What Government Officials Cannot Know

Mr. Summers is a member of the staff of The Foun­dation for Economic Education.

As government grows ever larger, it is perhaps wise to ask if there are limits to the knowledge of govern­ment officials. If there are limits, and if government has grown be­yond these limits, then government officials are intervening in matters of which they are necessarily ignorant—can have no knowledge.

Let us concentrate on a basic form of ignorance familiar to all: igno­rance of what other people are thinking. In particular, consider what a government official can and cannot learn about the preferences, expectations, and plans of his fellow men. Using our results, we can sketch some areas in which govern­ment officials cannot know what they are doing.

What can be learned about another person’s preferences? Sup­pose we see Mr. Smith buy rye bread for one dollar. This tells us that at the time of the purchase he valued the rye bread more than the dollar. If there was pumpernickel on the next shelf, and if we assume that he noticed it, we can also conclude that at the time of the purchase he pre­ferred rye bread over pumpernickel.

But will he prefer rye bread to a dollar tomorrow? What about his tastes in other goods? How about all his other preferences? And did he really notice the pumpernickel?

To find out, we can ask. But what a person says he would prefer in a given situation is often different from what he actually prefers when the time comes to choose. There are several reasons for this. First, a person’s preferences may change be­tween the time he is questioned and the time he faces a given choice. Second, when a person faces a situa­tion he may find that it is different from what he had imagined it would be when he was questioned. In addi­tion, he may not completely under­stand the questions, be in a rush to complete the questioning, give less than truthful answers, or simply have no idea of what he would prefer at some later date.

Thus, the only way to be sure of a person’s preferences is to observe his choices. And even then we can only learn a person’s past preferences, because the next time he might choose differently.

Suppose now that, instead of hav­ing the dollar available to buy rye bread, Mr. Smith had paid the dollar in taxes. Then we can never know what Smith would have done with the dollar, had he been free to use it as he saw fit. If the government gives the dollar to Mr. Jones, we can, of course, observe what Jones does with it and thus learn some­thing about his preferences. But, be­cause Smith no longer has the dol­lar, we can never learn the use he would have preferred.

Thus a government official, in re­distributing wealth, cannot know how the taxpayers would have pre­ferred to use the money they paid in taxes.

But doesn’t the government, in redistributing wealth, help poor Mr. Jones more than it harms rich Mr. Smith? That is, doesn’t the redis­tribution cost Mr. Smith less than it benefits Mr. Jones?

No Way to Know

We have seen that a dollar has gone from Smith to Jones. How much did this dollar benefit Mr. Jones? We can observe what he does with it, but what does that tell us? If he spends it for eggs, we know that at the time of the purchase he valued the dollar less than the eggs. But this tells us nothing about the value he placed on the dollar compared with the value Mr. Smith placed on the dollar. In fact, because Smith no longer has the dollar, we cannot even learn what he would have done with it. Smith himself may not even know, because who knows what he will ultimately do with every dollar?

Thus, there is no way to know if the redistribution of wealth bene­fited Jones more than it cost Smith. In fact, there is no way to know if any government intervention yields benefits which are greater than the costs.

This is because costs and benefits cannot be measured. In our exam­ple, the redistribution cost Mr. Smith whatever he would have done with the dollar. There is no way to measure how much he valued this lost opportunity. In fact, the value he placed on the dollar probably changed over time—as his income changed, prices changed, and other circumstances changed. Similarly, the value Mr. Jones placed on the dollar, and thus the benefit he ac­quired in receiving it, also probably changed over time.

If a government official cannot know how much an intervention costs one individual and how much it benefits another individual, he certainly cannot add unknown costs and benefits to find "social costs" and "social benefits." Costs and ben­efits have meaning only to individ­uals and can be known only by the individuals directly involved. In particular, we cannot assume that a "uniform" intervention, such as a uniform tax of $1000, imposes the same costs on every person, because we don’t know the value each person places on $1000—nor can we assume that these values are constant. Simi­larly, if everyone must wait half an hour to buy price-controlled gas, we cannot assume that waiting imposes equal costs on every person.

Consider, for instance, the costs of a volunteer army versus the costs of a conscripted army. It is often as­sumed that because a volunteer army requires higher military pay than a conscripted army, that a vol­unteer army "costs" more.

This is an unwarranted assump­tion. We cannot know the true costs of the higher taxes needed to attract volunteers because we cannot mea­sure the value each taxpayer places on the money he pays in taxes. In addition, we cannot know the costs that conscription imposes on draft­ees. We can assume that draftees would rather be civilians, otherwise they would have volunteered. But no one can measure the value each draftee places on a return to civilian life. In short, there is no way to com­pare the costs borne by taxpayers versus the costs borne by draftees.

Cost-Benefit Analysis

Conscription is an example of the error in trying to place a dollar fig­ure on the costs and benefits of gov­ernment intervention. "Cost-benefit analysis" cannot be used to justify government intervention because no one can measure the true costs and benefits.

In fact, cost-benefit analysis can be used to argue against all inter­vention. If someone wants some­thing enough to bear the costs, he will voluntarily take appropriate ac­tion. If he believes that the benefits do not exceed the costs, he will ab­stain from such action. Government intervention tries to force people to bear costs that, as those people see it, exceed the corresponding bene­fits.

Public goods are another example of intervention based on ignorance. It is argued that government must provide services such as welfare, parks, public schools, public housing, and public television because, although "everybody wants them," no one is willing to pay for them except if all his neighbors are taxed to help pay the bill.

This is a peculiar line of thought. We are told that people want some­thing, though they won’t voluntarily pay for it! Not only is this absurd, it assumes that the proponents of pub­lic goods know how people would use their money if it didn’t go for taxes. As we saw earlier, the only way to know how a person prefers to use his money is to stand back and let him use it.

Antitrust is another area where government is intervening in igno­rance. At the root of antitrust legis­lation is the assumption that, under certain conditions, a firm can raise its asking prices, sell less, and make more money than if it sold more items at lower prices.

In theory, this may occur. But no one knows if it actually happens or when it happens.

Suppose, for instance, that Na­tional Motors is the only manufac­turer of a unique sports car and produces a limited supply which all sell at a "high" price. Did they make more money than if they manufac­tured more of these cars and sold them all at lower prices?

No one knows. We know that each buyer, at the time of purchase, pre­ferred the sports car over the money it cost him. But we don’t know how many people would have bought such a car at a lower price. That is, we don’t know how many sports cars could have been sold at lower prices so we don’t know if National Motors made more money selling its limited-edition at "high" prices. No one knows if National Motors reaped "monopoly gains."

Measuring Expectations

Let us now consider what can be learned about another person’s ex­pectations. Although observing a person’s actions yields limited in­formation about his preferences, it tells us nothing definite about his expectations. The most we can say is that a person is acting "as if" he expects something. A man carrying an umbrella may be acting as if he expects rain, but he may just be returning it to a friend. The prob­lems involved in asking about pref­erences carry over to asking about expectations, with a new twist—asking about a person’s expectations may change his expectations.

For example, a temporary in­vestment tax credit may be used to try to stimulate investment. The first time such a tax credit is intro­duced it might come as a surprise and produce the desired results. But if the stimulant is used again and again, businessmen may learn that a lagging economy is likely to be followed by a tax cut. If they expect a tax cut, they will tend to postpone investment until after the tax credit is enacted. Thus, the expectation of a tax credit may reduce, rather than stimulate, investment. And clearly, if a government official asks a busi­nessman if he expects a tax cut, the businessman may suspect that one is forthcoming.

A similar situation occurs with price controls. If businessmen expect price controls, they may prepare for them by raising their list prices and offering discounts, free delivery, better guarantees, and other im­proved terms of sale. When the con­trols arrive, and the government continues to inflate the quantity of money, the improved terms of sale will tend to disappear. Thus, the anticipation of price controls may increase, rather than decrease, list prices.

The lack of knowledge of expecta­tions is especially evident in mone­tary policy—the government’s tam­pering with money and credit.

For several decades, Keynesian economists have told us that the cure for unemployment is for the government to inflate the quantity of money so as to keep ahead of union wage increases—based on the assumption that union leaders would not anticipate inflation. But, as years have passed, it is clear that unions have learned to anticipate inflation and now demand raises plus cost of living clauses. Govern­mental ignorance of personal expectations has been an important factor in producing our current high rates of inflation and unemployment—which according to Keynesians should not occur together.

Controlled Inflation

The Monetarists prescribe another form of monetary intervention—that the government inflate the quantity of money at a fixed rate so that everyone could better antici­pate changes in the "price level." In this way, it is held, economic booms and busts could be avoided.

But, changes in the "price level" are not the cause of booms and busts. Rather, booms and busts are the result of government manipula­tion of money and credit.

For example, suppose the gov­ernment tries to stimulate the X industry by pumping in money and credit. This creates a boom in the X industry, as prices tend to rise in that industry, and natural re­sources, labor, and capital goods are drawn into it. But the increased bid­ding for natural resources, labor, and capital goods causes their prices to rise and the X industry experi­ences increased costs. The increased costs spell the end of the boom, and if the government doesn’t continue pumping money into the X industry at a faster rate than is generally anticipated, costs catch up with prices and the X industry suffers a depression.

Clearly, the situation is only wor­sened if the government tries to stimulate more than one indus­try—or the entire economy.

Thus, it is unanticipated injec­tions of new money at specific points in the economy that create booms, and unanticipated increases in costs that end the booms. To avoid these booms and busts it is not sufficient to inflate the quantity of money at a fixed rate so that people could sup­posedly better anticipate changes in the "price level." Rather, booms and busts can be avoided in an interven­tionist economy only if everyone correctly anticipates where the new money will be injected, when it will be injected at each particular point, and how the prices of specific factors of production will react—clearly an impossible task. The only way to end economic booms and busts is to end government tampering with money and credit.

In sum, unavoidable ignorance of expectations can make a shambles of government attempts to quarter­back the economy.

Other People’s Plans

Let us finally consider what can be learned about another person’s plans. As with preferences and ex­pectations, asking a person yields nothing definite about his con­stantly changing plans. The most we can say is that a person is acting "as if’ he has certain plans. But even this yields nothing definite. If we see someone cutting down a tree, are we to conclude that he is doing it to clear a field, for firewood, for logs to build a cabin, to sell to a sawmill, or just for the exercise? Or is it for a combination of reasons? We can con­tinue watching and see what he does after the tree is felled, but what he does after is not necessarily what he planned to do before.

Having no certain knowledge of other people’s plans, a government official cannot know how an interven­tion changes their plans. In particu­lar, he cannot know what plans they would have carried out were it not for the intervention. For instance, no one knows how entrepreneurs would carry letters were it not for the government’s postal monopoly.

Other examples abound. For in­stance, we know that an increase in the minimum wage is a disincentive to employ workers with low produc­tivity. But no one knows how many people would have planned to hire workers were it not for the minimum wage. The minimum wage disemploys workers, but no one knows how many—so no one can defend the minimum wage be­cause he "knows" the resulting number of unemployed will be small.

Similarly for all other interven­tions. We know that unions tend to scare labor and capital from certain industries, but no one knows how much. Rent control scares away new housing, but no one knows how much. Social Security reduces the means and incentive to prepare for retirement by making productive investments, but no one knows how much investment is lost. And on and on. Government officials have no way to measure the damage caused by interventions because no one can measure how interventions change people’s plans.

Having shown how the limits of knowledge can be used to argue against government intervention, a few words should be said about the absence of intervention—the free market. In the free market, prices are determined by the preferences, expectations, and plans of everyone in the market. Thus, although we still possess very little direct knowl­edge of the preferences, expecta­tions, and plans of others, market prices convey them to us indirectly

in highly condensed form. For in­stance, consumer preferences are expressed in the selling prices of consumers goods, expectations are expressed in stock market prices and futures prices, and entrepre­neurs’ plans are expressed in the prices of capital goods. If a busi­nessman acts in harmony with these preferences, expectations, and plans, he earns profits; otherwise he suffers losses. Thus, the free market profit and loss system tends to coor­dinate businessmen’s decisions with the preferences, expectations, and plans of their fellow men.

When government intervenes in the market, prices are distorted and thus no longer convey the prefer­ences, expectations, and plans of others. When government com­pletely controls the economy, "prices" lose their meaning, and we are governed by ignorance. 




What individualism teaches us is that society is greater than the individual only in so far as it is free. In so far as it is controlled or di­rected, it is limited to the powers of the individual minds which control or direct it. If the presumption of the modern mind, which will not re­spect anything that is not consciously controlled by individual reason, does not learn in time where to stop, we may, as Edmund Burke warned us, "be well assured that everything about us will dwindle by degrees, until at length our concerns are shrunk to the dimensions of our minds."

F.A. Hayek, Individualism and Economic Order