Most economic forecast evaluations dating back 20 years show that prof
essional forecasters add little to the forecasts generated by the simp
lest of models. Using various types of forecast error criteria, these
evaluations usually conclude that the professional forecasts are littl
e better than the no-change or ARIMA type forecast. It is our contenti
on that this conclusion is mistaken because the conventional error cri
teria may not capture why forecasts are made or how they are used. Usi
ng forecast directional accuracy, the criterion which has been found t
o be highly correlated with profits in an interest rate setting, we fi
nd that professional GNP forecasts dominate the cheaper alternatives.
Moreover, there appears to be no systematic relationship between this
preferred criterion and the error measures used in previous studies.