An Empirical Estimation of the Investor Loss Function Associated with the Use of Analysts' Forecasts of Earnings

The aggregate investor loss function associated with the use of analysts' forecasts of earnings expresses the relation between loss and prediction error. Prior research suggested that the shape of this loss function should provide guidance in decision rules for investors. This study empirically estimates the investor loss function and finds an asymmetry in the relation between loss and analysts' prediction error. For overestimates of earnings, loss is a quadratic function of error; while for underestimates of earnings, loss is a linear function of error.