|
|
|
||||
|
|
|
||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Source: Economic Review, Federal Reserve Bank of Richmond, Jan/Feb 1985, p. 26. |
The average forecast obviously missed a large amount of the variation, especially of the rate of inflation. None of the individual forecasts, some of which are generated with the aid of the large and expensive computerized models, were substantially better than the average. Though economists once thought that they could, with enough time and effort, build models that would be highly accurate, many now doubt if such models are possible.
One reason economists may not be able to predict well is that the effects of policy depend on people's expectations. Because people's expectations depend partially on past policies, expectations are always changing.
To illustrate why expectations can be important, consider how a profit-maximizing business will react if it believes that price controls may be imposed in the future. It will be ready to raise prices but hesitate to cut them because if controls are imposed, raising them again will be very difficult. If it gets caught with prices high when controls are imposed, the business will not be hurt. But if it gets caught with prices down, it will be hurt. In making pricing decisions, the firm will have to consider these factors in addition to other factors in determining prices. As a result, a business that normally reduces prices in periods of recession may hesitate to do so if it fears price controls.
One might argue at this point that expectations should not matter, since businesses can always raise prices once it becomes evident that controls will be imposed. However, the government recognizes this possibility as well and, as a result, controls often have a retroactive feature in them--prices are fixed at values they had several months before the legislation takes effect, or at some average value of the previous period. Hence, the firm cannot be sure that lower prices now may not affect its ability to charge higher prices in the future if controls are a possibility.
The idea that expectations matter has many implications. It suggests that people will react differently to an anti-inflation policy if they believe that policy makers will stick to it than if they believe that policy makers will give up when unemployment rises. It suggests that people will react differently to a policy when they expect it than when they do not. And it implies that because expectations can change quickly and because expectations are difficult or impossible to include in computerized models of the economy, these models, no matter how complex, will not always give reliable forecasts.
A final problem in stabilization policy is that good economic policy only is enacted when it is also good political policy. A policy that looks wonderful to an economist may not look wonderful to a politician trying to stay in office. Economists have had a tendency to neglect this problem by assuming that economic policy makers are different from the rest of us, that self-interest motivates ordinary people but, by hidden assumption, the desire to serve the public good motivates policy makers.
Once one lets policy makers become interested in their own self-interest, two complications arise. First, politicians want to win elections. Incumbents do not win elections if economic conditions are bad and the electorate holds them responsible (which in the United States is usually true of the president's party). Further, some policies have short-run effects that are desirable but harmful long-run effects. Other policies have harmful short-run effects but desirable long-run effects. It seems reasonable for politicians to follow policies with good long-run effects immediately after an election and policies with good short-run effects right before elections. If they do, there should be movements in the economy that correspond to the election cycle. Economists have searched for these patterns, but have not been able to find them consistently in the data. Perhaps decision making is too split up in democracies to permit such patterns, or perhaps politicians and their economic advisors do not know enough about the effects of their policies to be able to successfully work this strategy.1
Second, some economists have argued that politicians tend to misuse economic theories about stabilization. One of the conclusions from Keynesian theory is that a government deficit can be, in times of recession, desirable. Politicians, however, like deficits all the time. They get votes by spending and lose votes by taxing. Hence they will use, or misuse, any theory that will justify a deficit. According to this argument, politicians will use stabilization arguments to expand the deficit during recessions. They will be reluctant to admit that a recession has ended, or they will foresee threats of future recessions because the stabilization theory says the deficit should be eliminated in periods of boom. This argument suggests that failure to include a theory of politician behavior in discussions of stabilization policy has contributed to the persistent deficit spending that many democracies have encountered.