Date of Original Version

6-2008

Type

Working Paper

Abstract or Table of Contents

This study examines how accounting conservatism may affect the information environment of analysts’ earnings forecasts, taking into account the interaction between unconditional and conditional conservatism. Unconditional conservatism preempts conditional conservatism in the later period and reduces the uncertainty in loss recognition associated with bad news. Through a simple analyst forecast model, I demonstrate that: 1) unconditional conservatism is negatively correlated with analysts’ forecast errors for good news or mild bad news firms, but positively correlated with analysts’ forecast errors for extreme bad news firms; and 2) unconditional conservatism reduces the overall uncertainty in analysts’ forecasts. The empirical results are consistent with the predictions. Moreover, the evidence shows that the impact of unconditional conservatism on analysts’ forecasts is greater for early forecasts, when the information uncertainty is high, than for late forecasts.