Overprecision in the Survey of Professional Forecasters

Abstract

Every decision depends on a forecast of its consequences. We examine the calibration of the single longest and most complete forecasting project. The Survey of Professional Forecasters has, since 1968, elicited predictions of key economic indicators such as unemployment, inflation, and economic growth. Here, we test the accuracy of those forecasts (n = 12,359) and measure the degree to which they fall victim to overconfidence, both overoptimism and overprecision. We find forecasts are overly precise; forecasters report 53% confidence in the accuracy of their forecasts, but are correct only 28% of the time. By contrast, forecasts show little evidence of optimistic bias. These results have important implications for how policy makers ought to interpret forecasts. Moreover, we employ novel methodology in analyzing archival data: we split our dataset into exploration and validation halves. The current manuscript presents results from the exploration half. The final manuscript will include results from the hold-out sample, as guided by editorial input.