Is analyst optimism intentional? : additional evidence on the existence of reporting, selection and cognitive bias in analyst earnings forecasts

Abstract

In this dissertation I examine whether the documented bias in analyst forecasts of earnings is intentional by specifically testing for the existence of reporting, selection and cognitive bias. The first two types of bias assume that due to incentives, analysts knowingly alter their forecasting process while cognitive bias assumes that analysts report their true forecasts, which happen to be optimistically biased. In order for bias to be intentional analysts must first be given incentives to bias their forecasts and second they should respond to these incentives by either inflating their true forecasts (reporting bias), or by selectively not reporting them when they are pessimistic (selection bias). In this dissertation I test for both of these premises. With respect to the first premise, I find that employers give analysts some turnover incentives to bias their forecasts. Specifically, the probability that an analyst is demoted (promoted) is significantly lower (higher) when the error is negative (i.e. optimistic) than otherwise. With respect to the latter premise, I find that analysts take into account the market's ability to adjust for the bias when forming their forecasts. Specifically, analysts reduce the bias in their forecasts as the market's ability to adjust for the bias increases. These results suggest that at least a portion of the bias is intentional. The fact that the relationship between the market's ability to adjust for the bias and the level of bias in the forecasts is negative provides support for the existence of reporting bias in particular.