Forecasts of Earnings Announcements Dates: Evidence from a Mandatory Regime

This paper studies the trade-os in information content learned from forecasts of earnings announcement dates obtained from a mandatory forecasting regime (China) versus a voluntary regime (US). First, we show that forecasts in China, like the US, embed unrevealed rm performance information and that investors respond to forecast disclosures in a manner consistent with their information content. Furthermore, unique to the US, both the voluntary forecast decision and the timing of the forecast convey incremental information about rm performance. Unique to China, the timing of the mandatory forecast is uniformly xed and on average much earlier than their endogenous counterparts in the US. This feature enables us to estimate the trade-o between timeliness and accuracy of forecasts { a 10% increase in timeliness of forecasts increases absolute forecast errors by 7%. Despite this trade-o, 25% of earnings-related news is learned on the date of the initial forecast in China compared to 11% in the US relative to news learned on the date of the earnings announcement. Finally we ask whether China rms who were ostensibly the target of the mandatory regulation { the likely silent rms under a voluntary regime do in fact provide informative forecasts under the mandatory regime. We construct and estimate a linear forecasting model and show that the precision of the manager's private information is associated with the decision to forecast in the US. Using this parameter, we identify the counter-factually silent China rms and nd that these rms also provide useful information via their forecasts. Our collective evidence informs policymakers of a) the role of the uniform timing of the mandatory forecast in improving the timeliness of information concerning rm performance and b) suggests that mandatory forecasting induced otherwise silent rms to provide useful information, but at the loss of information from the voluntary disclosure choice itself.