How Clear is the Crystal Ball?

Reflections on the accuracy of growth forecasts

Prakash Loungani

Published: March 2001

Two salie nt features of growth forecasts are discussed. First, recessions generally arrive before the forecast. Slowdowns are predicted but forecasters are unable or unwilling to call recessions. Second, private sector forecasts tend to be similar to those of official agencies. Some tips for forecast users are provided.

Download Paper in PDF format

More Papers From This Author in World Economics:

Economic Forecasts

Will the US and other economies slip into recessions this year? Which economies will decouple from a global slowdown? The authors suggest that the excessive caution shown by private sector forecasters limits the usefulness of their forecasts in answering these questions. Using growth forecasts for 14 major economies from 1990 onwards, they demonstrate that revisions made to forecast are extremely smooth. An implication of this smoothness is that forecasters do not call a recession until fairly late in the year in which the recession occurs. They are also slow to absorb news about developments outside their own economies, limiting the ability of the growth forecasts to predict the extent of decoupling.

Read Full Paper >

"There Will Be Growth in the Spring"

Forecasters are currently echoing Chauncey Gardner’s words that “There will be growth in the spring”. Or certainly by the summer. Are such forecasts credible? Yes. This article presents evidence that private sector forecasters have done a reasonably good job of forecasting recoveries in industrialised countries over the 1990s. Since recessions in these countries have tended to last under a year, forecasting a recovery in the following year has turned out to be a pretty good bet. However, a few recessions do end up lasting longer than a year: when that happens, the evidence suggests that forecasters are caught flat-footed.

Read Full Paper >