| Past studies have established links between analysts' research reports and a variety of market variables. Nevertheless, they fail to deal with the central issue of attribution and the possible confusion of cause and effect. The primary intent of the study is to examine the relationships between analyst earnings forecast revisions and the information reflected in prior stock price movements, analyst earnings forecasts, and analyst recommendations in a more dynamic representation.; The empirical evidence shows that the probability of downward consensus earnings forecasts is significantly increased following negative abnormal stock returns, negative prior percentage changes of consensus earnings forecasts, and prior downward analyst recommendation revisions in Logit models. Furthermore, consensus earnings forecasts are significantly explained by prior stock price movements and prior percentage change of consensus earnings forecasts, in terms of magnitude and frequency, in OLS regressions. Given longer revision periods for consensus earnings forecasts and a smaller number of analyst earnings forecasts, the impact on magnitude of consensus earnings forecasts from prior cumulative abnormal returns and percentage change of consensus earnings forecasts becomes stronger. In causality tests, the results fail to support the position that stock price movements add no useful information to analyst earnings forecast revisions. This study demonstrates that it is not necessarily true that financial analysts are always information generators. Some or many analysts may follow stock price movements or other analysts' opinions, rather than being responsible for them. |
