Forecast Accuracy Uncertainty and Momentum

Abstract

We demonstrate that stock price momentum and earnings momentum can result from uncertainty surrounding the accuracy of cashflow forecasts. Our model has multiple information sources issuing cashflow forecasts for a stock. The investor combines these forecasts into an aggregate cashflow estimate that has minimal mean-squared forecast error. This aggregate estimate weights each cashflow forecast by the estimated accuracy of its issuer, which is obtained from their past forecast errors. Momentum arises from the investor gradually learning about the relative accuracy of the information sources and updating their weights. Empirical tests validate the model's prediction of stronger momentum in stocks with large information weight fluctuations and high forecast dispersion. We also identify return predictability attributable to changes in the information weights.