Value-Relevance of Good-News Versus Bad-News Management Earnings Forecasts

We posit that the market views bad-news management earnings forecasts as more credible and value-relevant than good-news forecasts not because good-news forecasts tend to be biased, but rather because they are far noisier than bad-news forecasts. After controlling for noise, the difference in market response to good and bad news disappears. We demonstrate that bad-news forecasts show much lower dispersion around final earnings and, unlike good-news forecasts, they become considerably more accurate over time. These results provide direct support for the hypothesis that management perceives incentives to verify information when the news is bad as well as to delay bad-news announcements to allow time to gather additional information. These findings provide a rational explanation, not based on bias, for why on average the market’s reaction to good-news management earnings forecasts is weaker than to bad-news forecasts.