Endogenous Information Flows and the Clustering of Announcements

Viral Acharya is a Professor of Finance at New York University.

One of the most important ingredients to the process of price discovery in financial markets is the flow of new information, particularly apparent during times of market “crisis,” when it often seems that bad news is being reported simultaneously from multiple sources. While it is not surprising that firms’ news are affected by market and sector conditions (given the correlation of their cash flows), the timing of the announcements is suggestive that these disclosure decisions are not made independently. Indeed, recent empirical work suggests that corporate earnings warnings within an industry are clustered and that firms speed up their warnings in response to poor market conditions. Interestingly, however, such clustering is asymmetric in that good news does not generate such clustering, only bad news does. In the paper, Endogenous Information Flows and the Clustering of Announcements, forthcoming in the American Economic Review, my co-authors (Peter DeMarzo and Ilan Kremer, both of Stanford University) and I provide a theoretical explanation for this asymmetry.

We study disclosure dynamics when a firm possesses information that is correlated with market conditions, and explore incentives of managers – who maximize the present value of their compensation tied to firm’s market value – to delay the disclosure of news, until news become public knowledge. Because investors are uncertain whether a manager has learned the information, in equilibrium only those firms that have sufficiently positive news release their information.  Firms with more negative information prefer to keep their market value higher – at least temporarily – by claiming that they do not yet have any information to report.  Importantly, external public signals about market conditions that are correlated with firm’s own news are also likely to reach the market at future dates. Then, the firm has the opportunity to disclose its information before or after the public news announcement.

Our main theoretical insight is that because disclosure is irreversible, the firm faces a “real options” problem with regard to its disclosure decision: disclosing positive information may raise the stock price immediately, but this way the firm gives up the option that the external public news would have had an even more positive impact on the stock price if the firm had not yet disclosed.  As a result, information disclosure is delayed relative to the no public news case.  Indeed, there can be an “information blackout” period prior to the public news, when firms refrain from any voluntary disclosures. However, once the public news is released, if it is sufficiently negative it may trigger an immediate disclosure by the firm.

An important implication of our results is that the process of information arrival to markets is different from the process of information arrival to firms and managers. One, the underlying information process may have constant variability over time and no skewness, but this need not be true of the process describing disclosed information:  In periods without public news, stock returns will be positively skewed as the firm voluntarily releases good news; however, when public news is announced, returns will be negatively skewed.  Second,  as documented empirically, correlations between individual stock returns and the market returns will be much greater for downside moves, especially for extreme downside moves, than for upside moves, and  that these correlations will differ from the conditional correlations implied by a normal distribution.  Third, the asymmetry in response to public news and the resulting downside correlation of firm returns helps explain the empirical result that while individual stock returns will tend to be positively skewed on average, stock market indices will tend to have negatively skewed returns.  Finally, the arrival of adverse public news during market downturns will accelerate the disclosure of information by firms and result in greater volatility.

To summarize, skewness and volatility related patterns observed in stock returns are consistent with the dynamics of disclosures by firms and the incentives of managers who have discretion over disclosure timing.  In contrast to the existing literature which has often treated such patterns as a statistical artifact of data, we provide a common information-theoretic foundation for their existence.

The full paper is available for download here.

Both comments and trackbacks are currently closed.