Long-Term Economic Trends
If subsidized prediction markets are implemented, one of the most useful would be one that forecasts long-term economic trends. Policy debates frequently center around questions of the long-term economic outlook. For example, the political parties may debate the consequences of current spending for the budget deficit and the accumulated debt. Although the Congressional Budget Office estimates the effect of policies on the deficit and has established a reputation for relatively independent analysis, the office necessarily makes its projections on the basis of current law and current policy options, without ordinarily considering how decisions now might lead to legal changes later.51 A prediction market might better show how changes in policy have affected the outlook for the future. Moreover, the parties may disagree about how to calculate indirect effects of policies. For example, Republicans are more likely than Democrats to urge that tax cuts will in part pay for themselves by spurring economic growth and thus greater tax collection. Although the Congressional Budget Office can develop an approach to resolve such methodological disputes, there will always be the danger that politics will have affected decision making. And if the office consistently opts for the approach that leaves the least room for subjective decision making (for example, by adopting an assumption that tax cuts will not affect growth), that approach might not be the most accurate. The sponsor of a prediction market making such projections, in contrast, does not need to specify a methodology in advance. Instead, traders will have incentives to make their own assessments of the extent to which tax cuts will finance themselves. Even if privately sponsored, a prediction market projecting the debt in various years in the future could play an important role in public policy debates. For example, the media might come to report on how the prediction market’s projections have changed as the president announces new policies. The national debt, of course, is not the only economic variable of interest. Prediction markets also might be used to assess such variables as economic growth, the unemployment rate, the degree of economic inequality (measured, for example, by the variance of income across the population), and the savings rate. A separate prediction market contract might be used to measure each of these variables for each of a number of years in the future. This would provide a snapshot of the consensus view of how economic conditions are likely to change and make it possible to examine how this consensus view of the future has changed over time. It would also be possible, for one or more years, to use a number of prediction markets to estimate the probability that each of these variables will equal certain values. For example, one contract might estimate the probability that economic growth in 2020 will exceed 0.0 percent, another would estimate the probability that it will exceed 0.1 percent, and so on for a variety of plausible positive and negative values of economic growth. In effect, by expanding the number of prediction market contracts, it is possible to produce not only a point estimate of a future economic indicator but also a probability distribution. Another useful set of prediction markets might be used to help policy makers and individual investors anticipate the long-term future of Social Security. Advocates of a public social insurance system argue that such a system reduces the risk to future retirees, whose money will not depend on the gyrations of the stock market. Advocates of privatization respond that public social insurance introduces a new kind of risk–political risk–because Congress in theory can change the amount of Social Security payments at any time.52 It is difficult, however, to obtain a sense of the relative size of these risks. A set of prediction markets might forecast, again using multiple contracts to create probability distributions, what payments each of a variety of hypothetical individuals can expect from Social Security. At the same time, other prediction markets might produce estimates of the probability of different returns if a set amount of money were placed in some private investment vehicle such as an index fund. Of course, to obtain reasonable probability estimates, such a market would need a substantial subsidy, and still politicians and economists legitimately might offer a range of opinions. But if prediction markets become better understood, if only among policy elites, such markets could become standard reference points in political debates. Studies indicate that market-generated predictions of macroeconomic data releases avoid some anomalies previously found in surveys of professional forecasters.53 Markets could be used to predict noneconomic variables as well. There are many noneconomic contexts in which the task of assessment aggregation also might be useful. For example, prediction markets might be used, at least if information is available ex post, to predict the number of abortions that women in the
Leave a Reply