 3) has explained. You cannot get an individual's demand curve of the usual sort unless the average price does not vary with quantity--that is, average price must equal marginal price. Each demander is asked how much he wants at each average price. In other words, all intra-marginal units cost the same as the marginal unit. But, if the price of the intra-marginal units is up for negotiation, there is no demand curve of the usual sort, and buyers are not sure how much they want at various marginal prices because of the income feedback effect of the price of the intramarginal units. Buchanan (p. 44) suggests a convention that tax price per unit be uniform over various quantities for each person. He notes this is an arbitrary value judgment that distributes the gains of trade between the parties in a particular way. He suggests this needs to be a prior chosen constitutional rule or what here is called a property right. To choose a tax share or market differential pricing rule requires a moral choice in addition to that of money income distribution, reflected in the budget constraint. There is no reason based on Pareto-efficiency why one rule or another should be used for intra- marginal prices. MC = 0 breaks the necessary connection between bearing costs and receiving benefits (prices).

Are we to have a constitutional rule that says the person with the highest marginal value is to pay more? When society made a choice on money income distribution, did it have in mind that the person with the higher demand would have to pay more? This is the issue raised by Paul Samuelson (1955). He developed a general equilibrium model where both cost share and amount of the joint-impact good are established simultaneously, given a social welfare function, which makes the trade-off between competing interests explicit. This elegant solution is elaborated in the literature and need not be detailed here (Musgrave and Musgrave 1974, pp. 60-68 or Burkhead and Miner 1973, pp. 66-74). It highlights the fact that money income has no meaning for welfare until prices (tax share) of joint-impact goods have been set.

Samuelson and Buchanan make it clear that the marginal pricing rule feeds back on the welfare of the parties.14 Thus, we need a social welfare function and public choice. (In the Samuelson model, there are no prices, since the omniscient planner just delivers the net goods.) Money income distribution between parties does not speak to all of the ways they are interdependent. The problem is even more complicated in the case of nonoptional (high avoidance cost) joint-impact goods when one person places positive value on the good and another negative. What does the rule requiring the sum of marginal values to equal marginal cost mean in this case ? Those harmed want to veto production of the good or be compensated.

The discussion of broadcast television, above, has been in terms of the optimum physical units of output (numbers of channels). Since exclusion costs were high, it was assumed that everyone could use whatever good was available. The implications for joint-impact, low exclusion-cost goods should also be noted (which will also serve to reiterate the conclusion of the above discussion). Assume a scrambling device or cable delivery makes exclusion economic. If the marginal cost of another user is zero, what is the optimum number of users? Conventional welfare economics theory using Pareto-better logic suggests that the optimal pricing rule is price equal to marginal cost (P = MC). The logic can be seen by reference to Figure 1 in Chapter 4. (This figure is developed in terms of decreasing cost goods, but the same logic applies to zero marginal-cost goods.) Begin with a price on the average revenue curve (such as Pr) that exceeds MC. Additional users can be added who are willing to pay the additional cost of production, which in the case of pure joint-impact goods is zero. Apparently, it is possible to improve the condition of all additional users who want the product without making any other user worse off. With joint-impact goods, even if exclusion is economic, it is apparently the wrong thing to do. As already noted, exclusion won't reveal demand. And even if individual demands were known, their use is questionable. If the intra-marginal consumers pay what they did before (for example, Pr), the fixed costs will be covered, but the market must be differentiated and some who add nothing to cost will be excluded.

Implementation of the P = MC rule, however, is difficult in a market. If the same price is charged all buyers, then the "correct" price of zero will not keep the private firm in business. A private firm might cover total cost by market receipts if it differentiated its market and charged different prices to different buyers (as discussed in Chapter 4 and above in the context of price bargaining among consumers, Figure 6). See Rowley and Peacock (1975). This differentiation where feasible is often objected to by some buyers who regard it as unfair discrimination. We need a value judgment that decides whose interests are to count if we are to retain the private market firm in this case. The market failure literature suggests that the market be replaced by government provision via some taxation scheme that does not affect use at the margin (perhaps a poll tax). This solution implements the Pareto-better pricing rule of P = MC. However, inquiry must be made as to who pays the tax. The tax may be paid by those who do not want the good or who do not wish it at the "price" implied by the tax. Just as it required a value judgment on a market pricing scheme, it requires a value judgment on a governmental pricing (taxing) scheme. One can speak of a Pareto-better pricing rule of P = MC only if one presumes a property right in how total costs will be financed.

Pricing Goods With Overhead Costs

Goods produced under economies of scale and joint-impact goods are parts of a continuum related to marginal cost decreasing and becoming finally zero. The production of this group of goods was conceptualized by J. M. Clark as being marked by the existence of "overhead costs." He also included certain aspects of horizontal and vertical integration (see Chapter 6 on cost discontinuities and investment coordination), peak load situations (Chapter 7), and immobile assets (Chapter 6). Overhead costs refer to "costs that cannot be traced home and attributed to particular units of business," and thus "an increase or decrease in output does not involve proportionate increase or decrease in cost" (1923, p. 1). Conceptualized in this way, the policy problem for Clark was how to utilize underused productive capacity. If you could increase the output of a given product or a set of interrelated products without a proportionate increase in cost, this was a source of economic progress.

The same phenomenon can be looked at from the point of view of inputs or products. An input that cannot be traced to a particular unit or kind of output is an overhead cost. The products of these overhead (or joint) costs are either joint-impact goods or goods produced under decreasing cost. The pure joint-impact good is the polar case when another unit (user) can be obtained at no additional cost. Clark saw many governmentally produced goods such as technological knowledge as overhead inputs to industry, which could not be produced by a single firm because the inputs were not related to a particular firm's output (but contributed jointly to many firms' outputs). Each one of the users of a joint-impact good has joint costs with all the rest of the users. These joint costs cannot be mechanically traced to any one user. He saw that cost as a natural phenomenon had no meaning when overhead costs existed (p. 32). If the output observed to change with added units of capital or labor is attributed to that marginal unit, then none of the product will be available to cover overhead. The sum of marginal products would exceed total product.

Nevertheless, the overhead costs must be allocated and paid by someone. For further discussion of joint-cost allocation, see Eckstein (1965, ch. 9) and Bain (1966a). Clark argued that this allocation must be made with the objective of reducing idle capacity. And this could only be achieved by charging different prices to different users (that is, to attract a buyer to use the idle capacity required a lesser price than that charged to earlier purchasers). "The problem involves the allocation of certain general burdens where reasonable men may differ as to what is just apportionment. There is no natural system of prices in the old sense. Cost prices do not mean anything definite anymore. Efficiency requires discrimination and discrimination has no universally accepted standard to go by to keep it from degenerating into favoritism" (Clark 1923, p. 32). For price differentiation not to be seen as discrimination requires a widely shared value judgment. Clark argued that pricing under conditions of overhead costs was not something to be deduced by formula but was a cooperative choice, "a partnership between labor and the other parties in industry" in "amending the constitution of industry" (p. 482). Pricing policy during depressions, sharing of industrial knowledge, unemployment insurance, and a trade ethic for defining harmful price discrimination was up to public choice, which he hoped would be guided by enlightened pursuit of reducing idle capacity of the whole economy.

Writing in 1923, Clark believed that the business cycle could be controlled by social cost-keeping replacing the more common individualistic canons of industry and labor. Yet it was the Keynesian policies of a more centralized sort combined with the necessities for common action created by war that reversed the Great Depression and sprung the social trap. The attraction of mutual gain from fuller utillzation of capacity has not yet been sufficient without other community-creating forces. Until we pay more attention to distributive conflicts involving decreasing cost and joint-impact goods, this community will be hard to achieve in peacetime. It is not possible for every consumer to be regarded as the marginal consumer and only to pay marginal cost. It is unacceptable for some labor to be paid less than customary in order to achieve increased employment in a depression. It is not possible for every firm in response to a general decline in demand to lay off employees and expect product demand to be constant. Labor is an overhead cost for the total economy and can't be idled and stored until needed. Lay-offs also destroy morale and team work and thus the marginal value product of labor (Thurow 1985). If we do not agree on the distribution of the gains and costs of full utilization of our resources, full production will not be achieved.

In summary, this discussion is not an argument for a particular pricing scheme. It is an argument that any pricing scheme has distributive consequences and no statement can be made about the distribution of income unless property rights in the pricing scheme are made explicit. To understand wealth distribution, it is not enough to know who have money in their pockets if you do not know how the pricing policy will affect their real purchasing power and who gets to select the good. In short, there is a moral problem even if preferences are revealed and money income distribution is accepted. You cannot tell what a person is sovereign over until the price rule for goods with economies of scale and joint-impact goods is specified.

Policy and Theory

Some models are used to make actual empirical calculations and others are used to reach policy recommendations directly. The latter has been the historical use of the welfare model of pure competition. This model is used to argue, by abstract reasoning without any reference to empirical estimates, the supremacy of the market as a decision-making system. Most of the models employ utility-indifference curves that have no empirical content. If the assumptions hold, you get the best of all possible worlds regardless of whether half the world is unemployed or starving. The same uses are made of these "existence models" that have been examined here.15 They are not meant to be used to make empirical estimates but to argue for certain policies. This fact is made clear by Burkhead and Miner: "Because of the inability to exclude, however, this analysis is essentially irrelevant as an approximation to any viable decentralized voluntary solution. At the same time, it does provide a normative guide for the evaluation of those political arrangements that endeavor to simulate a consumer-preference-directed solution to the provision of public goods" (1971, p. 95). This position ignores the problem of which consumers we are speaking of if interests in joint-impact goods conflict.

The ways in which these existence theorems are used in normative arguments are subtle. An example is provided by professional reaction to the model developed by Paul Samuelson noted above. His model of joint-impact goods provision incorporates the pricing rule and goods ownership into choice of a social welfare function. The model focuses public choice issues on choosing directly between A's and B's total utility. The quantities of incompatible-use and joint-impact goods are derived from this choice. Resource allocation and income distribution are simultaneously determined in one overall moral choice between A and B. Richard Musgrave (1969) has objected to this abstraction. First he notes a practical objection that people cannot agree on a social welfare function. He ignores the equally practical objection to any existence theorem that people cannot agree on money income distribution (factor ownership) either. Second, he argues there is no practical way to specify a social welfare function in terms of utility. Money income can at least be seen. But no government can be very satisfied that real welfare is satisfactory by just agreeing on money income distribution. Further, Musgrave's own models contain equally nonobservable variables.

These practical objections are beside the point. What is really troublesome are the policy implications of different, but equally abstract and impractical models. Musgrave notes that if income distribution policy cannot be settled independently of taxation and public expenditure policy, there is no basis for distinguishing taxation for redistribution of income from taxation as a market price surrogate for purchase of joint-impact goods (p. 133). Economists would not be able to say anything scientific about the global correctness of tax policy. It is my opinion that in fact income distribution and resource allocation are inseparable. Economists are going to be forced to the conclusion that has long been painfully obvious to the person in the street. Welfare (or real income) is not only affected by the resources one owns or money in the pocket but also by pricing policies (including tax policies) and the amounts of government goods provided (including joint-impact goods).

Whatever the motives of economic theorists their theory shapes policy. For example, Tobin (1970) suggests that redistribution via public expenditure practices is not proper. Thus, only direct redistribution, such as the negative income tax, remains. There is something psychologically maddening about taking money out of people's pockets. Senator George McGovern, the unsuccessful Democratic candidate for president in 1972, found this out very dramatically when he suggested heavily redistributive taxes. If theory suggests this is the only way that income should be redistributed, then it constitutes support for the status quo, since such redistribution is unpopular.

Another implication has to do with our attitudes toward changing rights in the flow of government expenditures as opposed to factor ownership. The former can be changed drastically by any legislature, but court consent is required to change the latter. Some people have argued in terms of constitutional "welfare rights" so that the courts would prevent wide fluctuations in welfare transfer payments. People whose well-being is tied to public spending and tax policy are disadvantaged compared to those whose income flow is protected by the courts. The fact that some economic theory suggests that only direct income distribution is efficient contributes to maintenance of this disadvantage.


A political decision is Pareto-optimal only with a rule of unanimity. With less than unanimity, someone will pay for and potentially consume some governmental good that she does not want at the politically chosen price. Unanimity has its problems, too. First, it implies a preference for the status quo, which is a preference for the interests of A over B. Second, while unanimity can prevent some kinds of political externalities, where someone has to pay for something he does not want, it increases decision costs. Whenever the individual wants to obtain a governmental good, he will have to obtain the consent of every other person. He has to balance off political externalities, decision costs, and doing without some good he wants. Thus most people will find it to their individual advantage to support a decision rule of less than unanimity, and the rule will vary by product (Buchanan and Tullock 1962, chap. 7). One might expect the rich, who can more easily find private substitutes for government goods, to favor a high degree of unanimity, which protects them from the poor, who might try to vote a tax and decide the kind and level of a jointly used good.

There are many ways of keeping the political system safe for those with money to buy what they want. Some of these were discussed above, such as the question of agenda setting and determination of voting boundaries. Perhaps the biggest question of all is what to do about two people with money who want different amounts or kinds of non-optional or pre-emptive joint-impact goods.

There is another problem with the Pareto-rule that says resource and money-income ownership are inviolate. It is not popular in the western democracies to deny explicitly the one-person/one-vote rule with respect to political choice, but this ethic is potentially in conflict with the Pareto-criterion. For the two rules to be consistent, voting power would have to equal the distribution of income. Where they are unequal and unanimity is not required, any majority voting on the government budget and taxes is unlikely to be Pareto-better. While even the rich may agree to some rule of less than unanimity, it is unlikely that all people will want the same rule. If we turn goods provision over to politics, the Pareto-criterion will not be met.

There is discussion about which kind of voting system best represents the median preference with given money income or which one best reflects intensity of preferences.16 Still, few are willing to oppose the ethic of equality with respect to voting power, which is in conflict with protecting existing distribution of income (Habermas 1975, Wolfe 1977). We have seen some non-Pareto-better redistributions of income via extension of the voting franchise, though political choice is still reflective of income via campaign finance. Conservatives have long argued that joint-impact goods are a special case, and few real examples can be found. James Buchanan (1968, chap. 9) has developed a five-part categorization of goods in terms of jointness--he calls it "indivisibility."17 He argues that only the case of pure indivisibility may be reason for government provision, and that of course is not necessarily best in the face of political externalities. Richard Musgrave (1969, p. 142) has opposed Samuelson's definition of joint-impact goods as all goods not on the knife edge of no jointness characteristics. This definition makes incompatible-use goods the rarer, special case, with most goods having some degree of jointness. This definition even makes Samuelson uncomfortable, and he admits, "If the experts remain nihilistic about algorithms to allocate public goods, and if all but a knife-edge of reality falls in that domain, nihilism about most of economics, rather than merely public finance, seems to be implied" (1969, p. 109). Note that this is a nihilism over the high-priest function and not economics as a predictor of substantive performance.


It has been emphasized here than rights are antecedant to exchange and thus Pareto-better exchange can't be a guide to choice of rights. This does not mean that rights are settled once and for all or in discrete moments of court or legislative action. People discover further independencies in the context of everyday business. They not only exchange what they own, but continue to explore and decide what each owns. Many reach localized accommodation as they go along (private ordering) without any recourse to court or legislature (Williamson 1985, p. 164). The common law is a formal ratification of many of these accommodations. Different forms of business organization may produce more or different decentralized accommodations.

The fact that these decentralized accommodations have lower transaction costs is relevant, but not decisive for either welfare economics or prediction of institutional change. To find economizing behavior is not to settle the question of what to economize. One person's transaction cost is often another's opportunity. The choice among institutions is not alone one minimizing transaction costs if the total performance of who counts is thereby different.


Conventional theory draws attention to a very narrow set of institutional variables. It presumes that there are only two main sources of power: (1) the distribution of money income and factors of production and (2) the degree of competition as influenced by the number of buyers and sellers. By assuming that an ethical judgment had been made on income distribution via factor ownership, the welfare theorist felt free to advocate the competitive market as controlling all other sources of power. The institutional variables suggested by conventional theory are too limited for empirical research on substantive performance and for an adequate welfare economics.

The argument against the simple prescriptions flowing from consumer sovereignty concepts can be stated as follows: You cannot tell what a person is sovereign over until you specify fully how property rights affect all of the ways for one person to affect another. Some of the points can be summarized as follows:

1. The point of the argument is not to deny that competition or ownership of factors are relevant institutional variables. For incompatible-use goods, the degree of competition goes a long way in controlling the power that one person has over another even when one owns many factors and the other few. But this is far from the whole picture.

When incompatible-use goods are not produced under constant costs, the real income of a person is affected by the prices of goods, which are affected by the preferences and purchases of others. In other words, pecuniary externalities affect the real value of money income. The greater the variation in preferences, the less money income is a measure of real income. Further, there are many other ways that people are interdependent. Different rules affect the size and impact of information costs on different parties. The same point can be made with respect to all varieties of transaction costs. These costs affect how far a person's money income will go and what can be done with owned factors. The other varieties of interdependence discussed in Part II create additional sources of power of one person over another.

2. With respect to joint-impact goods, there are additional institutional variables. In order to describe a person's opportunity set, we must know who gets to choose nonoptional joint-impact goods and how the costs of optional joint-impact goods will be shared. Government could charge people different prices (taxes) for the same good. But this surely requires an additional ethical judgment beyond specifying money income. The real utility of this income is unknown until the cost share (tax price) for joint-impact goods is settled.

Samuelson makes it clear that decisions on both money income and the pricing rule for joint-impact goods are necessary when policy is made on welfare distribution. Thus are destroyed any simple institutional prescriptions that assume the government has settled all problems of sources of power and interdependence between people when it approves of money income distribution.

The discussion in this book has added several implications to those of the Samuelson model. In thinking about our ethical judgments concerning the relative welfare of different people, we must be concerned with not only money income and pricing rules for joint-impact goods but also a whole host of rules that control interdependence growing out of exclusion costs, transaction costs, scale economies, consumer and producer surpluses, and peak load conditions.

Thus the world becomes more complicated, and a person can think his welfare is improved when he gets one variety of rule going in his favor only to find that welfare lost by another sort of rule going in the opposite direction. The narrow set of institutional variables suggested by conventional theory works to some group's advantage. While others are guarding the rights identified in the conventional models, their welfare can be reduced by changes in implicit rights in other areas.

These general points are examined further below in the context of a number of specific propositions on institutional choice that can be found in the literature. They purport to be prescriptions scientifically deduced from a simple normative judgment of the relative ownership of income, factors, or goods. But because of the inadequacy of ownership specified in this way to control all sources of interdependence, it turns out that these propositions make a hidden presumption of additional normative judgments that must be accepted before one can deduce a prescription for choice among alternative institutions.




In the following section, a summary and critique will be made of some of the suggestions in the literature that have been put forth to guide institutional choice under the guise of increasing something called social welfare. They are not mutually exclusive. The first several of these are a summary of the rules implied in the previous two sections of this chapter.

Prohibit all but Pareto-better change. The Pareto-criterion does not escape the need for a moral choice among the conflicting interests of people. It implements the choice already implicit in status-quo property rights. To limit oneself to Pareto-better change only is to express preference for the current distribution of rights and income.

Pecuniary externalities should be ignored. This restatement of the rule above is subject to the same critique. As Armen Alchian (1965, p. 818) puts it, "Although private property rights protect private property from physical changes chosen by other people, no immunity is implied for the exchange value of one's property." Pecuniary externalities are those produced by the play of the market. If the value of your assets is destroyed by competition, that value is incidental to someone else's Pareto-better trade. One problem is that transactions often have third-party effects. Property rights determine whether these interests are to count and be included in the Pareto-better transaction. Some people's interests are protected by rules of theft against technological externalities. Other people's interests are protected by rules affecting market power, which in turn affects the value in exchange of one's assets. To argue only for protection from technological externality is to argue for the interests of one person over another (see Chapter 7).

The market will internalize all externalities that should be accounted for. This is another restatement of the Pareto rule. The market by definition will allow A to make a bid to B to avoid or secure some effect of B's actions. A's interest is internalized and becomes an opportunity cost to B (see Chapter 8). If A's bid fails, the consequences of the original distribution of rights are carried out. Another way to express this rule is to say that government regulation is bad. It is illogical to grant B a right and then take it away in a regulation. But public choice gave B the right in the first place, and this choice does not necessarily hold for all time. What the government gves it can take away. And when it wants to take away a part of what was given to B, rather than the entire right, it turns to regulation to give some rights to A. To argue against regulation is to argue for the supremacy of B over A. Regulation is not to reduce the freedom of all people. Freedom to choose within one's opportunity set must be distinguished from freedom to choose one's opportunity set. A's freedom is B's regulation.

The government must correct all externalities. Robert Haveman suggests that "when significant spillovers are found in the real world, collective action, usually through a government becomes necessary if efficient performance is to be secured" (1970 pp. 39-40). Another version of this is that the government must step in whenever there is a divergence between social cost and private cost. Harold Demsetz (1967, p. 345) says, "A primary function of property rights is that of guiding incentives to achieve a greater internalization of externalities." (See also Baumol 1965, p. 25.) But there is no way to remove all externalities; they can only be shifted. If A's acts are causing damage to B, a partial analysis may suggest that we do something about it. But if their interests conflict, the only thing to do is to choose between them. Either allow A to continue to affect B, or give B the right that leaves A exposed to B's acts. It loads the argument to label A's interest as a social cost and B's as a private cost.

Goods with high exclusion costs must be provided by governments. The market will not reflect everyone's willingness to pay for these goods, as people are tempted to be free riders or conclude their action would not make any difference anyway. The fact that people do not buy a good in the market does not necessarily mean they do not want it. However, if the government taxes to provide the good, the free rider may be turned into the unwilling rider. While some may hide their tastes, others truly do not want the good but may be forced to buy it (see Chapter 3).

Joint-impact goods should be provided by government. Whether exclusion is relatively cheap or not, utilization of a joint-impact good by A does not reduce the potential availability to B by an equivalent amount. To prevent B's use even where possible is wasteful where marginal cost if zero.18 Yet A and B may not agree on the quality and amount of the good that should be provided or on how they should share in its cost. A may want the good, and, if it exists, B may not be able to escape being affected by it even when it has negative value to B or B wants a different quality. Where interests conflict, any rule for provision of joint-impact goods whether private market or government must choose between the interests of the parties (see Chapter 5).

The usual prescription for allocative efficiency of price equal marginal cost (of another user) begs the question of who pays fixed cost. And, the usual prescription for production efficiency that the sum of the individuals' marginal values equal the marginal cost (of another physical unit) can't be implemented by public or private producers because of strategic non-revelation of demand. Even if known, the rule presumes the rights of those with negative values to be compensated (negative price). If a negative value is allowed to influence quantity, then it is a right whose value is a willingness to sell rather than a willingness to pay. This is a matter to be decided by public choice, not assumed.

Distribute ownership to minimize transaction costs. Where parties conflict over the right to exclude the other from use of an incompatible-use good, Richard Posner (1972b, p. 18) offers the following economic principle: "The right should be assigned to the party whose use is the more valuable .... transaction costs are minimized when the law ... assigns the right to the party who would buy it from the other party if it were assigned to the other party instead and if transaction costs were zero."19 Thus, in effect, if A has the necessary income and preference to buy a resource from B, A should be declared the owner so that the cost of transactions might be saved for the economy! B, of course, now has nothing to sell and may go hungry, but Posner evidently judges this to be of no consequence since transaction costs are minimized. The question of income distribution is only a zero-sum game anyway and of no interest to economists interested in maximizing total product. He is entitled to his value judgments as between the interests or A and B, but these should not be glorified by calling them an "economic principle." The so-called principle provides a method for allocating ownership to new resources at the margin but is silent on how ownership of previous goods (which supports present ability to bid) was determined. The government is not ethically neutral whether it declares A the owner of a new resource, auctions it to the highest bidder, or shifts rights from B to A even if the Kaldor-Hicks potential compensation test is met.

Some suggest that total wealth first be maximized and then redistributed according to some value judgment; thus some people would get their income via court-protected private property rights while others would get theirs only by the grace of some legislative body. Such a situation represents a very unequal status and requires an additional value judgment.

The minimization of transaction costs can be applied to a number of cases of institutional choice. It was noted in Ch. 6 above that collective (non-competitive) modes of organizing work, including unions, facilitate trust and enable cost savings in training and supervision. But, cost minimization has a context and the resulting sticky wages and macro effects are not to everyone's liking. Cost minimization by one set of parties often creates cost for others and it is rights that determine whose costs count.

Common property is inefficient. If goods are owned in common, trade is either prohibited or very costly since the agreement of all is required. This situation can prevent some asserted Pareto-better trades. Such a rule expresses the value judgment that the rights of third parties whose interests are furthered by high transaction costs are not to count.

A variant of the above rule is that private ownership is most productive. As Alchian puts it, "Under public ownership the costs of any decision or choice are less fully thrust upon the selector than under private property. In other words, the cost benefit incentives system is changed toward lower costs" under private ownership (1965, p. 827).

This conclusion depends on how productivity (output) is defined and what the rules require either public or private decision makers to take into account (that is, what costs or inputs are relevant). There is no such thing as the cost. Costs that count are what they are because of rights and personality. Both are malleable. The private owner may make more profit, but there are as many profitable solutions as there are property rights sets. The rules shape what actions are profitable. The public choice question is "What is to be made profitable?" (See Samuels, Medema, and Schmid, 1997)

Option demands must be filled by government.20 Where resource use is irreversible, present values may not reflect actual future demands, which if they develop will be frustrated by today's consumption. However, one person's options kept open are another's options foreclosed. If the Grand Canyon is kept wild for the grandchildren of current members of the Sierra Club, there will be higher prices for today's electric consumers. One person's options for children's future enjoyment is another's pain in the neck. To meet an option demand, like any other, is to choose between the competing interests of people (see Chapter 6).

Government investment planning is necessary with economies of scale and cost discontinuities. Where one firm uses inputs produced by the other, investment timing coordination can reduce both firms' unit costs and avoid unused capacity while the other firm reacts to price reductions. Such coordination is possible with a market firm by increasing the scale of the firm and sometimes by market contract. But decision costs may be significant in either private or public planning. If planning is done by vertical integration of ownership control, it may lead to restraint of trade, and if it is done by government, different groups will have an opportunity to shape the benefits of the coordination (see Chapter 6).

It is better to redistribute wealth in money than in kind. "General taxation, positive and negative, is the best way to moderate the inequalities of income and wealth generated by a competitive economy" (Tobin 1970, p. 276). Also see Collard (1981, Ch. 12). Henry Simons (1948, p. 38) states, "It is urgently necessary for us to quit confusing measures for regulating relative prices and wages with devices for diminishing inequality. One difference between competent economists and charlatans is that, at this point, the former sometimes discipline their sentimentality with a little reflection on the mechanics of an exchange economy." It can be easily shown that the recipient of a grant can reach a higher utility with the flexibility of money rather than the gift of a good. But the fact is that often the grantor's utility may be enhanced by changing the recipient's specific consumption. Consumption frequently has a degree of joint impact. What is efficient depends on whose interests count. Welfare recipients want money to spend as they wish, but some givers want to attach strings so the recipients' behavior will be changed.

Some writers say that redistribution is not a question that economic theory has any answers for. Yet they do feel free to pronounce that some techniques for redistribution are good and others are bad. If the prevailing public opinion is against direct grants via taxation because the receiver is seen as not earning them, then to restrict policy consideration to this method is in effect to argue for the status quo.

A free person should choose a unanimity rule. It follows that any government action by less than unanimity is a diminution of freedom created by force. This rule overlooks decision costs. There is a necessary trade-off between decision costs involved in getting unanimity for those things a person wants and the exposure to having the group choose something the person does not want. (Buchanan and Tullock 1962) Depending on one's image of this trade-off, a free person may vote for a majority rule. No voting rule is going to please everyone because everyone's trade-offs between reduction of external costs and decision costs including political externalities will differ.21 Again public rules will have to make a choice between people's competing interests. People may differ with respect to the voting rule they prefer.

A public referendum is the best reflection of public interest. One writer suggests that we should "pose issues of public policy in terms of whether society does in fact hold certain value judgments rather than in terms of the demonstrable inherent legitimacy of certain activities" (Steiner 1970, p. 39). But how shall we know these social value judgments? The same writer suggests that "survey data about public attitudes on issues exist and provide some sort of a base" (p. 45). Leaving aside all of the problems of who is to frame the questions for the survey, we still have the problem of what to do with conflicting attitudes. When pressed, most Western economists and political scientists fall back on democracy, and, the purer the better, after compromising with decision costs. But the concept of democracy does not settle the agenda-setting problem and the question of voting boundaries (see Chapter 8). Democracy is not merely to be in favor of the people. "Power to the People" is a demagogic slogan. The issue is always which people are to count when interests conflict.

The point is not to argue for or against the above propositions. Some of them fit my own value judgments, and some fit better than many others I could imagine. They are rejected as value-netural rules of choice, and as such they are lies. They would be honest only if they were put forward as the explicit value judgments of the speaker.


The lessons of the above point-by-point critique of rules for institutional choice can be summarized and extended by a critique of the conventional performance categories of freedom, efficiency, and economic growth. This critique serves to make the point of this chapter in a nontechnical way. It also illustrates why the conventional performance categories are not satisfactory for empirical studies such as those in Chapter 12 below.


What is meant by the common assertion that the competitive market maximizes freedom?22 Voluntary trade gives the appearance of freedom. If each party did not think he were better off, no trade would take place. Voluntary trade contrasts to government regulation, which has the outward appearance of force. But appearances often mask underlying factors of a different sort. The model outlined above inquires into a step previous to observed trades and asks how the parties obtained the rights to be traded. After it is decided that A has many rights and B has few, we can note B's efforts to rearrange his few goods into a better mix. If A agrees to trade, both are better off, but B's pile is still small relative to A's. Who says that this has to be? Why is it when A's interests conflict with B's in access to resources that A is more often the one found to have the right? Why is B mostly in the position of seldom creating a cost for A, while A has much that is a cost to B? Antecedent to the observed mutually beneficial trade is the prior mutual coercion that is given shape by the inescapable public choice of property rights. Where interests conflict in terms of this antecedent question of the rights of each party, increased freedom for A is more exposure for B. This is the significance of the quotation from Isaiah Berlin at the beginning of this chapter, "Freedom for the pike is death for the minnow."

The recipient of an order in an administrative transaction cannot voluntarily choose to abstain from the transaction. This situation nominally contrasts to a bargained, market transaction where parties are free to trade. However, where opportunities are unequal, the result is coercive. Robert B. Seidman asks, "Is it significantly different to say to a man, I will pay you a wage if you will work for me?' rather than 'You will get no wage unless you work for me?' Two hundred years ago Lord Chancellor Northington expressed it succinctly: 'Necessitious men are not, truly speaking, free men. ' At law, to treat unequals notionally as equal is in fact to elevate the stronger to a position of domination" (1973, p. 556). (This is also a major theme of Ely [ 1914, pp. 555-618] .) For a man of little property, the freedom not to agree to a wage offer is the freedom to starve.

There are some macro-descriptive uses of the term freedom. We may wish to contrast two countries with respect to the extent of participation in decision making. Some indexes of pluralism are possible without presuming a value judgment with respect to the conflicting participants. The issue is one of whose freedom rather than freedom in the abstract. The great moral choice in any society is whose freedom counts when interests conflict in the face of scarcity. Where people conflict, global freedom is without meaning and can only obfuscate the real conflict and the ethical question. When we look at human beings face to face as their personalities and preferences interact, it can be seen that the mutuality of the market is the second step built on the antecedent public decisions assigning the property rights that are then traded and rearranged. Freedom in the abstract is not a satisfactory performance variable.


Neoclassical theory has suggested that the market gives us not only freedom but also efficiency. A producer must search for the cheapest method to create his product or be driven from the market. The theory of the firm and production economics is a set of calculations necessary to determine optimal input and output choices. But, again, appearances are deceptive. The earlier discussion of costs and externalities (Chapter 1) is relevant. It is one step to minimize a set of costs and another to determine what effects to include in that calculation. Contrary to the suggestion of neoclassical theory, costs do not simply exist in nature but are selected by the public choice of property rights. Calabresi and Bobbitt (1978) refer to this as "tragic choices". It is property rights that determine what physical effects are to be accounted for by decision makers. It is property rights that exclude some people's misery from the cost-minimization calculations of business. Of course, to include it for A is to shift it to B.

A third step focuses on the rules for making rules, or what is sometimes referred to as the constitutional level. This hierarchy of decision levels has been implicit in the discussion thus far. (A similar distinction is made by Ciriacy-Wantrup 1967, pp. 181-84, and 1963, chap. 3.) The first level is that of specific combination of inputs and outputs by the private or public firm or individual. The property rights rules of the second level are used by decision makers of the first level in making their optimal choices. Decision makers of the first level try to influence these property rights by participation at the third level according to political rules. Efficiency calculations are most meaningfully applied to decisions at the first level.

Economic efficiency is a concept from engineering and physics with values attached. It is simply an abstract expression of the ratio of values of selected input to selected output. It is property rights that do the selecting. It is the relative opportunity sets of individuals based on choices of antecedent rights and their choices from within those sets that determine the content of the input and output categories and influence their market value. In other words, choice in period two is a function of choice in period one. If effects X, Y, and Z are included in one efficiency ratio but only X and Y in another, then the efficient use of resources will be different with each (in the second set, effect Z is priced at zero as seen by a decision maker). Efficiency is the derived result of the prior choice of the content of the input and output categories, and not a guide to the choice of alternative sets of categories. In short, there are an infinite number of efficient solutions, each subsidiary to the choice of property rights, which define the content of input and output. Rights determine efficiency, not the other way around.

It is useful to speak of the efficiency with which a given institutional rule achieves a given performance objective. But, for clarity, objectives need to be explicit. Efficiency and rationality are closely related. We often speak of someone being irrational when we presume some objective is being sought but do not see that the decision maker had other things in mind. A favorite academic critique of public policy involves the discovery of apparently inconsistent choices. For example, the United States had a policy of subsidizing irrigation development at the same time it was paying other farmers to reduce acreage planted because of surpluses. These choices are inconsistent in terms of overall food policy. Policies such as this emerge from the composite processes of the Congress as members search for laws that will achieve a winning coalition of votes (Ingram 1969). Such policies are hard to rationalize in any terms except that of political expediency, but they may have utility in keeping the peace and preserving willing participation in the group. It may be necessary to buy support of Westerners for an agricultural bill by paying them off in irrigation projects that reduce the overall effect of the bill. There may be other ways to make this trade, but then again there may not or they may be too costly to transact. Not all payments can be made directly in money but must be made in terms of policies that establish simultaneously partially conflicting rules and public spending.

One of the performance elements of political transactions is what Talcott Parsons (1965, p. 16) refers to as system maintenance. It is no good to maximize some material output if the country then erupts into civil war. System maintenance is hard to test empirically; it is difficult to be sure how any given trade-off contributed to maintenance of the community as a going concern. (This was discussed in Chapter 2.) This concern can be indiscriminately used to rationalize any decision a public body makes. Still, for all of its difficulties, it cannot be ignored.

Some people prefer a strong role for central planning to avoid inconsistent policies. Theoretically, central planning implies that one group's interests have been chosen along with consistent means thereto. However, most existing planned economies seem to have their share of contradictory policies also. Sometimes conflicts of interest are resolved by having clear winners and losers on each separate issue. A given group either loses consistently and its participation is coerced, or it wins some and loses some, so that overall it agrees to play the game. However, one can observe that agreement and acceptance are obtained where something is done to favor A, and at the same time the effect is lessened by simultaneously passing a partially conflicting policy on the same issue to favor B.

In sum, where there are conflicts of interest, it is not possible to ask only in general whether an institution is effective (efficient), but one must also ask of its effectiveness for whose interests. Efficiency in the abstract is not a satisfactory performance variable.

Productivity and Economic Growth

These performance concepts are closely related to that of efficiency already discussed. It is common to say that property rights should be structured in such a way as to promote economic growth or maximize the net value of production. But there are problems in using economic growth as a performance variable because of its ambiguity.

How shall we know if institutions are enhancing productivity? A number of neoclassical writers have suggested a conception involving the divergence between private and social cost (Pigou 1946, chap. 9). The idea is that an individual may not account for the effects of all of his acts, and thus there is a difference between what he accounts for and the effects on society. The criteria for evaluation then focus on the costs and benefits of removing the difference between private and social costs.

Douglass North and Robert Thomas (1973, p. 4) have developed this theme and provide an example from the economic history of Spain. The earliest use of land was for sheep herding. The crown had granted the right of grazing to the shepherds' guild. As population increased, the products of a fixed-place agriculture became more valuable. Profit was possible in farming. However, anyone who improved and planted a field ran the risk of its being trampled by wandering herds. Evidently, the right was nontransferable, and there was no way for the farmers to acquire the secure use of the land, or the cost of organizing bids and communicating them to itinerant shepherds was too high. The conception of these writers is that the benefits captured by the shepherds did not consider all of the costs to society that seem to prefer less sheep and more agricultural products. Does this gives us a basis for saying that such a common property rights arrangement is detrimental to society and economic growth?

It is hard to argue against a consideration of all costs and an arrangement that maximizes net social benefit. But the language is better suited for its emotive than for its analytical value. If interests conflict in the face of scarcity, what does it mean to account for all costs?

When A's opportunity means a cost (forgone opportunity) for B and vice versa, then it is impossible to consider all costs simultaneously. What North and Thomas call considering all costs is a narrow theoretical concept defined as the ability of one party to make a bid for the rights of another. A potential bid of farmers frustrated by market prohibitions is labeled a social cost. But this lost opportunity for farmers is selectively perceived while other forgone opportunities are ignored. If land is made marketable, who gets to participate in the decision to sell? Some shepherds may have large fixed assets or great psychological commitment to herding as a way of life, while others may not. Thus, reservation prices (acceptable bids) will differ among individuals. Whose opportunity will be forgone? The right of participation will affect whether the farmer's bid will be accepted. Value of net output is not something independent of rights something that can be used to deduce a correct rights system, but is itself a partial function of rights.

The so-called social perspective is an attractive one, especially to consumers who apparently have let it be known in the market that they prefer more agricultural products and fewer sheep. Apparently the value of output would be higher and consumers would benefit if the land were owned by farmers--especially if the consumers do not care about any losses to fixed assets incurred by the shepherds or any costs to the mental health of the shepherd from being uprooted and changing jobs. One can imagine a Roquefort cheese lover who will not appreciate the reduced supply and higher price he has to pay for a staple of his diet. It is fine to say that rights should be a function of relative costs, but which costs are to be considered? Is the farmer also to pay the shepard's and cheese lovers costs when acquiring the land? If so, it is possible that what first appeared to be a profitable output-expanding transfer is that no longer. It depends on what is included in output, and what is considered as growth. Cost is not something given but something selected by public choice of property rights.

The above should not be interpreted to mean that nearly universal gains from the battle against nature are impossible. But no analyst can sit on high and guess that the gains exceed the costs and blithely call it social progress, leaving distribution issues to others. Some want to say that the first consideration is to maximize output and then let the political process or whatever available institution decide how to divide output. But it is the political process of establishing property rights that defines growth, efficiency, and maximum output in the first place. These terms derive their content from property rights decisions. Those who are necessarily displaced when one type of growth is chosen over another must be asked if they prefer a share of the new product to the personal peace they enjoyed before (Morgenstern 1972, p. 1169). Only then are we sure everyone gains rather than just hiding our preference for one group's taste over another's. However, such a requirement preserves the status quo.

To conclude, the conventional performance variables of freedom, efficiency, and economic growth are misleading when employed for normative purposes and are unsatisfactory for empirical use. The concepts of social cost and economic growth turn out to be ambiguous. While it is useful for the analyst to search for common and widely held values, be careful of a partial analysis that ignores human interdependence in the face of conflicting interest. It is often argued that the test for an acceptable change in rights is whether any losses can be fully compensated by the gainers. But who gets to say what effects must be accounted for? This is part of the problem of public choice, not something to be presumed.


Some scholars argue that institutional alternatives should be and are chosen in terms of total (global) benefits and costs.23 This argument is reflected in some of the examples used above such as the Spanish sheep-herding conflict, and its shortcomings can be further illustrated with another example. North and Thomas (1973) utilize a total benefit-cost analysis both to explain historical changes in rights and to justify them. They refer to piracy, which once preyed on the shipping of the Mediterranean. They note that shippers paid bribes to avoid a worse fate. These are not neutral words of description. This language assumes which acts are theft and thus makes a presumptive value judgment as between competing interests. By this logic, the subject of Russell's cowboy and Indian painting described in the Introduction would be labeled as theft by the Indian rather than a rent collection for use of the Indian's land.

An alternative is to regard the "pirates" as owners of the sea; then the bribes become market rents. North and Thomas (p. 4) see everything as a function of natural costs. The bribe was used is long as it was cheaper than building a navy to drive away the pirates. They see the navy as ensuring a right, while it can as well be seen as theft, which violates the previous rights of the pirates. They say, "Ultimately priacy disappeared because of the international enforcement of property rights by navies." They assume the subject of their analysis before they start by regarding the traders and shippers as the owners of the sea and pirates as thieves who couldn't compensate the shippers. They see the demise of the pirates as economic growth. Yet it can also be seen as a redistribution of rights at the expense of the former sea owners. Who considered the loss in food for the pirates' families as a cost? If the "pirates" own the sea, then the navy is the thief. Can we say at that point that the shippers were doing something useful while the pirates were not? On that ground, no one could own natural resources and collect rents from them. Any payment for use would be seen as a bribe or tribute to landlords.

Initially, bribes were cheaper than maintaining a navy. "However, with the expansion of trade it ultimately became evident that the complete elimination of piracy was the cheapest alternative" (p. 4). Cheapest for whom? For shippers or pirates? Pirates could be like any landlord. If landlords were eliminated, the distribution of wealth would be different. But that requires a value judgment and would not have the widespread appeal that efficiency in the abstract attracts.

The Circularity Problem

The school of thought that argues that institutional alternatives should be and are chosen in terms of total benefits and costs and given relative prices relies on potential Pareto-improvement tests (Kaldor-Hicks) and has all of the problems noted above plus the fact that some people are actually left worse off (compensation not paid). The potential compensation test can't be a guide to shifting rights without presuming who had rights before the shift. All propositions related to the analysis depend on an implicit assumption of who counts when interests conflict. The analysis attempts to choose alternative rights at one level while taking another level as given and unexamined. Efficiency calculations always depend on where you start, but they can never validate that starting place. Therefore, a benefit-cost analysis of alternative rights is always a partial analysis. Efficiency calculations always presume some set of rights and therefore cannot be a guide to rights, unless the prior rights are legitimated. Partial analysis has its place as long as its presumptions are clearly labeled. The "circularity problem" can be avoided only by analyzing how rights affect a particular group interest (substantive results).

Another example is provided by what is termed the theory of "induced institutional innovation" (Hayami and Ruttan, 1985). Institutions are seen as endogenous to the economic system and change results from exogenous change in factor prices influenced by technology and population. This theory predicts that efficient institutions will develop to make it possible for marginal cost to equal marginal revenue which may have been thrown into disequilibrium by exogenous changes.

Hayami and Ruttan (pp. 99-103) illustrate their argument by reference to a rice farming village in the Philippines between 1956 and 1976. The traditional property institution provided that people of the village who participated in the harvest and threshing received one-sixth of the rice. In 1958, a publicly built irrigation system was made available to the village, and in the late 1960's, high yielding rice varieties were available accompanied by fertilizer, pesticides and improved cultural practices. The effect of these new technologies was to raise farm income and given the traditional rights, tenants and other laborers got some of this gain as well as the landlords. In Hayami and Ruttan's view, this created a disequilibrium and labor received more than its marginal product and more than its opportunity cost in other occupations (declining as a result of population growth).

Hayami and Ruttan argue that this situation induced institutional change, namely a system where labor could get a share of the harvest only by giving uncompensated weeding labor. This in their view lowered the marginal cost of labor to its opportunity cost. The resulting performance gave all of the benefits of technological change to the landlords. Labor is still to be thankful, however, because if the old rights had prevailed and labor cost kept high, there would have been less employment. Thus the capture of the gains by the landlords achieved efficiency by restoring the equilibrium between marginal revenue and cost.

What would have been the substantive performance if landless labor had been given part ownership in the benefits of technology? It is only selective perception that regards a claim on net gain to the firm as changing the marginal cost of labor. A share of net return is not a marginal cost. Both landlord and labor will decide to work in the firm only if their marginal value product is greater than the wage they could earn elsewhere. Being owners need not change that calculation. A laborer who is also a stockholder does not change their own or the firm's calculations of resource allocation. Equilibria between marginal cost and marginal revenue is not unique to one ownership interest and thus cannot explain change in ownership. The particular equilibrium among many possible is rights dependent and can't explain change in rights.

A test is provided by the same Philippine case. A law was passed giving former tenants the right to pay a fixed rent. As a result, these tenants assessed their opportunity costs (including value of leisure) and subleased the land at a rate that gave the hired labor only low non-farm wages. This contracting achieved marginal equilibrium again with landless labor getting its non-farm opportunity cost only. This equilibrium gave the original tenant a share of the productivity gains along with the landlord. If the original tenant can be made a part owner why not all the landless labor in the village? There is no theoretical reason that they cannot be beneficiaries of public investments in irrigation and new plant varieties as well as landlords and original tenants.

Induced institutional innovation theory is a hidden presumption of desired income distribution. It tries to portray rights as derivative of only natural forces. Any human and cultural factors therefore only affect the speed by which presumptively efficient institutions are reached or sometimes result in bad institutions which do not fit nature. This ideology masked as science is part of the power struggle used by different groups to obtain institutions favorable to them. There is no way to have a welfare economics that does not require the taking of sides.


Society puts pressure on its scholars to provide authoritative answers to policy questions in the face of conflict. It is a measure of the insecurity of our civilization that these demands are made. An example of one such demand is contained in a letter dated October 12, 1966, sent to experts in the field of land-use policy by letter Milton Pearl, the director of the Public Land Law Review Commission.

To assist the Commission in achieving its objective, we are seeking

the advice and counsel of persons like yourself, who are concerned with government, policy development, and public administration as they affect our political, social and econoinic well-being. Specifically, we solicit your assistance in helping us establish criteria that will serve all members of the Commission alike in reaching conclusions as to what constitutes "the maximum benefit for the general public." Such criteria would put decision-making within the Commission on a plane above reliance on divergent opinions arrived at without reference to a common base.

The psychologist Erich Fromm (1941) describes points in history where people run

from freedom and seek authoritative guidance. There always seems to be some dictator ready to answer our call. When people cannot take the responsibility for their destiny and live with the fear that they might later wish they had chosen another path, someone will come forward to announce that they have found the true way. Today's demands on scientific authority are not as dramatic as the German turn to Hitler or its equivalent in other countries suffering the economic crises of the late 1920s and 1930s. Still, economists, social scientists, and others are called upon for scientific answers to today's conflicts.

The title for this section is taken from a book by Joan Robinson, who writes, "The Keynesian revolution has destroyed the old soporific doctrines, and its own metaphysics is thin and easy to see through. We are left in the uncomfortable situation of having to think for ourselves" (1963, p. 95). She continues,

Perhaps all this seems negative and destructive. To some, perhaps, it even recommends the old doctrines, since it offers no "better 'ole" to go to. The contention of this essay is precisely that there is no "better 'ole." The moral problem is a conflict that can never be settled. Social life will always present mankind with a choice of evils. No metaphysical solution that can ever be formulated will seem satisfactory for long. The solutions offered by some economists were no less delusory than those of the theologians that they displaced (p. 146).

The message of this chapter is similar. The rules put forward by the new welfare

economists are simply disguised value judgments and preferences for the interests of A over B. It would seem elementary logic to note that there cannot be a total benefit-cost calculus when interests conflict. Yet we seek it with the fervor of a sea captain searching for the lighthouse beacon.

When interests conflict, there must be a weighting of these interests. But who is the authority for that weighting? Whenever we speak of society, of the public or the people wanting something, we implicitly have chosen one side or the other of the interests at conflict. We have made a selective value judgment without saying who gave us that right. If we say we have just observed some basic agreement and have derived our conclusion from it, we may be guilty of self-delusion. To attach the term "collective will" to some existing governmental decision or result that oozes out of a series of interacting individuals' decisions is only to glorify what exists, not to provide a test for it. The decisions of Congress can hardly be a guide to what Congress should do. The same is true of any decision. It is always a function of some set of rules that are selective of which conflicting interests to count. There is no way to break out--one is always chasing one's tail. You can use a vote to confirm the legitimacy of a market rule, and you can use a public opinion poll to confirm the rightness of the voting rules, and so on ad infinitum: But you are still left with the need to choose between conflicting interests.

Peter Steiner (1970, p. 40) suggests that "we agree first (some way, any way) on collective values and then use them to make social choices." Any way, indeed! He suggests that we research the "revealed objectives of society instead of the derived ones" (p. 51). What he fails to see is that all revelations come through some institutional system and reveal as much of the institution as of some mystical public will. Change the rules, and you will get another revelation and another public will. Steiner was close to the truth when he noted, "Obviously the question 'What is the public interest?' has no simple answer. Indeed, asking the question invites the sort of smile reserved for small children and benign idiots" (p. 54). Indeed it is an illogical question altogether and not just a matter of acknowledging the complexity of the answer and promising that one day clever research will finally reveal the true rock of common value.


There is no better hole to go to. There is not even any given hole. When interests conflict, to ask the question of what constitutes social welfare in total is to deny that conflict. There is no better hole because the question is inconsistent with the reality of human differences. (This is not to deny that at any given moment a given society has widely shared values. In this case, the rules affecting who chooses do not matter since the choice is the same whoever chooses.) In the face of differences, the question assumes away the agony of moral choice of who is my brother. We stand naked in our differences without the clothing warmth of the high priests, economist or otherwise. People must and will choose their accommodation, or they fight, or they live in suspended hate and unwilling participation. If we insist on cloaking our choice in the metaphysics of social welfare, global efficiency, freedom for the people, or nearness to God, I can appreciate the chill that drives us. Perhaps the human animal can never stand complete self-consciousness and needs self-delusion and ceremonialism to get up in the morning. Still, they make our discourse more difficult.

It is not in the realm of science to abolish conflict. Differences there will always be, but we are capable of understanding conflict.24 So much human conflict over proposed rule changes is simply uninformed.25 People have little warranted knowledge of the range of real performance consequences of alternative rules. Such knowledge does not remove the conflict, but at least if one has such knowledge, one can make sure the conflict is over real differences in the type of world different people want to live in rather than over some mistaken notion of what kind of world a new rule would likely produce. The scientist can play the role of provider of consumer and voter information, which can be the background for people choosing (learning) their preferences both of goods (performance) and of institutions. This role does not short-circuit the process by which people make choices by telling them they have already really made the choice that will now be made known to them along with the derived deductions as applied to the case in question. Of course, provision of information is not value free, but it is of a different order than the value presumptions hidden in the choice rules discussed above. Information on impacts of alternative rights is itself subject to interdependence created by information costs. Those with access to cheap impact data can use it to influence public choice of rights.

To conclude, there is nothing in this chapter that should be interpreted as being against the calculation of specific efficiency when objectives of the individual chooser are made explicit. For the most part, the chapter is an argument against presumptive choices among conflicting interests contained in theories and calculations of global efficiency. Whenever there is conflict of interest, to speak of global efficiency is to make a value judgment weighting the interests of one party over another. The plea of this chapter is for the speaker to make explicit the weighting expressing his or her own value judgment or that of the client being served. Inquiry into which institution/right is better can be clarified by asking, "Better for whom?" Each must apply his or her own answer. This chapter is an argument against the high priest role of economic analysis but not an argument for or against a different set of values than those presumed in neo-classical theory of any other. Much still can be done in predicting the substantive consequences of alternative property rights for the various parties.

It is to this task of providing warranted, empirical information of rules consequences that we now turn in Chapter 12. Our objective, however, is modest. There is no presumption that informed choice assures wisdom or peace.


1. What is included in this section would be placed in most texts under the heading of "market failure." The term has been avoided so as not to prejudge the issues.

2. Since concepts are value imbedded, any analysis unwittingly has judgmental elements.

3. Or government if controlled by a unanimity rule.

4. Which some people also regard as theft. Recall Pierre Joseph Proudhon's (1970) definition of property as theft.

5. The payment that is made is an externality, but Pareto irrelevant. The Pareto-relevant have been internalized.

6. "Economists can also discount the idea that it is necessary to know who counts, or who is imposing upon whom or who has the property rights before one can usefully make efficient analyses of these troublesome joint products" (Seagraves 1973, p. 619).

7. Economists are nearly united in their opposition to common property. See, for example, Cheung (1970) and Demsetz (1967).

8. The symmetry could be restored if it were equally easy for B to interfere with A's use of the sun. But, given their location, this is impossible. However, B could do other things, such as blow up A's building. But these opportunities are usually subject to injunction or criminal penalties. But if B could create a harm to A and the court would assess only a $10,000 damage, and if A is aware of this possibility, A will not interfere with B's sun rights and B is the resource user whether A or B is the original owner. B creates and pays $10, 000 damage to A to prevent the loss of a resource worth $15,000 or buys the resource outright for $10,000.9. This point is extensively developed in Bartlett (1973). Bartlett asks what Pareto-optimality means when resources are used to influence demand as well as cater to it. Also see Goldberg (1974), pp. 464-74.

10. There is nothing above that cannot be found in the literature. Even writers sympathetic to Coase will note in passing or in a footnote the many qualifications of the Coase rule. Yet the message that they wish us to remember seems to obscure these qualifications. For example: "To sum up, then: when negotiation is possible, the case for government intervention is one of justice not of economic efficiency" (Turvey 1963, pp 309-13).

11. One such meter is "the demand revealing process" (Clark tax). See Tideman (1977).

12. If A refuses to buy three channels for a price of $5.50, the producer may only provide two channels, which is not "efficient."

13. If those who don't work get into the movies free, maybe those paying will decide not to work.

14. Nancy Ruggles (1968, p. 38) says, "Every pricing system results in some sort of income distribution .... In choosing a pricing system it thus becomes necessary to make specific assumptions about interpersonal comparisons of utility, and then to judge the pricing system in relation to these assumptions as well as in relation to the marginal conditions."

15. A term used by Breton (1974, p. 4) to indicate the definition of equilibrium conditions that are incapable of implementation by any known institution.

16. For a review of the literature, see Winfrey (1973, ch. 3), Ochs (1974, ch. 3), and Mueller (1979, ch.6).

17. This has been sharply analyzed by Burkehead and Miner (1971, pp. 138-39).

18. Francis Bator (1962, p. 7) says, "The point is clear enough--public good and decreasing cost phenomena cause private decisions to go wrong." He hastens to add that inefficient markets may still do better than any alternative.

19. Posner suggests that since courts have used this right in the past in formulating the common law of nuisance, they should use it in the future. Since he criticizes a random assignment of property rights in the face of high transaction costs, it follows that he regards an equal initial distribution as inefficient.

20. Krutilla and Fisher argue that early discussions of wilderness retention policy "could be based only on what some persons might regard as rather vague equity considerations." But they claim that "although we do not make any assumptions about the priority of rights, results (favoring nondestructive use) . . . are obtained solely from considerations of efficiency" (1975, p. 73). They do not make clear whose efficiency they are implementing.

21. John Rawls (1971) suggests that widespread agreement on constitutional rules could be obtained if people had to decide them before they knew their own position and payoffs in the ensuing economic game. The suggestion of Rawls is that people will favor equality. This seems doubtful, for people can and do differ in their concept of a fair game since some are risk-averse and others seek the thrill of gambling for an improbable, but large payoff. For a critique, see Gordon (1976).

22. See, for example, Hayek (1944); Friedman (1962); and Stigler (1975). Friedman says it is a loss of freedom for government to explicitly have an income distribution objective (equity). "Freedom is my god." This implicitly accepts current rights distribution and does not avoid equity choices. For a critique, see Samuels (1976).

23. This work has been collected in Erik G. Furobotn and Svetozar Pejovich (1974). Also see Anderson and Hill (1974), and Posner (1977). For a critique see Ogus (1983), Veljanovski (1981a), and Kornhauser (1980 pp. 163-80).

24. "It is not the business of political philosophy and science to determine what the state in general should or must be. What they may do is to aid in creation of methods such that experimentation may go on less blindly, less at the mercy of accident, more intelligently" (John Dewey 1922, p. 34). Also see Johnson (1986, Ch. 9).

25. Kenneth Boulding (1973, p. 63) has formulated what he calls the law of political irony: "Political conflict rests to a very large extent on a universal ignorance of consequences, as the people who are benefited by any particular act or policy are rarely those who struggled for it, and the people who are injured are rarely those who opposed it. .... Bad definition and the failure of perceptual discrimination [ are] perhaps the most important source of bad politics."

If you have any questions or comments, please email

Back to Main Page