[Author’s Name]
[Institution’s Name]
Essay on Macroeconomic Forecast
Introduction
Analysts at a number of federal agencies are required to generate published but highly uncertain forecasts of economic and demographic conditions well into the future. The Office of Management Budget (OMB) reports underlying economic assumptions that affect aggregate spending and revenue estimates. The Social Security Administration (SSA) relies on economic and demographic assumptions to estimate the long-term solvency of the retirement and medical security trust funds. The Congressional Budget Office (CBO) produces independent measures of many of these estimates (and has also initiated an effort to estimate the uncertainty underlying the projections).
How accurate is economic assumptions and long-range projections produced by these agencies? Macroeconomists interested in the performance of private forecasters have generated a substantial literature and a set of tools for evaluating forecast performance. There have been a few efforts to make similar assessments of government forecasts. We nevertheless know very little about the extent to which performance varies across government agencies or the types of systematic bias observed in government forecasts.
This paper compares actual and projected outcomes for a limited number of macroeconomic indicators the inflation rate, a short-term interest rate, annual percent change in the real gross domestic product, and the unemployment rate for three agencies. The bias and efficiency of the forecasts are estimated using tests drawn from the macroeconomic literature on private forecast performance and some alternative explanations for observed bias are considered.
The principal substantive conclusions of the paper are that many short-term forecasts are unbiased while long-term forecasts are severely biased. The absence of short-term bias, a surprising result given that private sector forecasts are often biased, suggests the absence of strong incentives for forecast distortion. But a number of agency forecasts are inefficient in the sense that information available to other agencies is not fully incorporated in each forecast.
