A multivariate analysis of annual earnings forecasts generated from quarterly forecasts of financial analysts and univariate time series models / BEBR No.561

Includes bibliographical references (p. 26-27)."The study compares the forecast accuracy of financial analysts, ARIMA models, and various permier models considered in the literature in the predicting of annual earnings per share. Various refinements were made of previously used methodologies. The results of the multivariate analysis indicated that financial analysts provide the most accurate forecasts. In addition, the divergence in accuracy between the various sources of forecasts tend to decrease as the end of the year approaches, while at the same time there is a general increase in accuracy. Also specific results are provided for individual model performance.