Daniel J.H. Greenwood

Home | Previous Page

Law Matters: A Critique of the Coase Theorem

Daniel J.H. Greenwood -- DRAFT -- October 25, 1990

The Coase Theorem has caused an uproar in legal academia by its suggestion that in the absence of transaction costs, legal rules would not matter.1 Transaction costs are as central to the real economic world as gravity is to the physical; policy decisions can no more be based directly on a transaction cost free model than NASA could base its planning on a model of a world without gravity: in such a world a cow could jump over the moon.

Nonetheless, friction-less models can teach us about our very different world. Coase contends that his model's main contribution is to demonstrate that transaction costs are important. Coase, p.8. It can also teach us that seemingly neutral legal rules are powerful social judgments with far-reaching consequences for the society we live in -- and that those judgments cannot be avoided. This conclusion is precisely opposite to the one Coase draws from his theorem; if correct, it casts doubt upon the entire law and economics enterprise.2

Let us return to Coase's world of no transaction costs.3 Assume a situation in which one individual, Capitalist, wishes to build a factory that will emit vast quantities of acidic pollutants. A second individual, Green, prefers to earn her living writing novels in her pristine pine woods some 200 miles down wind. Further assume that pollution control equipment is available for a price of $1 zillion ($Z). If Capitalist installs the equipment, no noticeable pollution will reach Green's forest; if he builds the factory without the equipment, Green's forest will die. In order to emphasize that neutral, efficiency promoting law is impossible even without considering externalities or other market failures, assume as well that no one other than Green will be bothered by the pollution or take pleasure in Green's forest (or Capitalist's factory) except to the extent that such pleasure is reflected in the respective market prices.

I intend to show that even in the absence of transaction costs, it is the legal rules, not the naked preferences of these two actors, that will determine whether the factory is built. Coase's contrary conclusion stems not from his assumption of no transaction costs but from other distorting assumptions.

Consider first a rule giving Capitalist a property right to pollute. Under these circumstances, in the absence of transaction costs, the factory will be built unless Capitalist is willing to sell his right to pollute and Green is willing to buy it. Coase's key insight is that $G, the amount that Green is willing and able to pay Capitalist to save her forest, is an opportunity cost for Capitalist: by building the factory without pollution controls, Capitalist is forgoing $G. As a wealth maximizer, Capitalist should be indifferent between receiving $G from Green or an equal amount of profit from the factory. Accordingly, Capitalist should be willing to sell his right to pollute if $G is greater than the lesser of $Z, the cost of the pollution controls, or $P, the profit he expects to make from the factory in the absence of pollution controls.4

In short, the key issue is the relation of $Z and $P to the amount Green is willing and able to pay to preserve her forest ($G). If $G is greater than either $P or $Z, the optimal result is for Green to purchase either Capitalist's pollution rights or his building rights and save the forest. If the factory is only marginally profitable or pollution controls are cheap, relative to Green's means, it is efficient for Green to save her forest. Absent transaction costs, she should be able to strike a bargain to allow her to do so: she will pay Capitalist either for his lost profits or for the pollution control device, whichever is less, and both parties will be better off.

Coase does not discuss the possibility that Capitalist might not be merely a wealth maximizing rational capitalist. If Capitalist is a true believer, he may actually want to destroy Green's forest. Perhaps he believes, like our former President, that trees cause pollution, and wishes to atone for the anomie his overly-regimented factory will cause by eliminating other socially undesirable miscreants, such as the tree-polluters. In this case, he may not be willing to forgo his right to pollute except for a payment substantially greater than $Z. Indeed, there may be no amount that Green can pay him that would make it worth his while to sell his right to emit his pollutants.

Similarly, Capitalist may value his factory for reasons other than its expected return. Even if his projected return, $P, is quite small, Capitalist may view the factory as a public service. Perhaps he wishes to serve his community by providing hundreds of jobs. He might then perceive Green's offer to pay him $(P + 1) to refrain from building the factory as, at most, a gratuitous insult to be summarily rejected.

Alternatively, Capitalist may not accept the standard economic assumption that opportunity costs are the same as real costs. Ordinary people generally do not treat forgone income as the same as actual costs -- as should be obvious to law professors, nearly all of whom have forgone far larger salaries as lawyers and yet do not think of themselves (and are not treated by others) as major donors to their employers. Forgoing $G may strike him as quite different from incurring a cash cost of $Z installing pollution controls and he may not treat the two numbers as comparable.

It is worth noting that whether for these reasons or simply because Green is unable to raise enough money for $G to be larger than the lesser of $P or $Z, if Capitalist is unwilling to sell his right to pollute, the Pareto optimal result, assuming the background property law, is for the factory to operate and Green's forest to die. This result remains optimal even if Green is also a true believer willing to do anything within her power to save the trees, so long as her power is not sufficient to make $G an attractive offer to Capitalist.

Coase, however, contends that in the absence of transaction costs this result is independent of the legal regime. This seems incorrect. Indeed, he "efficient" result only becomes determinate once the legal regime is specified. (See Leff; Kelman). That is, far from being irrelevant, the legal system does all the important work even in the imaginary friction-free world.

To see this, consider the alternative legal regime. This time, imagine that the property right is Green's. That is, Capitalist can only destroy Green's forest if he purchases the right to do so. If Capitalist doesn't buy Green's land, Green will be entitled to an injunction barring Capitalist's factory from emitting pollutants that will affect it.

Here, the analysis is different. First, there is no a priori reason to believe that Green will be willing to sell her forest's survival for the price she was willing and able to pay to preserve it. Green might be poor but committed, for instance, and thus unable to raise much money to save her forest but prepared to forgo great sums to avoid selling it. Thus not $G, the price Green was able and willing to pay to preserve her forest, but $S, the price she demands to sell its life, which is probably larger than $G, is the number Capitalist must consider.

The problem here is not one of transaction costs but opportunity costs. While profit-maximizing buyers may treat opportunity costs the same as real costs, sellers -- especially if the item being sold has value beyond its profit potential -- rarely will. The opportunity cost of a option forgone is a theoretical number. The cost to buy an opportunity is a real one. In a world of limited resources, limits apply to the latter that don't apply to the former. Torts lawyers know this intuitively: the price that someone would demand to sell their right eye is always more than the amount they'd be willing to spend out of their own resources to save it. Which question you ask will affect which answer you get. So should law professors: few law professors would be willing (or able) to pay to retain their jobs as much as the opportunities they forgo by leaving private practice.

If Green is a true believer, $S, her sale price, may be so large that it exceeds any possible $P, the profit from the factory. In this case, the economic results will be exactly opposite to those described above: the factory will not be built without pollution controls regardless of Capitalist's desires or the size of $P.5 A Capitalist tree-hater will be out of luck. As in the case discussed above, this result is also Pareto optimal -- no voluntary agreement can be reached to change it, given the background legal regime.

Second, unlike in the first legal regime, here the relative size of $Z and $P is critically important. In the first regime, the factory was built only if Green was willing to pay for the cheaper of the pollution controls or Capitalist's lost profit. Here, the payments go the other direction, but they are not symmetrical.

If the factory is profitable even with pollution controls (that is, $Z is less than $P), a rational, wealth-maximizing Capitalist will build the factory regardless of Green's offers or desires.6 The only question is whether it is cheaper to install the controls or buy off Green. Either he will buy the right to destroy Green's forest (if $Z is greater than $S) or he will install the pollution control device (if $Z is less than $S).

On the other hand, if the factory is profitable only without pollution controls (that is, $Z is greater than $P), then Capitalist will build it only if $S is smaller than both $P and $Z.

Coase argues that the two legal regimes result in the same allocation of resources, "neglecting wealth effects." This, however, is not true when opportunity costs are not the same as real costs or, what is almost the same thing, buy prices differ from sell prices. If $P > $S > $Z > $G or $Z > $S > $P > $G, the legal regime is the sole determinant of the Pareto optimal result: the factory will pollute and the forest will die if the property right belongs to Capitalist, whereas if the right belongs to Green, the factory will not pollute and the forest will live. See Appendix I.

The distinction between buy prices and sell prices is no mere theoretical anomaly. Rather, people are often unwilling to sell a right they already possess except for a sum far larger than they would be willing (or able) to pay to acquire the same right. The true believer, who will not sell at any price (S = infinity) but obviously cannot pay more than he or she has available (so that necessarily S > G) is merely a special case.7

* * *

Furthermore, in the long run, there is an even more important distinction between these two regimes: the cascading results of the "wealth effect" will have ever increasing effects over time.8 If Capitalist has a right to pollute, Green's attempts to preserve her forest result in Capitalist earning at least as much as he would have had his actions never impacted on Green (regardless of whether the factory ultimately pollutes or not). That is, $P is the floor of his earnings; Green must pay him more than $P to cause him to change his plans. In contrast, if Green has a right to clean air, Capitalist earns no more than if the factory were non-polluting. In other words, $P is his earnings ceiling; he must pay Green a part of $P in order to proceed.

In the next round, and the rounds after that, this difference in the direction of the side payments makes all the difference. If the right is Capitalist's, his economic power increases and Green's decreases with each succeeding transaction, regardless of whether the level of pollution does as well. If the right is Green's, the reverse is true. Thus, in the long run, either Capitalist or Green will end up with a vastly greater ability to determine the social choice between factories and forests. The first legal regime empowers polluters while the second one empowers environmentalists.9

If one also assumes, as seems reasonable, that Green will value clean air more as her wealth increases, so that both $G and $S increase along with Green's wealth, the cascade effect becomes even more pronounced. With each succeeding round the side that was given the benefit of the initial legal rule will become more and more likely to win. The two rules lead to dramatically different societies -- one much dirtier than the other -- with different enriched elites -- in one case, those who build factories, in the other case those who guard pristine nature. Critically, it is the law that determines what will be built and who will become rich, not entrepreneurial abilities, merit or the abstract preferences of the citizens taken individually.

* * *

As Guido Calabresi has emphasized, property regimes enforceable by injunction are not the only legal possibility. Liability principles allow a party to use another's property provided compensation is paid. Thus, a liability regime may be thought of as reversing the results in the true-believer situation. In a property regime, the property right-holder need never sell, regardless of the price offered. Therefore, the property right-holder can impose its will on others regardless of the intensity of their desire to buy. In a liability regime, in contrast, the buyer is privileged: by simply taking the property right and then paying the appropriate compensation, the buyer can always force the right-holder to sell.

The compensation to be paid under a liability rule is socially determined, and thus might seem to remove some of the sources of the differing results in differing property regimes.10 This, however, is not the case. Compensation and liability rules necessarily reflect the same distinction between buyers and sellers we have seen in property rules.

Thus, one possible liability rule is that Capitalist can pollute only if he compensates Green for the damage to her trees. This rule gives Green a partial property right in clean air: although she cannot guarantee the survival of her forest, at least she need not pay to preserve it and will be paid if she loses it.

Within this regime, there are several possible compensation rules: Green could be paid the cost of repairing the damage, she could be paid the price for which she, or an imaginary objective seller, would have sold her right to clean air, or she can be compensated the price Capitalist, or an imaginary objective Capitalist, would have been willing to pay to purchase the right (disgorgement). Our tort law uses all three principles and the frequent disagreement between judges and juries on the size of verdicts reflects, I suspect, an underlying lack of consensus as to which is most appropriate in which circumstances.

For our purposes, it suffices to see that the latter two principles are fundamentally similar to the two property regimes outlined above, except that they force the true believer to part with his or her property right at an objectively determined price. The two rules therefore will lead to the cascade effect discussed above. That, in turn, invalidates Coase's claim that the law can be neutral.

The first compensation rule is somewhat different. If Green is paid her damages in the sense of her cost to repair the forest ($D), yet a third set of equations will result, in which the relation of this new number, $D, to $P and $Z will determine whether the factory is built. If $D is smaller than both $P and $Z, the factory will be built without pollution controls and Green will be compensated. If $D is larger than $Z, then Green will not be compensated and the factory either will be built with controls or not at all, depending on whether $Z is larger than $P. This is yet another different allocation of social resources, although like the regime in which Green had property rights, it is generally favorable to the greens of the world against the capitalists.

Since $D is necessarily less than $S,11 the liability regime is less favorable to Green than the property one, and the cascade in favor of Green will be slowed. Green will, however, continue to have increased economic power with each succeeding round of transactions. Thus, even if this forest is destroyed, Green will be more able to create or preserve other forests with each succeeding round.

Finally, the liability regime could be set up to compensate Capitalist for Green's interference with Capitalist's right to use his property to pollute. Green, thus, would have a right of "reverse condemnation:" or modified eminent domain: she could prevent Capitalist from polluting but only by compensating him for his lost rights. This is the position that so-called "property rights" advocates have taken in resisting both zoning and environmental rules. It is a play on the action/inaction distinction that should be quite familiar to any student of Coase: one of Coase's central points is that interference can always be seen symmetrically. Problems of your fist in my face can always be recharacterized as my face in your fist. Here, rather than seeing Capitalist as polluting Green's forest or destroying her property, the property rights advocate sees Green officiously interfering with Capitalist's use of his property.12

This regime would be the inverse of the Green-liability rule discussed above. Capitalist's damages would be the lesser of $P or $Z: the cost to him of not polluting. The economic results would be yet a fourth variation. As in the regime where Capitalist had property rights to pollute, the factory will be built whenever both $P and $Z exceed $G; built with pollution controls where $G exceeds $P and $Z is less than $P; and will not be built where pollution controls make the factory unprofitable ($Z > $P) and $G is greater than $Z. However, in each case Capitalist would earn a total of $P, the expected profits from the polluting factory.

This regime, then, is similar to granting Capitalist a property right to pollute in that it favors Capitalist in both the first and subsequent rounds. However, it does not favor Capitalist to the same extent as the property regime -- under the property regime Capitalist earned more than $P as a result of Green's attempt to preserve her forest, whereas here he earns only the same $P he would earn in the absence of Green. Accordingly, the cascade in favor of Capitalist will be somewhat slowed. Nonetheless, Capitalist's wealth relative to Green's will increase with each round, and Green may eventually be unable to pay the damages necessary if she is to invoke her right to prevent pollution and maintain her forest.

* * *

Transaction costs may radically alter the above analysis. In particular, collective action problems are likely to accentuate the burden placed on the party without rights, particularly if Capitalist is the right-holder, and Green is one of many victims each of whom views ending the pollution as a public good. Thus, the model may be understating the cascade effect.13

But even this simple model is enough to make several critical points. First, property does not come from the state of nature. It is defined by legal rules. Merely invoking Capitalist's or Green's right to property will not resolve the dispute between them. They each have property and each must infringe on the other's proposed use in order to use his or her own property the way he or she wishes. Invocation of property rights merely restates the problem: defining who is the injured party here and determining who the law will empower relative to whom.

Second, economic efficiency won't do the trick either: all of the different results outlined above are economically efficient given the legal starting point. Until that starting point is given, the efficient outcome is indeterminate. More important in real world debates, if the legal background is simply assumed, the unrestrained pursuit of efficiency will merely enrich the rich, empower the powerful and entrench the status quo.

Third, the law matters. If we give rights to Capitalists, we get more pollution and more factories. If we give it to Green, we end up with more forests and more novels. This is the result of the difference between buy prices and sell prices or opportunity costs and real costs, and it is accentuated by the cascade effect of ever increasing wealth. If we give the initial right to Capitalist, he will have more resources and Green will have fewer at the beginning of the next round and Capitalist will, therefore, be more able to outbid Green. We thus end up with even more factories and pollution than we might have anticipated simply based on the difference between S and G. Indeed, Capitalist may accumulate sufficient wealth to succeed in changing the underlying rules to help promote his vision of the world even more -- say, by changing from a pro-Capitalist liability rule to a pro-Capitalist property rule, in order to assist the socially productive efforts of factory builders. If we give the initial right to Green, the inverse will be true (although perhaps to a lesser degree, if Green suffers from collective action problems and other transaction costs that Capitalist does not, or if Green starts off with fewer economic resources than Capitalist).

Finally, there is no neutral point. We cannot be "fair" between Capitalist and Green. Any rule will advantage one over the other. We cannot, as Coase suggests, pick the rule that in the presence of transaction costs will mimic the results that would occur in the absence of transaction costs, because the rule largely determines even those results. The only way to choose between the alternative liability rules is to come to a collective, political judgment about whether we want factories or forests, pollution or air, to empower Capitalists or Greens. Nominally neutral principles, even in a cost-free world, lead ineluctably to non-neutral results; moral and political judgment cannot be escaped.


Footnotes:

1. The classic article and several of Coase's more recent restatements of his thesis are reprinted in R. Coase, The Firm, The Market and The Law (1988) (hereafter cited as Coase, p. ).

2. The political (as opposed to analytical) power of the law and economics movement lies in its contention that the law can avoid judging between competing value systems by promoting efficiency. Thus, the neutrality sought by liberal theorists since Hobbes can be found in a free market approach. However, the Coase Theorem, as I reinterpret it, suggests that efficiency as a goal is either indeterminate or a hidden empowerment of the powerful. In either case, it loses its privileged status as a governmental goal that is compelled by reason, not politics, and must join the fray as an ideological stance in the interest of only certain people, institutions and value systems.

3. The no transaction costs assumption is a powerful one: we are simply assuming away all externalities, social action problems and imperfect information. However, there is no reason to believe that any of these facts of real life will correct the problems I show in the economists' world.

4. For simplicity's sake P should be thought of as the expected return from the factory less any opportunity costs for alternative investments. Thus, P > 0 means that the factory is the most profitable investment available to Capitalist, not merely that it will have a positive economic return. It follows that P means only that if Capitalist installs pollution controls, the factory will no longer be his most attractive available investment. It might still have a positive economic return.

5. Of course, if P > Z, Capitalist may decide to build the factory with pollution controls.

6. More precisely, Capitalist will proceed so long as the factory remains his best available opportunity. See footnote 4, supra.

7. Technically, the divergence of S from G results from the divergence of social price from individual preference (i.e., any given individual would often be willing to pay more or less than the current market price of an item) combined with limited individual resources. Because each person's own bundle of rights is in effect a partial monopoly, but rights to be acquired are not, individual preferences can be effected on the selling end but not the buying end.

8. Coase simply neglects wealth effects without any discussion of why they do not swamp his theorem even in the imaginary no-transaction costs world.

9. This is an example of the phenomenon known to chaos theorists as a "cascade". A small difference in initial conditions -- whether Green or Capitalist is granted the property right -- leads to increasingly large differences in results with each succeeding iteration.

10. At first glance, it might seem that socially determined compensation would not have the divergence between S and G that caused some of the divergence between the two property regimes described above. However, the liability rules may simply duplicate the divergence. Thus, the relatively low damages awarded by workmen's compensation relative to jury awards could result from the two systems asking different questions, paralleling S and G: Workmen's compensation asks, how much is society willing to pay to purchase the right to cause this injury, while the jury asks, how much would the injured party have required ex ante in order to agree to sell the right to remain uninjured.

11. Without transaction costs, if $D > $S, the parties would agree to settle by Green selling her rights for a sum greater than $S but less than $D. Courts approximate this result: if $D were greater than $S, Green would be required to mitigate her damages by selling the right to destroy the forest rather than repairing it.

12. If all the Greens got together and passed a zoning law, for instance, that merely generalized Green's right to clean air, it would be obvious that Green's forest was interfering with Capitalist's factory.

13. See, e.g., [Schattshneider]; M. Olson, The Logic of Collective Action: Public Goods and the Theory of Groups (1968).


Appendix I

Regime I refers to a legal system that grants Capitalist an unrestricted (but transferrable) property right to pollute, enforceable by injunction.

Regime II refers to a legal system that grants Green an unrestricted (but transferable) property right to prevent harmful invasions of her property, including by pollution, enforceable by injunction.

1.

$P > $Z > $S > $G
(highly profitable factory, less expensive pollution controls, minimal Green resistance)

Regime I

Regime II

built

built

no pollution controls

no pollution controls

no payoffs

Capitalist pays Green $y, where $Z > $y > $S

Capitalist earns $P

Capitalist earns $P-$y

 

2.

$P > $S > $Z > $G
(cheaper pollution controls)

Regime I

Regime II

built

built

no pollution controls

pollution controls

no payoffs

no payoffs

Capitalist earns $P

Capitalist earns $P-$Z

 

3.

$P > $S > $G > $Z
(cheapest pollution controls)

Regime I

Regime II

built

built

pollution controls

pollution controls

Green pays Capitalist $X where $G > $x > $Z

no payoffs

Capitalist earns $P+$x

Capitalist earns $P-$Z

 

4.

$Z > $P > $S > $G
(pollution controls would make factory unprofitable, but without controls, factory is more profitable than Green resistance)

Regime I

Regime II

built

built

no pollution controls

no pollution controls

no payoffs

Capitalist pays Green $q where $P > $q > $S

Capitalist earns $P

Capitalist earns $P-$q

 

5.

$Z > $S > $P > $G
(higher Green resistance)

Regime I

Regime II

built

not built

no pollution controls

 

no payoffs

no payoffs

Capitalist earns $P

Capitalist earns $0

 

6.

$Z > $S > $G > $P
(highest Green resistance)

Regime I

Regime II

not built

not built

 

 

Green pays $x where $G > $x > P

no payoffs

Capitalist earns $x > $P

Capitalist earns $0

 

Notes to Appendix I

- The payments are different in every case, and thus the starting points for round two will be different and a cascade effect is likely.

- Furthermore, Coase's specific claim that the legal rule will not change the allocation of resources, leaving aside wealth effects, is contradicted even in this first round by scenarios 2 and 5, due to the divergence between S and G.