The purpose of this chapter is to use some of the ideas developed in previous chapters to understand the behavior of political institutions. It contains three parts. The first is an analysis of the effects of tariffs--in particular, the question of whether imposing a tariff is a Marshall improvement or a Marshall worsening. The second is a sketch of two variants of public choice theory--the economic analysis of government--intended in part to explain the inconsistency between the sorts of tariffs that economists can defend as efficient and the sorts that exist. The third part uses the concept of rent seeking, introduced in Chapter 16 in a different context, to analyze the cost of government activity.
Chapter 5 introduced the principle of comparative advantage and showed why such standard arguments for tariffs as "The Japanese can produce everything cheaper than we can" or "Tariffs protect American jobs" are wrong. Showing that particular arguments for tariffs are wrong is not the same thing as showing that tariffs are undesirable. We saw why mutual gains from trade were possible, but we did not, at that point, have the tools necessary to determine whether those gains were maximized by free trade, or might be increased by appropriate tariffs or trade subsidies. Now we do. As you will shortly see, the answer to that question depends both on what we assume about international markets and on whose interests we take into account in judging one arrangement superior to another--only the interests of Americans or the interests of both Americans and the people we trade with.
In the first section of this part of the chapter, I will prove that if America as a whole is a price taker in international markets, then American tariffs are undesirable even if we take into account only the interests of Americans--or, in other words, that the abolition of American tariffs would produce net benefits for Americans. The result will be proved once graphically and once verbally, using a number of simplifying assumptions. In the next section, I will show several exceptions to the general rule that tariffs are undesirable; in each case, the exception depends on dropping one of the assumptions used in the proof. Some of the exceptions are cases where tariffs may be desirable if we consider only the interests of Americans but not if we include the effect on foreigners; others are cases where the imposition of a particular tariff is a Marshall improvement even when we count effects on everyone.
Having answered the question of what tariffs we should have, I will then, in the second part of the chapter, take up the question of what tariffs we do have and why.
I will start by listing the assumptions that will be used in the proof. We assume only one good is imported (autos) and one good is exported (wheat). We assume that America is a price taker in international markets: Changes in our production of wheat and consumption of autos are not sufficient to change the rate at which autos exchange for wheat abroad. The wheat and auto industries in the United States are price-taking industries with no substantial net externalities. Transport costs are zero.
The Geometric Proof. Figure 19-1 shows the supply curve for American production of automobiles and the demand curve for American consumption of automobiles, both before and after the imposition of a tariff. PA is the market price before the tariff, P'A after the tariff. QA is the quantity of imported cars before the tariff, Q'A after. Figure 19-2 shows the corresponding curves, prices, and quantities for wheat.
The first thing that should strike you about Figure 19-1 is that at neither PA nor P'A is quantity supplied equal to quantity demanded. This is because the price at which U.S. quantity supplied would equal U.S. quantity demanded is above the world market price; the United States therefore imports autos. Quantity demanded is equal to quantity supplied (by the U.S. auto industry) plus imports. Similarly, in Figure 19-2, the price at which quantity of wheat supplied and quantity demanded in the United States are equal is below the world price of wheat.The United States therefore exports wheat. Quantity produced (by U.S. farmers) equals quantity demanded (by U.S. consumers) plus exports.
The next question that might occur to you concerns Figure 19-2. The tariff is on autos; why should it affect the price of wheat? The answer is that wheat is what we are sending foreigners in exchange for the autos they are sending us. If, because of the tariff, fewer dollars go abroad to buy foreign cars, then foreigners will have fewer dollars with which to buy American wheat. Foreign demand for American wheat falls; the price of wheat in America drops. This effect is shown on Figure 19-2.
The effect on the domestic auto market of a tariff on autos. D and S are the domestic demand and supply curves for autos. QA is the rate at which autos are being imported before the tariff is imposed, Q'A the rate after. PA is the U.S. price of autos before, P'A the price after.
The effect on the domestic wheat market of a
tariff on autos. D and S are the domestic demand and supply
curves for wheat. QW is the rate at which wheat is being
exported before the tariff is imposed, Q'W the rate after.
PW is the U.S. price of wheat before, P'W
after.
The colored area U1 on Figure 19-1 is the increase in (American) producer surplus as a result of the tariff; U1 plus the shaded area R1 + S1 + T1 is the reduction in (American) consumer surplus. The shaded area is the net loss (to Americans) of surplus on autos as a result of the tariff. Similarly, on Figure 19-2, U2 is the gain in (American) consumer surplus as a result of the fall in the price of wheat produced by the tariff on automobiles, and U2 + R2 + S2 + T2 is the loss of (American) producer surplus. The shaded area R2 + S2 + T2 is the net loss (to Americans) of surplus on wheat as an indirect result of the tariff on autos.
There is one more term to be considered in calculating the net effect of the tariff on Americans: the money actually collected by the tariff. If the tariff is $t/auto, the government collects t dollars on each of Q'A autos imported each year, so its revenue from the tariff is t x Q'A. If that is larger than the sum of the two shaded areas, then the tariff produces net gains and is a Marshall improvement; if it is smaller, then the tariff is a Marshall worsening and its abolition would be a Marshall improvement. What I am going to show is that t x Q'A, the revenue collected by the tariff, is equal to the area S1 + S2. Since S1 + S2 is only part of the cost of the tariff, that will imply that the total cost is larger than the revenue collected, and hence that the tariff produces net costs rather than net benefits.
I will use two relations implicit in the situation I have described. The first is that since America is assumed to be a price taker in international markets, the tariff does not affect the relative prices of autos and wheat outside the United States Before the tariff, the price ratio is PA/PW. After the tariff, the price of wheat abroad (in dollars) is P'W, the price of autos abroad is P'A - t, so the price ratio is (P'A - t)/P'W.
How do I know that the world price of autos is P'A - t? P'A is the price of autos in the United States In order to get foreign autos into the United States, you must pay their world price plus the tariff t; the price in the United States is P'A, so the world price must be P'A - t.
Since the price ratio outside the United States is the same before and after the tariff, it follows that:
=
(Equation 1)
Autos are, by assumption, our only import and wheat our only export, so the total number of dollars foreigners get for the cars they sell to us must equal the number of dollars they spend for the wheat they buy from us. Using prices and quantities after the tariff is imposed, this gives us:
P'W x Q'W = $'s spent on wheat by foreigners = $'s received for cars by foreigners = (P'A - t)Q'A. (Equation 2)
(We spend P'A on each car, but since t goes to the government to pay the tariff, only P'A - t goes to foreigners).
Finally, from Figures 19-1 and 19-2, we have:
S1 + S2 = (P'A - PA)Q'A + (PW - P'W)Q'W. (Equation 3)
Equations 1 and 2 imply that:
Substituting this into Equation 3 gives us:
S1 + S2 =
Q'A(P'A - PA) + Q'A
(PW - P'W) =
Q'A{ P'A - PA +
(PW - P'W)} = Q'A{ P'A -
PA + PA - P'W 
}
.
If we cancel out - PA + PA
and use Equation 1 to replace 
with 
,
we get:
S1 + S2 = Q'A{
P'A - P'W 
}
=Q'A(P'A - P'A + t) = Q'A
x t (Equation 4)
Since S1 + S2 is only part of the net loss to American consumers and producers caused by the tariff and Q'A x t is all of the revenue collected by the tariff, net loss is larger than revenue; the imposition of the tariff is a Marshall worsening and its abolition would be a Marshall improvement. This assumes that the triangles R1, R2, T1, and T2 are not all equal to zero. If they were--if, for instance, all of the supply curves were perfectly inelastic, which seems highly implausible--then the tariff would produce net benefits of zero. The tariff cannot make things better and is almost certain to make them worse.
The Verbal Proof. I have now proved my result--that if the United States is a price taker in international markets and American firms are price takers in domestic markets, American tariffs on net injure (or at best do not benefit) Americans--mathematically. Next I will prove it again in another language: English.
From the standpoint of the United States, foreign trade is a technology for turning wheat into autos at the rate PA/PW. We proved in Chapter 16 that a competitive industry (without externalities) is efficient. Hence the result of the competitive industry for turning wheat into autos is efficient. A tariff alters that result, taxing the conversion of wheat into autos and so reducing the quantity of wheat used and autos produced. That change could be made by a bureaucrat-god. A bureaucrat-god cannot improve an outcome that is already efficient--that is the definition of "efficient." So a tariff cannot be a Marshall improvement. Since it alters a situation that is already efficient, it is almost certain to be a Marshall worsening.
Capital in Action. There are two things I would like you to notice about what we have done so far in this chapter. The first is that our proofs are themselves examples of the use of capital in production. We have spent the previous 18 chapters accumulating intellectual capital, learning a complicated set of ideas, many of which must, at times, have seemed entirely useless. Using that capital, we have now, with a few pages of high school geometry and algebra plus a paragraph of reasoning, proved one of the more important practical results of economic theory--twice.
The second thing I would like you to notice is the contrast between the two proofs, a contrast typical of the differences between the two languages we used. The advantage of the verbal proof is that it helps us to intuit why tariffs are undesirable--provided we have previously learned to intuit why a competitive industry is efficient. Trade is simply a technology for converting exports into imports; a competitive industry uses that technology up to the point where the benefit of one more unit of imports is balanced by the cost of producing the exports that must be exchanged for it. A tariff adds an additional cost of production; the industry reduces its output, depriving some consumers of imported goods that they valued at more than their cost but less than their cost plus the tariff. The tariff is simply a tax on a particular way of producing things; the net loss is the resulting excess burden, just as with any other tax.
Does this conclusion depend on assuming that the United States is a price taker in international markets? For the mathematical proof, the answer is yes; that is how we got Equation 1, which was used twice in the proof. In the verbal proof, however, I said nothing at all about whether the United States as a whole was a price taker; all I assumed was that the export and import industries were price takers within the United States (and therefore efficient). That is not at all the same thing. If U.S. agriculture consists of a million small farms, each farmer is a price taker; but if the United States produces 90 percent of the world's wheat, the United States as a whole is not-- changes in how much wheat we export will affect world prices.
In fact, the verbal proof does depend on the United States being a price taker, but the reason is a somewhat subtle one. If the United States is not a price taker, then the quantity of wheat exported (and autos imported) affects the price ratio abroad, which means that it affects the rate at which we can convert wheat into autos. From the standpoint of the United States, that is an externality; when I buy autos abroad, I drive up their price (and drive down the price of the wheat I use to pay for them), making it more expensive for you to buy autos abroad. If we think of trade as a way of converting wheat into autos, this is like a situation where my increased production of widgets somehow makes your widget factory less productive--which would be an externality. We know from Chapter 18 that a competitive industry with externalities does not generally produce an efficient outcome. So if the U. S. is a price searcher, the initial situation (without a tariff) is not efficient, and it is possible that a tariff may improve it.
From the standpoint of the world as a whole, the externality in question is a pecuniary externality, so it may be ignored; if my purchases of automobiles drive up the world price, that is a loss to the other buyers but a gain to the sellers. But if the buyers are Americans, the sellers are foreigners, and we consider only the interests of Americans, there is a net externality--we count the loss and ignore the gain. So if the United States is a price searcher in international markets, the outcome without tariffs is efficient if all interests are considered but inefficient if only American interests are.
Three important assumptions went into the proof that tariffs were undesirable: that the United States was a price taker in international markets, that American import and export industries were price takers within the United States, and that they had no important externalities. We will discuss the results of dropping two of them.
America as a Monopolist. Suppose the United States is not a price taker in international markets; suppose, for example, that we have a monopoly on wheat. Individual farmers are still price takers, so they produce up to the point where MC = P. All of the farmers taken together, however, would do better if they acted like a monopoly or a cartel, restricting their production and driving up price to the point where MC = MR. The government can produce that result by imposing an export tax--a sort of backwards tariff--on wheat. The export tax drives up the price of wheat abroad; Americans as a whole (including the government collecting the tax) gain just as they would if the farmers had gotten together and raised their price. The same result can also be produced by a tax on imports--an ordinary tariff. Importing and exporting are two sides of the same transaction--trading wheat for cars--so it does not matter at which point you impose the tax. It follows that if we are price searchers in international markets, a tariff may produce net benefits for us.
The same argument applies if we are price searchers as consumers of automobiles: monopsonists. In that case, a tariff on automobiles drives up their price in the United States, lowers U.S. consumption, and so drives down the price of automobiles abroad. Since the United States is a net importer, we benefit by the lower price.
What happens in both of these cases is that without a tariff, individual Americans function as price takers in international markets even though America as a whole has some monopoly power. The tariff, in effect, creates a monopoly (or monopsony) out of a multitude of small firms. The result is a net gain at the expense of our trading partners. When we drive the international price of autos down (by imposing a tariff that decreases our consumption) or the international price of wheat up (by imposing an export tax on wheat), we benefit, since we are sellers of wheat and buyers of autos. Our trading partners lose, since they are buyers of wheat and sellers of autos. Just as in the (single-price) monopolies discussed earlier, the result is a net loss but a gain for the monopolist. Such a tariff is a Marshall improvement if we consider only gains and costs to Americans; it is a Marshall worsening if we include gains and losses to foreigners. Just as with other cases of monopoly pricing, demand and supply curves are likely to be more elastic in the long run than in the short run, so our gains from the tariff may be only temporary. One other problem with such a tariff is that we may not be the only country with a monopoly or a monopsony. If we use a tariff to exploit our monopoly power against our trading partners, they may do the same thing to us.
Protecting Infant Industries. A tariff designed to create monopoly profits for the nation that imposes it is one example of a tariff that may be efficient, provided that we ignore the effect on foreigners. A second example, and one often used by supporters of tariffs, is a tariff to protect an infant industry. This can, under some assumptions, result in an improvement even if we include the effects on foreigners.
Assume the United States has the potential to produce tin but does not yet have a tin industry. A company that tries to start a tin foundry in the United States will have a hard time of it--American workers do not know how to work with tin, American railroads have no experience shipping tin and no special freight cars designed to carry it, and American coal mines have no experience producing the particular kinds of coal needed to refine tin from tin ore. Until all those problems are solved, American tin will be more costly than imported tin. If only the tin industry could get established, it would be profitable, but nobody wants to be first.
One problem with the argument as stated is that if the tin industry is going to be profitable in the long run, tin companies should be willing to accept losses in the first few years, treating them as an investment to be paid back out of later profits. If companies are not willing to do that, perhaps the profits are not large enough, or certain enough, to make the losses worth taking.
To get around this argument, one must assume that the process of development occurs within the industry but outside the firm. No firm can do it by itself, but if they all do it together, workers will become skilled in working tin, subsidiary industries will grow up to support tin manufacture, and so on. Another way of putting this is that there are large positive externalities produced by the first few firms in an industry; while they are losing money producing tin, they are also producing a body of skills and knowledge in their employees and suppliers that will lower the costs of future producers.
If this is true, then since the initial firms do not include the (external) benefits they produce for others in calculating the value of what they produce, they may never start production unless they are subsidized by a temporary tariff that raises the cost of imported tin. This is the argument for an infant industry tariff. We have dropped the assumption that the firms in the industry have no important externalities; the result is that a tariff may be desirable--imposing it may be a Marshall improvement. In this case, unlike the previous one, a tariff may be desirable even if we take into account the interests of everyone concerned. If the United States has the potential to produce tin less expensively than it can be imported, the gains to the U.S. producers and their customers may ultimately outweigh the losses to foreign producers.
Should vs Will. So far, I have shown that tariffs are usually undesirable but that there are exceptions--situations where tariffs are desirable, at least from the standpoint of the countries that impose them. Is that why the United States, and most other countries, have tariffs? Apparently not. The tariffs we observe in the real world have little resemblance to the ones that can be defended as economically desirable. It is not infant industries that get protection but senile industries--American auto, shoe, and steel producers, for example. Why?
To answer that question, we need to understand not what laws should exist but what laws will exist. We need, in other words, an economic theory of politics. The branch of economics that deals with such questions is called public choice theory, presumably because it explains public choices, while ordinary economics deals with private choices. The name is somewhat deceptive, since what makes public choice theory part of economics is that it analyzes the behavior of political institutions as the outcome of the choices of rational individuals, each seeking his own objectives. It is a theory of the behavior of political institutions that results from private choices on the political marketplace.
In the next section, I will sketch one version of public choice theory; after doing so, we will see how that theory can be used to explain the observed pattern of tariffs. One of the questions we will be interested in is whether the political system can be expected, like the competitive market of Chapter 16, to generate efficient results. If so, we would expect the tariffs actually observed in the real world to correspond fairly closely to the "good" tariffs discussed above.
If, on the other hand, the political market--like the private market with monopoly, externalities, or public goods--produces inefficient outcomes, then there is little reason to expect the tariffs that economists observe to correspond to those they recommend. The problem is then to predict the outcomes of the political market--to show which particular industries will or will not get tariff protection. Having done so, we can compare the predictions of the theory with what we actually observe, in order to test the theory and perhaps, if its predictions turn out to be correct, understand the reasons for the outcomes we observe.
Public choice theory is simply economics applied to a market with peculiar property rights. Just as in the economic analysis of an ordinary market, individuals are assumed to pursue their separate objectives rationally; just as in that analysis, one may first make and later drop simplifying assumptions such as perfect information or zero transaction costs. The property rights on the public market, however, are different from those on the private market; they include the right of individuals to vote for representatives, of representatives to make laws, of government officials to enforce the laws, of judges to interpret them, and so on.
Ordinary economics is greatly simplified if we treat firms as imaginary individuals trying to maximize their profits; in this way, we convert GM from several hundred thousand individuals into one. There is some cost to the simplification, since it ignores the conflicts of interest within the firm among managers, employees, and stockholders. So far, no alternative simplification seems to work as well. So economists continue to analyze an economy of profit-maximizing firms, except when the particular problem being considered hinges on intrafirm interactions--as it does, for instance, in the theory of the firm, of which the optional section of Chapter 9 and the discussion of mergers in Chapter 18 provide brief samples.
One of the ways in which different public choice theories differ from each other is in what they take to be the equivalent of the firm on the political market and what it is assumed to maximize. For the moment, I shall consider the entrepreneurs on the political market to be elected politicians and limit my discussion to the market for legislation. This is a simplification of the political market and only one of several possible simplifications--two others will be discussed in later sections--but it provides a convenient way of sketching the theory.
Consider, then, the market for legislation. Individuals perceive that they will be benefited or harmed by various laws. They offer payments to politicians for supporting some laws and opposing others. The payments may take the form of promises to vote for the politician, of cash payments to be used to finance future election campaigns, or of (concealed) contributions to the politician's income. The politician is seeking to maximize his long-run income (plus nonpecuniary benefits, one of which may be "national welfare"), subject to the constraint that he can only sell legislation for as long as he can keep getting elected.
Is It Efficient? To see whether we can expect the outcome of this market to be efficient, let us consider a simple example. A legislator proposes a bill that inefficiently transfers income from one interest group to another; it imposes costs of $10 each on a thousand individuals (total cost $10,000) and grants benefits of $500 each to ten individuals (total benefit $5,000). What will be "bid" for and against the law?
The total cost to the losers is $10,000, but the maximum amount they will be willing to offer to a politician to oppose the law is very much less than that. Why? Because of the public-good problem. Any individual who contributes to a campaign fund to defeat the bill is providing a public good for all thousand members of the group. The same arguments used in Chapter 18 to show that public goods are underproduced apply here. The larger the public, the lower the fraction of the value of the good that can be raised to pay for it.
The benefit provided to the winners is also a public good, but it goes to a much smaller public--ten individuals instead of a thousand. A smaller public can more easily organize, perhaps through conditional contracts ("I will contribute if and only if you do"), to fund a public good. Even though the benefit to the small group is smaller than the cost to the large one, the amount the small group is able to offer politicians to support the bill will be more than the amount the large group will offer to oppose it.
The effect is reinforced by a second consideration--information costs. Assume that information about the effect of legislation on any individual can be obtained, but only at some cost in time and money. For the individual who suspects that the bill may injure him by $10, it is not worth obtaining the information unless it is very inexpensive. His possible loss is small and so is the effect of any actions he is likely to take on the probability that the bill will pass. The member of the dispersed interest chooses (rationally) to be worse informed than the member of the concentrated interest. This is rational ignorance; it is rational to be ignorant if the cost of information is greater than its value.
What Does Concentration Mean? So far, I have discussed only one characteristic of a group--its size. It is useful to think of the terms "concentrated" and "dispersed" as useful shorthand for the whole set of characteristics that determine how easily a group can fund a public good; the number of individuals in the group is only one of those characteristics.
Consider, for example, a tariff on automobiles. It benefits hundreds of thousands of people--stockholders in auto companies, auto workers, property owners in Detroit, and so forth. But GM, Ford, Chrysler, American Motors, and the UAW are organizations that already exist to serve the interests of large parts of that large group of people. For many purposes, one can consider all of the stockholders and most of the workers as "being" five individuals--a group small enough to organize effectively. The beneficiaries of auto tariffs are a much more concentrated interest than a mere count of their numbers would suggest. That may explain why such tariffs exist, even though the costs they impose on consumers of automobiles and American producers of export goods, both dispersed interests, are larger than the benefits to the producers of automobiles.
The reason the public-good problem leads to inefficiency in ordinary private markets is that the amount a group can raise to pay for a public good benefiting that group is less than the total value of the good to the members of the group; hence some public goods that are worth more than it would cost to produce them fail to get produced, which is inefficient. The reason it leads to inefficiency in public markets is that both costs and benefits are only fractionally represented on the market, because of the public-good problem. If potential gainers and losers from proposed legislation raise different fractions of their gains and losses to bid for and against the laws, as will usually be the case, laws that impose net costs may be passed and laws that impose net benefits may not be. This again is inefficient.
Predictions. What predictions can we make on the basis of this simple model of individuals and interest groups bidding for legislation? One is that legislation will tend to benefit concentrated interest groups at the expense of dispersed interest groups--where "concentrated" and "dispersed" describe the bundle of characteristics that determine how large a fraction of the benefit that the members of the interest group would receive from legislation can be raised by the group to buy the legislation.
A second prediction is that although the system may frequently generate inefficient outcomes, nonetheless more efficient outcomes will be preferred to less efficient, all other things being equal. Consider, for instance, a politician choosing among different ways of subsidizing a particular concentrated interest at the expense of a particular dispersed interest. One scheme will provide the beneficiaries with $1 million and cost the victims $10 million; an alternative will provide $1 million and cost $5 million. The amount spent to oppose him will be less in the second case than in the first, so he prefers it; he is choosing a transfer with an overhead of 80 percent ($0.20 return for every dollar of cost) over one with an overhead of 90 percent. The same argument applies if both schemes cost the victims the same amount but one provides $1 million to the beneficiaries and one provides $2 million. The larger the benefit, the larger the amount he can get paid, in one form or another, by the beneficiaries.
So far, we have two predictions. Transfers go from dispersed interests to concentrated interests, and they are made as efficiently as possible, all other things being equal. This raises an obvious problem: Why do we ever observe inefficient transfers, such as tariffs? Why do not politicians always prefer to simply tax the proposed victims and turn the receipts over to the proposed beneficiaries, thus reducing transfer costs to the unavoidable minimum: the administrative cost of collecting the tax and paying out the benefits, and the associated excess burden?
One answer is that there is a third prediction implicit in our model. Politicians will prefer transfers for which the information cost of figuring out what is really happening is as high as possible for the victims and as low as possible for the beneficiaries; if the cost is the same for both victims and beneficiaries, high information costs will generally be preferred to low ones.
The first half of this is obvious: The harder it is for the victims to know they are victims, the less they will spend trying to prevent the legislation; and the easier it is for the beneficiaries to know they are beneficiaries, the more they will spend supporting it. The second half, the general preference for high information costs, follows from the fact that beneficiaries are generally more concentrated than victims. As I pointed out earlier, it is easier for a concentrated interest to overcome the problems associated with information costs. So if information costs are high, it is likely that the beneficiaries will still pay them--and support the legislation--while the victims will fail to pay them and so fail to oppose it.
The preference for high information costs helps to explain the existence of inefficient forms of transfer. Given the choice, the sponsors of legislation designed to benefit some people at the expense of others would prefer to disguise it as something else. A bill to tax consumers and give the money to GM, Ford, Chrysler, American Motors, and the UAW is likely to encounter more opposition than an auto tariff designed to do the same thing--because the auto tariff can be (and is) defended as a way of "protecting American jobs from the Japanese."
We now have three predictions about the outcome of political markets: They favor concentrated interests, they prefer more efficient to less efficient transfers, and they prefer transfers disguised as something else. How do these fit what we observe?
Tariffs in the Real World. One common observation about real-world tariffs is that they tend to go, not to infant industries, but to senile ones. In part, this is what we would expect from our discussion of concentrated versus dispersed interests. The American steel industry is a powerful concentrated interest; potential infant industries that do not now exist but could be created by an appropriate tariff are not. So it is the old industries that get the protection.
This explains why infant industries do not get tariffs, but it does not explain which industries do get them. If tariffs tend to go to declining industries, a satisfactory theory should explain why. The discussion of sunk costs in Chapter 13, combined with the prediction that politicians will prefer transfers that give the highest possible ratio of benefit to cost, all other things being equal, can do so.
Suppose a tariff is imposed on imports that compete with a growing, competitive, domestic industry. Before the tariff, price was equal to average cost, so economic profit was zero. The tariff reduces the supply of imports, so prices and the industry's output rise. But once enough new firms have entered the industry to reestablish equilibrium, average cost is again equal to price--profit is again zero. There is no gain to the industry, hence no reason for it to reward the politicians who imposed the tariff, save during the adjustment period.
If some of the inputs used by the industry are in fixed supply, such as certain types of land, their value will be bid up; their owners may be willing to offer part of the increase to get the tariff passed and maintained. If the inputs instead have a highly elastic supply curve, or if their ownership is divided among many individuals, no one of whom finds it worth his while to work for a tariff, then only transitional profits are available to reward the supporters of the tariff.
Consider next the case of a tariff on a declining industry. In such an industry there is usually an important resource in fixed supply: factories that produce enough revenue to be worth keeping but not enough to be worth building. The ownership of that resource is as concentrated as the industry is. The tariff increases demand for domestically produced goods by raising the cost of the competing imported goods and so increases the present value of the factories. In this case, unlike that of a growing industry, a large part of what the consumers lose in higher prices the producers receive in increased wealth.
The cost of the tariff is still larger than the benefits--but the cost is spread among many consumers and the benefits are concentrated on a few producers. Since the benefits to the industry are much larger in the case of a declining industry than in the case of a growing one, declining industries will be willing to work much harder to get tariffs. It is not surprising that they are generally more successful. The result is a pattern of tariffs almost exactly opposite to the pattern that could be justified as efficient.
The same analysis explains why tariffs on agricultural products are common--not so much in the United States, which is a net exporter of farm products, as in Japan and the countries of the European Economic Community, which are net importers. In the case of a tariff on farm products, the relevant fixed resource is land; increased demand for domestic crops drives up its price. Just as in the case of a declining industry, the producers get a large fraction, although not all, of what the consumers lose. If the fraction is large enough, and if the producers are sufficiently concentrated and well organized in comparison to the consumers, the result may be a tariff.
So far, my discussion of public choice theory has focused on the market for legislation in a way that appears to downplay the arrangements by which we are taught, in high school civics classes, our society is run: democratic elections. My reasons for doing so are in large part implicit in the discussion. Information costs make it difficult for voters to know which politicians are really acting in their interest, and the public-good problem means that it is rarely in the interest of voters to pay the costs and buy the information necessary to recognize and support "good" politicians.
This is one approach to public choice theory, but not the only one. There are other approaches, some of which choose to ignore such problems and analyze a democratic government in terms of the outcome of a system of majority voting among voters who correctly perceive their own interests and the positions of the candidates. The following is an example, applied to a recent election.
Hotelling and Hayden. You are planning to build a store on the block shown in Figure 19-3. After you build your store, your competitor will build his. The customers are evenly distributed along the block; each customer always goes to the nearer store. Where do you build?
One wrong answer is shown in Figure 19-3a; your store is A, and your competitor's store is B. By locating his store as shown, he gets all of the customers to his right. He maximizes the number of customers he gets by building next to you, on the side toward the center of the block. Your correct strategy is to build in the center, as shown on Figure 19-3c, forcing him to build on one side or the other. You each get about half the market.
The Hotelling Theorem applied to stores and to
politicians.
A few years ago, while teaching one of the courses out of which this textbook developed, I was reminded of this simple but elegant argument (originated by Harold Hotelling, the economist whose analysis of depletable resources was discussed in Chapter 12) by the campaign advertisements of Tom Hayden and his Republican opponent. Hayden was running a very left-wing campaign for the California legislature from a district including the city of Santa Monica--sometimes referred to by those unhappy with its politics as the People's Republic of Santa Monica. His opponent appeared to be taking very left-wing stands for a Republican: support of rent control, opposition to offshore drilling, and the like. The obvious interpretation is shown in Figure 19-3b, which is merely 19-3a relabeled. With Hayden far to the left on the political spectrum, his opponent maximizes his votes by being almost equally far left. Voters to his right have nowhere to go.
The same analysis explains the tendency of the American political system to nominate two similar candidates, both near the political center. That corresponds quite nicely to the prediction shown on Figure 19-3c. The fact that presidential candidates are not always at the center, like the fact that Hayden ran (and won) on a noncentrist platform, may be a result of additional complexities in the system. For one thing, issues cannot be perfectly represented as a one-dimensional left-right spectrum. For another, political support is not limited to voting; a voter near one end of the spectrum may be willing to vote for a candidate he perceives as only a little less bad than his opponent, but be reluctant to donate time or money to his campaign. Finally, political parties are not infinitely flexible, at least in the short run; the Republican party of California may have been unable to field a candidate who could convince the voters that he was almost as far left as Tom Hayden.
Hotelling's argument, as applied to politics, is called the median voter model. The idea is that on an issue such as the amount of government spending, for which positions line up in a one-dimensional pattern, the outcome will always represent the desires of the median voter. Any position on one side of that favored by the median voter (such as a proposal to spend slightly more money than he wishes spent) will be defeated by a coalition consisting of the median voter and everyone on the other side of him (those who want to spend less than he does)--just over 50 percent of the votes. So will a position on the other side of him. So the median voter gets exactly what he wants.
Here again, the conclusion depends on issues having an unambiguous one-dimensional ordering and on the voters correctly perceiving their own interests and voting accordingly. In a more realistic model, the results become much more ambiguous. If, for example, the issue is income redistribution, the voter with the median income might be defeated by a coalition of the two extremes--rich and poor voting to tax the middle class for their joint benefit. That coalition could in turn be defeated by a coalition between the middle and the poor (or the rich), that by still another coalition, and so on without end. This is the same endless cycle that we saw in Chapter 11 when we analyzed the game of three-person majority vote.
The Revenue-Maximizing Bureau. As a final example of the diversity of public choice theories, consider William Niskanen's theory of bureaucracy. In his analysis, the essential actor is the government bureau: the U.S. Department of Defense, the Water and Sewer Department of the City of New Orleans, or any other organized body of bureaucrats with a common interest. The objective of each bureau is to maximize its budget in order to maximize the power and income of its members.
Since bureaus usually cannot impose their own taxes, they must get money from a legislature--which can. They do so by offering the legislature a selection of price/output packages, each consisting of a certain amount of output (defense, water and sewers, schooling, or whatever) to be produced with a budget of a specified size. Since, in Niskanen's model, bureaus know their own cost functions but legislatures do not, the bureau can and does lie to the legislature about the cost of alternative levels of output. Its objective is to get the largest possible appropriation. To do that, it understates the amount it could produce with lower budgets, in order to make those packages less attractive in comparison to the (high-budget) package it is trying to sell the legislature. The strategy of the revenue-maximizing bureau turns out to be very similar to the strategy of the monopolist with perfect price discrimination. Each is trying to offer its customer a package that he will just barely accept, in order to transfer as much of the resulting gain as possible to itself.
This model too has its problems. For one thing, it is not clear that bureaucrats would want to maximize their budgets even if they could do so. If in order to get the maximum budget, the bureau must promise the legislature--and produce--a large output, it might be better off with a slightly smaller budget and a much smaller set of obligations. If less has to be produced, more money will be available to be spent on the salaries (and other perquisites) of the bureaucrats.
In Chapter 16, I introduced the idea of rent seeking in order to explain why perfect discriminatory monopoly might, under some circumstances, be the worst rather than the best way of organizing an industry. The term "rent seeking" was actually introduced to economics in a context more appropriate to this chapter. The analysis goes as follows.
A government has valuable favors to give out--import permits, licenses, or the like. They take the form of pieces of paper giving the recipient permission to do something. Each piece of paper is worth a million dollars; a potential recipient would, if necessary, pay up to that amount to get it. A thousand such pieces of paper are to be given out.
If you are giving away something worth a million dollars, there will be no shortage of claimants. Some way must be found to choose among them. Suppose, to begin with, that the permits are supposed to be given out to those firms that will use them "in the public interest." The society is a democratic one; government officials try to give the permits to the firms that the voting public prefers.
Firms can and do act to influence the public's perceptions. If your firm wants a permit and does not expect to get it, it may be worth spending some money on improving your public image--perhaps by advertisements telling the general public how important your product is to the national welfare, how many jobs depend on you, and how crucial it is that you get the permit.
How much will you be willing to spend on such advertising? If it makes the difference between getting and not getting the permit, anything up to the value of the permit--$1,000,000. Initially, perhaps, you can get away with less. But when other firms observe that your $100,000 ad campaign is going to result in your getting one of the permits and their not getting one, they start their own ad campaigns--budgeted at $200,000. You reevaluate the situation and increase your budget. They do the same.
As long as, on average, an expenditure of less than $1 million on advertising gets a government favor worth $1 million there will always be more firms willing to enter the game. By doing so, they either raise the amount that must be spent or lower the probability of success. Equilibrium is reached when each firm, on average, spends as much to get the permit as the permit is worth. If the firms that spend more are certain to get the permits, the result is that 1,000 firms spend $1 million each. If the situation is more uncertain, there may be 2,000 firms spending $500,000 each and each ending up with a 50 percent chance of success.
From one standpoint, the result is unsurprising; in equilibrium, marginal cost (as usual) equals marginal value. From another, it is very surprising indeed. The government is giving out, for free, a billion dollars worth of special favors, and the recipients are ending up with nothing--the full value of the favors is used up in getting them. I sometimes describe this conclusion as Friedman's Second Law: "The government cannot give anything away."
I have assumed that potential recipients of the permits can influence their chances by using advertising to improve their public image. The result does not depend on that particular assumption. One can imagine a variety of other ways in which potential recipients might influence the result, including political donations to the party in power and bribery of the officials allocating the permits. The logic of the situation remains the same--the firms spend as much getting the favors as the favors are worth.
The term "rent seeking" was introduced to economics in an article by Anne Krueger. The principal example she considered involved countries that maintain an official exchange rate for their currency higher than the market rate and use import permits to ration the resulting shortage of foreign currency. If a native of (say) India earns dollars by selling goods abroad, he is supposed to turn them over to the government at the official rate--which means that he gets fewer rupees for his dollars than they are worth on the market. Indians who are given import permits by the government can then buy those dollars with rupees at the official rate, and use the dollars to buy goods abroad and import them into the country. Such permits are worth a great deal of money. Krueger concluded that a conservative estimate of the market value of the permits and other favors given out by the governments of Turkey and India, and hence the amount wasted on rent-seeking activity, was about 7 percent of national income for India and 15 percent for Turkey.
Rent seeking is not limited to poor countries with exchange controls. In analyzing the market for legislation, I concentrated on the outcome of that market without discussing its costs. If special interests buy legislation from politicians, that increases the value of being a successful politician, which in turn increases the amount spent on getting and keeping political office. This brings us to an interesting puzzle.
It is common, especially around election time, to read articles lamenting how much is spent on campaigning. What surprises me is how little is spent on political campaigns, considering the stakes. In a presidential year, total expenditure by both parties on the presidential race and all congressional races is on the order of a few hundred million dollars. The prize is control of the federal government for several years, during which that government will spend several trillion dollars--or about ten thousand times total campaign expenditures.
One explanation for the disproportion between the prize and what is spent to get it is the public-good problem faced by even a relatively concentrated interest group. If a group can only raise, for political contributions, 10 percent of the value to its members of what it is buying, then the ability to deliver a dollar's worth of benefits is worth only $0.10 to the politician delivering it. A second explanation is the inefficiency of even relatively efficient transfers; a government expenditure of $10 million on behalf of some interest group may provide them with only $1 million worth of benefits. The difference between cost and benefit represents, in effect, the cost of hiding the transfer; a more direct and less inefficient arrangement would also be more obvious to the victims, hence less politically attractive. Combining the two effects would mean that a politician controlling $10 million in expenditure would end up with only $100,000 in campaign contributions.
A final explanation is that much of the cost of buying a political office never appears in records of campaign expenditure, not even the politician's private records. It consists of promises of a share of the loot--or, to use less loaded language, political commitments given to individuals and groups in exchange for their support.
1. Suppose the United States, with a monopoly of wheat production, imposes an export tax on wheat in order to raise the world price and exploit its monopoly position. The revenue from the tax is used for general government expenditure. Are American wheat farmers better or worse off as a result of the tax? You may assume that they are a very small fraction of the population and so get a negligible fraction of the benefit from the money collected.
2. It is easy to prove, for the wheat-auto case analyzed in this chapter, that a tariff and an export tax have the same effect; they tax the same transaction (trading wheat for autos) at different points. Similarly, an export subsidy and an import subsidy have the same effect. Yet we observe that tariffs and export subsidies are reasonably common, while export taxes and import subsidies are rare. Why?
3. In proving the inefficiency of tariffs, we used ideas from many parts of the book. Draw a diagram showing the relevant ideas and how they connect. Each box should be labeled by a chapter and an idea in that chapter. While I do not expect a complete diagram, you should show the major items; I would expect at least ten boxes, and probably more.
4. We have proved that tariffs--taxes on imports--are typically a Marshall worsening. Does that imply that subsidies to imports would typically be a Marshall improvement? Discuss.
5. It is commonly said that we need auto tariffs to protect the jobs of American auto workers. Where in our calculations did we include the value to the workers of the jobs they will lose if tariffs are abolished? Explain.
6. Is the fact that so many people bother to vote evidence against the analysis of voting in this chapter? Discuss.
7. Plumbers were for many years successful in getting city building codes to prohibit new technology, such as plastic pipe, which reduced the demand for their services. The analysis of this chapter suggests that a declining industry with lots of capital in the form of sunk costs, such as obsolete factories, would be relatively successful in obtaining favorable legislation. What form do you think the sunk costs of the plumbers took?
8. When a tariff is imposed on some particular import good, the beneficiaries are the domestic producers that the imported good competes with. The victims include not only consumers of that good but also producers of the export goods that would have been exported in exchange for the imports. There is no particular reason to expect the typical export industry to be any less concentrated than the typical import-competing industry. Nonetheless, exporters, like consumers, are a dispersed interest. Explain.
9. When auto tariffs are debated in congress, the principal opponents tend to be distributors and retailers of foreign autos--even though the cost to them must be a small fraction of the cost to consumers. Explain.
10. Discuss the rickshaw surplus of Chapter 7 in terms of rent seeking.
The article that first used the term "rent seeking" is:
Anne Krueger, "The Political Economy of the Rent-seeking Society," American Economic Review, Vol. 64 (June, 1974), pp. 291-303.
Other forms of essentially the same idea appeared earlier in:
Gordon Tullock, "The Welfare Costs of Tariffs, Monopolies and Theft," Western Economic Journal, Vol. 5 (June, 1967), pp. 224-232.
David Friedman, The Machinery of Freedom: Guide to a Radical Capitalism (New York: Harper & Row, 1971; Arlington House 1978; Open Court, 1989), Chapter 38.
Other sources for public choice theory include: James Buchanan and Gordon Tullock, The Calculus of Consent (Ann Arbor, MI: University of Michigan Press, 1962); Anthony Downs, An Economic Theory of Democracy (New York: Harper & Row, 1957); William Niskanen, Bureaucracy and Representative Government (Chicago: Aldine-Atherton, 1971); and Mancur Olson, The Logic of Collective Action (Cambridge, MA: Harvard University Press, 1965).
A discussion of the market for legislation similar to that in this chapter appears in Gary Becker, "A Theory of Competition Among Pressure Groups for Political Influence," Quarterly Journal of Economics, Vol. 98 (1983), pp. 371-400.
Hotelling's original contribution is in Harold
Hotelling, "Stability in Competition," Economic Journal, Vol.
39, No. 1 (March, 1929), pp. 41-57.