Whom Do You Trust? An Analysis of Executive and Congressional Economic Forecasts

Abstract

Economic forecasts play an increasing role in U.S. budgetary and fiscal policies. This paper analyzes the accuracy and bias of economic forecasts prepared by the Executive Branch and Congress. Short-run forecasts by the Executive Branch for the forthcoming year do not appear biased. They are as accurate as private forecasts and the forecasts of the Congressional Budget Office (CBO); they are not sensitive to political factors such as the size of the deficit. At the same time, the accuracy o f Executive branch's short-run economic forecasts has not improved over the period from 1962 to 1984, and the forecasts themselves can be well approximated by simple, single-equation models that use a small number of variables. In contrast to short-run forecasts, long-run economic forecasts of both the Executive branch and CBO are consistently optimistic. The degree of optimism increases with the length of the forecasting horizon and the degree o f fiscal pressure the year in which they are issued. The Reagan Administration's long-run forecasts, while typically optimistic, have not been rosier than those issued under previous administrations.

Political distortions in state forecasts

“…Thus we expect 3 See Poterba (1994) and Rubin (2006). 4 Kamlet et al (1987) examine the accuracy of Executive Branch and Congress forecasts of real GNP growth, inflation, and unemployment for the years 1962 to 1984. The authors expect forecasts to be more optimistic in years when an unbiased forecast would predict a lower level of growth.…”

Organizational Pressures on Forecast Evaluation: Managerial, Political, and Procedural Influences

“…Their ®ndings supported the earlier work and found that forecast accuracy is enhanced in an environment with strong institutional pressures and highly partisan politics. Kamlet, Mowery, and Su (1987) found no evidence that partisan politics biased short-term economic forecasts by the executive branch or Congress. On the other hand, Smith (1984, 1989) and Feenberg et al (1989) discovered a conservative bias or underestimation in state revenue forecasts explained by the need to generate a hedge or buer against future revenue uncertainty.…”

How good are US government forecasts of the federal debt?

“…Kamlet, Mowery, and Su (1987) compare one-yearahead and multi-year-ahead forecasts from CBO, OMB, their own ARIMA model, and the ASA/NBER survey for the real growth rate, inflation rate, and unemployment from 1976 to 1984. 3 They find that neither agency outperforms the other for short-term forecasts.…”