A review of Bergman, N. K. and S. Roychowdhury, 2008, “Investor Sentiment and Corporate Disclosure,” Journal of Accounting Research 46 (5), 1057-1083.
I. Research question and its importance
Drawing on the literatures of behavioural finance, this paper investigates the relation between investor sentiment and firm disclosure policies. Because investor sentiment induces expectation bias over future earning, it leads to mispricing which managers have the incentive to respond strategically via their voluntary disclosures. Specifically, the authors focus on long-horizon disclosure because the desire to preserve high sentiment is more concentrated than it is in short-horizon disclosures. For example, managers might be incentivized to issue walk down forecast in short horizon to protect themselves from litigation. Moreover, the authors investigate whether such disclosure policy is associated with analyst pessimism in order to separate the target audience from the general investors. In brief, this paper is amongst the few studies that delve into the link between macroeconomic factor and firm disclosure.
II. Method, finding, and limitation
The authors follow conventions in the literature. For one, they proxy investor sentiment using quarterly average Michigan Consumer Confidence Index. For another, they proxy firm disclosure policy using frequency of long-horizon forecast (90 days or more) over all forecasts. Disclosure data was retrieved from First Call, a database where firms can voluntarily disclose without regulatory restriction prior to the Regulation Fair Disclosure (FD) that came effective in the last quarter of 2000. Controls such as seasonality (via GDP and market index prior to the current quarter), institutional ownership and analyst pessimism (indicator that equals to 1 if the estimate is below medium analyst forecast) are included in the main specification. Their findings unveil an increase in long-horizon forecast during periods of relatively low sentiment, and a decrease during periods of relatively high sentiment. Moreover, after controlling for analyst pessimism, such association remains significant, pointing to the attempt of managers to communicate with investors in large. However, the association between investor sentiment and short-horizon disclosure is mixed, as presented in an additional analysis. Nonetheless, the authors deliberately cut analyst estimate after three quarters without presenting the distribution of it. Hence, reader is left unaware about how much data was excluded for this variable.
III. Future research
The paper document the association between investor sentiment and firm disclosure. Future research shall advance this research by investigating the causality and effectiveness of disclosure over investor behaviours. However, as mentioned in the paper, Regulation FD imposes restriction over voluntary disclosure and inevitably increase the frequency of optimistic forecast. Hence, future work shall either correct for this skewed forecast distribution or adopt new measure such as natural language algorithm applied (e.g., with R or Python) to corporate press releases and lands at an optimistic score that enables the exploration of association between firm disclosure and investor behaviours.