The Information Content of Accounting Disclosures for Forecasting Sales of Manufacturing Firms

Abstract

We identify accounting disclosures of manufacturing firms that may be useful for forecasting year-ahead sales and develop a forecasting model incorporating these disclosures. We find that the most useful disclosures are prior sales growth, days sales outstanding, employees, margin and order backlog. We find that our forecasting model incorporating these disclosures provides forecasting ability incremental to that from analysts’ forecasts alone. In return-based tests, we find that a strategy of buying (selling) the 20% of firms with the largest positive (negative) difference between the model-based sales growth forecasts and analysts’ sales growth forecasts results in an annualized risk-adjusted return of 6.04%. The pattern of these returns is consistent with analysts over-extrapolating prior sales growth and markets discovering the analysts’ errors during the subsequent year, and with our model identifying firms where this has occurred. We conclude that the information in accounting disclosures about future sales is not fully impounded in analysts’ forecasts or in stock prices.