The next few chapters are concerned with two very simple questions. The first is under what circumstances rational individuals, each correctly acting in his own interest, produce an inefficient outcome. To put it differently, under what circumstances does individual rationality fail to lead to group rationality—to the best possible outcome for the group, with "best" defined in terms of economic efficiency? The second question is what can be done about it—in particular, how can legal rules be designed to minimize such problems.
Start with the simplest case, a situation in which I bear all of the costs and benefits produced by my actions; whatever I do, nobody else is either worse or better off as a result. Being rational, I choose the action that maximizes the net benefit to me. Net benefit to everyone is net benefit to me plus net benefit to everyone else. Since my action maximizes the former and is irrelevant to the latter, it also maximizes the sum.
For a simple example, consider my decision of which computer game to play this evening. I own all of the alternative games already, so my decision has no effect on the revenue of the companies that produce them. My office is far enough from the bedrooms so that my decision has no effect on how well my wife or children can sleep. The only effect is on me and, being rational, I choose the game that I currently find most entertaining.
We can expand the range of examples by considering cases in which my action affects other people but the effects cancel. Suppose I decide to put my house up for sale. One result is that my neighbor ends up selling his house for five thousand dollars less than he otherwise would have because he had to cut his price in order to keep his customer from buying my house instead. My neighbor is worse off by five thousand dollars as a result of my action but the person who bought his house at the lower price is five thousand dollars better off, so the net effect on other people is zero. Such effects are called "pecuniary" or "transfer" externalities—effects on other people that result in a net transfer between them but not a net cost to them. Since I impose no net cost on others, when I take the action that maximizes net benefit to me I am also maximizing net benefit to everyone.
Situations in which individual rationality leads to group rationality are surprisingly common in market settings. In a competitive market, the price at which something sells equals both the cost of producing it (more precisely, the cost of producing one more unit of the good—marginal cost) and the value to the purchaser of consuming it (more precisely, the value of consuming one more unit of the good—marginal value). When you buy an apple or an hour of labor on a competitive market, what you pay just compensates the person who sells it; when you sell an apple or an hour of labor, what you receive just measures the value to the purchaser of what you sold him. Prices end up transmitting all of the costs and benefits associated with your actions back to you. Thus, when you make the decision that maximizes your benefit, you are also making the decision that maximizes net benefit. It is this fact that is at the heart of the proof that a perfectly competitive market produces an efficient outcome. For the details, see a good price theory text.
While such situations are common, they are by no means universal. When my steel mill produces a ton of steel, I pay my workers for the cost of their labor, I pay the mining firm for the cost of its ore, but I do not pay the people downwind for the sulfur dioxide I am putting into their air. In deciding how much steel to produce and how to produce it, I make the decision that maximizes net gain to me, that maximizes net gain summed over me, my suppliers and customers, and everyone upwind, but not the decision that maximizes net gain to everyone.
Since the cost to me of producing steel is less than the true cost, I end up selling the steel at below its real cost; since the amount people buy depends in part on price, other people end up buying too much steel. Total value summed over everyone concerned, including my downwind neighbors, would be larger if we produced and consumed a little less steel and substituted some other material instead.
The existence of such external costs results in my producing too much steel. It also results in my producing too little pollution control. There may be ways to reduce the amount of pollution my mill produces: a different production process, cleaner coal, taller smokestacks, filters. If I can eliminate two dollars worth of pollution damage at a cost of one dollar of pollution control, it is worth doing from the standpoint of efficiency—cost one dollar, benefit two. But the cost is paid by me, and the benefit goes to the people downwind, so it is not in my interest to do it.
The problem is not simply that pollution is bad. All costs are bad; that is why they are costs. We are willing to bear some costs because we get benefits in exchange; we are willing to work, even when we would rather play, because work produces useful goods. The problem with external costs such as pollution is that they get left out of the calculation of what things are or are not worth doing, with the result that we end up with not only efficient pollution, pollution whose prevention would cost more than it is worth, but inefficient pollution as well.
One solution is direct regulation: some government agency such as the EPA makes rules requiring steel mills to filter their smoke, or build high smokestacks, or in various other ways reduce their pollution. While this is an obvious solution, it has some serious problems.
The first is that the EPA may not be interested in maximizing efficiency. Steel producers, coal producers, and the makers of filters and scrubbers are also voters and potential campaign contributors. It is not obvious that the summed effect of their political activities will make it in the interest of the politicians who ultimately control the EPA to use their power to produce an efficient outcome. If (to take a real-world example), senators from states producing high-sulfur coal have sufficient political clout, the regulations will be rigged to encourage the use of scrubbers even when switching to low sulfur coal would be a more efficient way of controlling pollution.
The second problem is that, even if the EPA wants to maximize efficiency, it does not know how to do it. Figuring out what pollution control measures are or are not worth taking and how much steel ought to be produced after properly allowing for the external costs of producing it are hard problems. Answering them requires detailed information about the costs and benefits of different measures for different steel mills, most of which is in the hands of the firms, not the regulators. If the EPA simply asks a firm whether it knows any cost-effective way of controlling its pollution, the obvious answer is "no," since if the EPA believes that there is no cost-effective form of control, the firm will get to pollute without having to pay any control costs at all. The answer may even be true. There is, after all, very little reason to learn things when the knowledge will only make you worse off.
Because of such problems, economists interested in the problem of controlling negative externalities, of which pollution is one example, usually favor a less direct form of regulation. Instead of telling the firm what it must do, the regulator simply charges the firm for its pollution. If making a ton of steel produces twenty pounds of sulfur dioxide, which does four dollars worth of damage, the firm is billed for its sulfur dioxide output at twenty cents a pound.
This approach, labeled "effluent fees" in the context of pollution (more generally, "Pigouvian taxes," after A. C. Pigou, the economist who thought up the idea), has several advantages over direct regulation. To begin with, the regulator does not have to know anything about the costs of pollution control; he can safely leave that to the firm. If the firm can reduce its emissions at a cost of less than twenty cents per pound, it is in its interest to do so. If the firm protests that there is no way it can produce steel without pollution, the EPA politely accepts its word and sends it a bill for the pollution it produces.
A second advantage is that this approach generates not only the right amount of pollution control but the right amount of steel as well. When the firm produces steel, its costs now include both the cost of controlling pollution and the cost of any pollution it fails to control. So the price steel is sold at now represents the true cost of producing it. When steel is cheaper than concrete, buildings will be built of steel; when concrete is cheaper, they will be built of concrete.
Effluent fees do not solve all of the problems of controlling pollution, unfortunately. For one thing, they do not solve the problem of making it in the political interest of the regulators to do the right thing. They might deliberately set the fee too low in exchange for political contributions or future jobs for the regulatory officials, or deliberately set the fee too high to punish firms for making contributions to the wrong candidate.
Even if the regulators are trying to produce the efficient outcome, it may not be easy to measure the damage actually done by each additional pound of sulfur dioxide, or CO2, or whatever. But at least the informational problems are less than with direct regulation, since the regulators no longer need to know how pollution can be controlled or at what cost. And the political problems, although they still exist, should be reduced, since it is harder to provide special favors to your friends when decisions are made pollutant by pollutant instead of firm by firm.
So far I have been describing Pigouvian taxes in the context of a regulatory agency such as the EPA. But the same analysis can be used to explain large parts of tort law. Instead of having the EPA impose effluent taxes, we permit the people downwind to sue the steel mill for the damage its pollution is doing to their houses, laundry, and lungs. The steel mill has the choice of eliminating the pollution, paying damages, or reducing the pollution and paying damages on what is left.
The analysis can be applied to parking fines and speeding tickets as well. When I drive fast I am imposing a cost, in additional accident risk, on other drivers. The law forces me to take account of that cost in my actions by fining me when I am caught exceeding the speed limit.
There are, of course, differences among these examples. An effluent fee goes to the state. Tort damages go to the victim and his attorney. Fines go to the state. There are other important differences as well, many of which will be discussed in later chapters.
But the fundamental logic of all three cases is similar. Someone takes an action that imposes costs on others. It is in his interest to take the action as long it produces a net benefit for him, even if including the effects on the rest of us converts that to a loss. We solve the problem with legal rules that force the actor to bear the external cost himself, to internalize the externality. His net cost now equals net cost to everyone, so he takes the action if and only if it produces a net benefit. Individual rationality has been harnessed to produce group rationality. It is an elegant solution although far from a perfect one, as you will see in subsequent chapters.
I mentioned earlier a special sort of externality called a pecuniary externality, one that imposes no net cost, since the effects on other people cancel out. Unlike other externalities, a pecuniary externality does not lead to inefficiency, since the actor's private net costs are equal to total net costs, just as they would be if there were no externality at all. My example was the externality I imposed on my neighbor by putting my house on the market when he is trying to sell his.
The implication of the argument is that my neighbor ought not to be able to collect damages from me for the reduction in the price of his house. Competition should not be, and is not, a tort. That particular legal principle appears in the common law at least as early as 1410, when the owner of one school sued a competitor for taking students away from him—and lost.
A clever—and mischievous—reader might suggest extending the argument from competition to theft. When a pickpocket steals fifty dollars from me, he is fifty dollars better off and I am fifty dollars worse off. The net effect is only a transfer, so there is no reason for an economist, concerned only with efficiency, to object. Does it follow that, just as competition ought not to be a tort, theft ought not to be a crime?
It does not. A pecuniary externality occurs when an action by A results in a transfer from B to C. The logic of the situation is rather different when an action by A results in a transfer from B to A.
Picking pockets is not a costless activity; it involves time, training, and a variety of risks. Suppose the total cost to me of picking your pocket works out, on average, to twenty dollars. I gain fifty dollars at a cost of twenty, so it is worth it for me to pick your pocket; I am thirty dollars better off. But the net effect is to transfer fifty dollars from you to me at a cost of twenty, making us on net twenty dollars worse off.
To carry the argument a little further, consider what happens if there are lots of other people about as talented as I am at this particular activity. Since twenty dollars worth of effort yields fifty dollars worth of income, picking pockets is an attractive opportunity for them too. The number of pickpockets increases rapidly.
As the number increases, the profitability of that line of work declines. Someone who carelessly flashes a big roll of bills is instantly targeted by six pickpockets, leaving numbers two through six out of luck. As the risk of having your pocket picked rises, so does the incentive to take precautions. People start carrying money in their shoes, which are harder to pick than their pockets.
The process stops only when it is no longer profitable for any more people to become pickpockets. We end up with lots of pickpockets, most of whom have abandoned productive jobs in order to make a little more money picking pockets. Most of the money the victims lose goes to reimburse the criminals for their time and effort, making the net loss roughly equal to the amount stolen—a little less, since some especially talented pickpockets make more in that profession than they could in any other.
So far we have omitted the cost to the victims of their precautions—sore feet from walking on small change and sore eyes from trying to keep a close watch on everyone around them. If we include that, the net cost of theft increases; it may well be more, not less, than the total amount stolen.
The general term for this phenomenon is "rent seeking." It occurs when there is an opportunity for people to spend resources transferring wealth from others to themselves. As long as the gain is more than the cost, it is worth making the transfer—from the standpoint of the recipient. As more people compete to be recipients, the gain falls. In equilibrium the marginal recipient (the least talented pickpocket) just breaks even. The inframarginal recipients (especially talented pickpockets) make some gain, although less than what the victims lose (even talented pickpockets have some costs). The victims lose both the amount transferred and the cost to them of their defensive efforts.
Rent seeking occurs in a wide variety of contexts, many relevant to the subjects of this book. The term was coined to describe competition for government favors—in the original example import permits that permitted the holders to buy foreign currency at an artificially cheap price. Firms competed to get those valuable favors by public relations efforts, lobbying, campaign contributions, bribes. As long as a million-dollar favor can be obtained for substantially less than a million, some other firm is willing to bid a little higher, so winning firms end up, on average, paying about what the favors are worth. Anne Krueger, who originated the term "rent seeking," estimated that the governments of India and Turkey, two poor countries that at the time had systems of exchange controls and import permits, were burning up between five and ten percent of their GNP in such unproductive competition.
For a somewhat less clear example, consider litigation. Each side is spending money on lawyers, expert witnesses, and the like, in order to increase its chance of winning the case. The plaintiff is spending money to increase the chance of a transfer from the defendant to him, the defendant to decrease it. The transfer itself is neither a net gain nor a net loss; the expenditures on litigation are a net loss.
This is a less clear example of rent seeking because it is at least possible that the expenditures are producing something valuable: an increased probability of a correct verdict, resulting in better incentives elsewhere in the system. Controlling pollution by allowing suits against polluting factories by people living downwind works only if factories that pollute are more likely to lose suits than factories that don't, or, in other words, if the court has at least some tendency to reach the correct verdict. Expenditures by both sides on producing evidence and argument may make that more likely.
The CEO of a company gives an optimistic speech. Six months later, the newest product turns out to be a flop and the company's stock falls. An enterprising lawyer files a class action suit on behalf of everyone who bought stock in the company between the speech and the stock drop. His argument is that the speech, by omitting relevant facts that might have led to a more pessimistic conclusion, fraudulently induced people to buy the stock for more than it was worth and that the company should therefore make up their losses. Such "fraud on the market" suits occasionally succeed. Since the potential damages are enormous, even a small chance of success is enough to make it in the interest of some defendants to settle out of court, making such litigation a profitable enterprise for entrepreneurial attorneys.
There are a number of problems with the theory underlying such suits. CEOs are no more omniscient than anyone else, making it a considerable stretch to treat an optimistic statement that turns out to be wrong as actionable fraud. Customers are free to decide for themselves whose predictions to believe. An investor who takes optimistic speeches and press releases for gospel might be wiser to keep his money under his mattress instead.
One problem particularly relevant to this chapter has to do with the calculation of damages. Even if we concede that the speech was deliberately fraudulent, liability ought to depend on net damage done. Suppose I bought a share from you for a hundred dollars and it later fell to fifty. If the speech (and my purchase) had not been made the stock would still have gone down; the only difference is that you, rather than I, would have been holding it. I am fifty dollars worse off as a result of my believing the CEO, but you are fifty dollars better off.
Insofar as the CEO's optimism merely resulted in different people holding the stock when it fell, the externality is purely pecuniary and should give rise to no liability at all. We get a net externality only if I bought the stock from the CEO, making the effect of his speech a transfer from him to me and converting the situation from pecuniary externality to rent seeking.
Courts, in accepting the plaintiffs' theory of how damages should be calculated, have produced an economically inefficient verdict. One result is to penalize executives for making predictions that might turn out to be wrong, thus reducing the total amount of information available to investors. Another is to divert real resources from producing useful goods and services to producing litigation.
Defenders of that theory might—and sometimes do—argue that even if the damage payment is unrelated to the actual damage done, it provides injured stockholders an incentive to sue and thus deters fraudulent statements by corporate executives. The problem is that the higher the potential damage award, the higher the incentive to sue even on a weak case—and suits with large sums at stake produce high litigation costs. One might, on similar grounds, argue that since we want to deter illegal parking, anyone who finds an illegally parked vehicle should be allowed to claim it for his own. What we want, here as in many other cases, is not an incentive but the right incentive, a point we will return to in later chapters.
Question: I have just described a case not only of pecuniary externalities but of rent seeking as well—and not by the CEO. Explain.
The term "rent seeking" was originated by Anne Krueger in "The Political Economy of the Rent-Seeking Society, American Economic Review 64 (June 1974)." The idea, however, was presented earlier and more generally by Gordon Tullock in "The Welfare Consequences of Tariffs, Monopoly and Theft," Western Economic Journal, vol. 5 (June 1967), pp. 224-232.