July 11, 2003

At the beginning of the first and third quarter of every year, we take the average of the individual forecasts in "The Wall Street Journal's Forecasting Survey" as a measure of the economists' consensus forecast. From this extraction, we then determine when an individual forecast is significantly different from that of the consensus. Fifty-four panelists participated in the most recent survey, each panelist was asked to make 10 forecasts, hence we have 540 separate forecasts. Under the null hypothesis of a normal distribution, and assuming the forecasts are independent of each other, we would expect that out of 540 "random" drawings we would get 5% or 27 observations that we would consider significantly different than the mean forecast.

You must have an active account to view these reports. You may register for a trial here

Download Complete Report in PDF Format

Download Complete Report in Word Format