Investor Information Gathering and the Resolution of Uncertainty

Gathering additional information is an instinctive response to uncertainty. This behavior is found in many settings, perhaps most pervasively in capital markets. For example, if investors observe an earnings number that differs from what they expected, they might seek to better understand the number by gathering additional information and context from financial statements or corporate disclosures. Such efforts may not fully resolve the uncertainty that prompted them, particularly if uncertainty is more difficult to resolve as it increases. Thus, the intensity of investors’ information gathering efforts may reflect both the uncertainty that motivated their search and the residual uncertainty that remains unresolved after searching.

I find evidence consistent with this pattern, using data on public internet access of SEC filings around firms’ earnings announcements. Investors search more when there is greater uncertainty leading up to a firm’s announcement and when the announcement itself exacerbates investor uncertainty. Information gathering leading up to the announcement appears to reflect uncertainty about what the impending earnings will be, while information gathering after the announcement appears to reflect uncertainty about how to interpret the announced number. Further results suggest that these information-gathering efforts are not fully successful in resolving the uncertainty that prompted them.

What Are Potential Pitfalls in Analyzing Investors’ Information-Gathering Behavior?

This observation, that information gathering may not fully resolve the uncertainty that prompted it, suggests potential difficulty in studying the effects of investor information gathering. Recent research in accounting and finance has used a growing number of direct measures of investor information gathering, such as searches for stock tickers on Google, public internet access of firm-specific SEC filings, and firm-specific search activity on Bloomberg terminals. Because uncertainty likely motivates much of this behavior, it is difficult to interpret correlations between these measures and capital market outcomes, such as stock returns.

Studying the effects of information gathering would be much easier if investors sought information about random firms. We could then potentially assign a causal interpretation to correlations between measures of information gathering and capital market outcomes. For example, if the intensity of information gathering is correlated with a more efficient price response to earnings news, one might conclude that information gathering causes more efficient processing of earnings news by the market (presumably because investors have more information and can thus better interpret aspects of the earnings number, such as its persistence). The problem, of course, is that investors don’t search randomly.  When an increase (decrease) in uncertainty drives an increase (decrease) in the value of information, investors demand more (less) information.

Thus, proxies for information gathering may capture unique aspects of unresolved investor uncertainty, so it is difficult to interpret tests involving such proxies. For example, if the intensity of information gathering after an earnings announcement is correlated with a stronger response to earnings news, is this an indication that more investors knew about the earnings news or an indication that the announcement resolved more uncertainty? Or are investors gathering information in response to the stock returns themselves? In other words, it is unclear if correlations between these proxies and capital market outcomes reflect something about information gathering or something about uncertainty. One of the contributions of my paper is to highlight potentially problematic associations between proxies for information gathering and stock return patterns around earnings announcements.

Is There a Way Around These Pitfalls?

A potential empirical approach for addressing this problem is to control for the uncertainty that motivates information gathering. However, further analyses indicate that firm-specific information gathering reflects unique information about firm-level investor uncertainty that isn’t captured by commonly used measures of uncertainty. This finding is valuable because, although capital markets are rife with uncertainty, it is notoriously difficult to empirically measure investors’ uncertainty. Thus, proxies for information gathering may offer a unique additional dimension for capturing investor uncertainty.

Another empirical approach is to explore situations where information gathering by investors is unrelated to their uncertainty. For example, if something other than uncertainty limits an investor’s ability to understand a firm’s earnings, it may give researchers a way to control for the uncertainty that motivates information gathering. Prior research has found that investors have a limited ability to process a firm’s earnings news when they face a great number of same-day earnings announcements by other firms. This is an instance where something other than uncertainty may change the amount of information that investors seek about a company, particularly if the other same-day announcements don’t contain much information about that company.

I use the number of announcing firms outside the subject firm’s industry to capture variation in investor information gathering that is unrelated to uncertainty about the subject firm’s earnings. I find some evidence that this plausibly exogenous variation in investor search does not reflect investors’ uncertainty about the subject firm’s earnings. The identification of further settings like this is a promising direction for future research.

What Is the Takeaway?

Paradoxically, because information acquisition is a direct response to uncertainty (yet may not fully resolve it), proxies for information gathering may reflect meaningful and unique aspects of investor uncertainty. This complicates the interpretation of analyses involving such proxies, particularly because common measures of investor uncertainty don’t seem to capture aspects of uncertainty reflected by investors’ information gathering efforts. The identification of  settings that offer variation in information gathering unrelated to investor uncertainty is a promising direction for future research on the consequences of information gathering.

This post comes to us from Professor Jed Neilson at Pennsylvania State University. It is based on his recent article, “Investor Information Gathering and the Resolution of Uncertainty,” available here.