The Superiority of Analyst Forecasts as Measures of Expectations: Evidence from Earnings

If both producers and consumers demand forecasts based solely on their forecasting ability, then the equilibrium employment of analysts, a higher cost factor than time series models, implies that analysts must produce better forecasts than time series models. Past studies of comparative earnings forecast accuracy have concluded otherwise. Using nonparametric statistics that provide proper yet powerful tests, we find that Box-Jenkins time series models consistently produce better forecasts than martingale and submartingale earnings models; but Value Line Investment Survey consistently makes significantly better earnings forecasts than the Box-Jenkins models.