ECONOMISTS' FORECASTS have gotten systematically worse over the years, right? Not at all, says Victor Zarnowitz, a research associate of the National Bureau of Economic Research. But the forecasts haven't gotten much better, either.
In an NBER working paper, Mr. Zarnowitz takes a historical look at various kinds of forecasts - quarterly, annual, consensus, individual and service bureau - and concludes that they have not become systematically less accurate or more biased over the last 30 years. In fact, he notes, "the annual forecasts of nominal and real gross national product growth may actually have improved, at least since the late 1950s."For example, the average absolute error in the predictions of annual GNP by business economists was 1.2 percent between 1959 and 1976 but only 1 percent between 1977 and 1984. On the other hand, their predictions of inflation were off by an average 1 percent from 1959-76 but 1.1 percent from 1977-84.
The major forecasting foul-ups, Mr. Zarnowitz finds, occur during recessions. During expansions, the typical forecaster misses the mark on real GNP by 1.2 percent and on inflation by 1.1 percent. During economic downturns, however, those forecasts were off by 3 percent and 2.5 percent respectively.
The biggest boo-boos occurred between the fourth quarter of 1981 and the fourth quarter of 1985. That time span also had the most unanticipated turning points: a severe recession in 1981-82, a slowdown in mid-1984 and a period of strong disinflation.