THE EFFECTS OF MANAGEMENT FORECAST PRECISION ON EQUITY PRICING AND ON THE ASSESSMENT OF EARNINGS UNCERTAINTY

This study examines the effects of management forecast precision (i.e., lack of uncertainty) on equity pricing and the assessment of earnings uncertainty. Kim and Verrecchia (1991) modeled the price reaction to the public release of information as a positive function of both the unexpected component of the information and the information's precision. We test these predictions with a sample of 868 management forecasts for 1983-1986 annual and interim earnings. The use of management forecasts rather than actual earnings to test the precision hypothesis has the distinct advantage that the level of forecast precision is not directly regulated and thus may vary across forecasts. Further, managers explicitly disclose their level of uncertainty. Both this study and Pownall et al. (1993) document that most forecasts are open-interval (minimums and maximums), closed-interval (ranges). or general impressions rather than point estimates. The method used to test the precision hypothesis removes restrictions on the traditional regression of unexpected returns on unexpected earnings. Specifically, the slope and intercept coefficients that map unexpected earnings into unexpected returns can vary in the cross-section as a function of forecast precision. Our results support a direct relation between forecast precision and the importance of management forecasts for security pricing. Holthausen and Verrecchia (1990) and Morse et al. (1991) modeled a decrease in investors' consensus as a positive function of the magnitude of signal surprise and the dispersion of the perceived precision of the signal. We examine these predictions with a sample of 221 point and closed-interval (range) forecasts. We calculate whether the range of outcomes disclosed by a manager exceeds the range of Institutional Brokers Estimate System (IBES) analyst forecasts. We find this variable and the magnitude of unexpected security returns (a proxy for signal surprise) to be positively associated with increases in the standard deviation of IBES analyst forecasts. Morse et al. (1991) found the hypothesized relation between signal surprise and increase in analyst forecast variance, but were unable to separate the precision effect from the signal surprise effect. Managers' explicit labeling of forecasts as more uncertain through range disclosure permits the direct calculation of management forecast precision relative to analyst forecast precision. Our tests involve joint hypotheses of the effects of forecast precision on security prices and the credibility of managers' disclosures of forecast precision. Ajinkya and Gift (1984) developed and tested the ''expectations adjustment hypothesis'' which posits sufficient incentives for credible, symmetric forecast disclosure. King et al. (1990) argued that expectations adjustment also suggests credible labeling of the precision of forecasts.