Files
Abstract
Optimal, rational forecasts are often biased and thus the empirical finding that actual forecasts are biased is not necessarily evidence of irrational behavior. Using an asymmetric loss function and a regression forecasting problem, these points are explicitly demonstrated.