We show that positive daily stock returns contain more information on the long-term change in stock value than do negative daily stock returns that are noisier on average and more prone to subsequent reversals. This difference in the informativeness of negative and positive returns is larger on nondisclosure days and decreases significantly on disclosure days. Our findings suggest that while positive information about firms is constantly flowing to the market, the flow of negative information is more limited and mostly confined to disclosure days.

There are reasons why negative information is more likely to reach investors on disclosure days rather than on nondisclosure days. Conservative accounting as well as litigation exposure compel managers to immediately disclose negative news while postponing the disclosure of positive information. Even if managers decide to withhold negative information, they are likely to do it only when investors cannot independently obtain this negative information. During nondisclosure days, managers could leak information to investors, but if they do, they are less likely to leak negative information before disclosing it to the public in the official channels. And even if investors obtain negative information on nondisclosure days, short-selling constraints could prevent them from profiting on the information. So disclosure and trading constraints can affect the extent to which negative and positive information are incorporated into prices – these two explanations are not mutually exclusive, and we examine the role of both in the differential informativeness and reversals of positive and negative returns.

To measure the information content in daily stock returns, we use two alternative measures. The first is return reversals. Stock returns reflect changes in fundamental value of the stock but can also be driven by noise. Noisy trading is expected to generate temporary price changes that will later reverse. Hence, larger (and more frequent) reversals of negative stock returns relative to positive stock returns would suggest negative stock returns are noisier and less information-driven than positive stock returns. We find reversals of negative stock returns are larger than reversals of positive stock returns mostly on nondisclosure days, and the difference in reversals of negative and positive returns decreases substantially on disclosure days.

Our second measure is the precision of information in daily stock returns. This variable is based on the association between short-term and long-term stock returns. We find that positive daily returns reflect more precisely the change in long-term value of the stock, compared with daily negative stock returns. That is because negative daily stock returns contain noisier information about the long-term value of the stock. In addition, the higher precision of positive stock returns relative to negative daily returns is more pronounced on nondisclosure days, and it decreases substantially on disclosure days.

Overall, both larger reversals and lower information precision of negative daily returns indicate negative returns are less informative about the change in the value of the stock mostly on nondisclosure days. These findings suggest investors get less negative information on nondisclosure days. We argue that leaks of more positive than negative information from firms is driving the difference in informativeness of positive and negative returns during the quarter.

To test our argument, we use Regulation Fair Disclosure (Reg FD). Reg FD, which became effective in October 2000, disallowed selective disclosure of information to investors. If indeed the information that firms leak is more positive than negative, we would expect Reg FD to reduce the information leakage mostly in positive daily returns. Using a difference-in-differences analysis, we indeed find that Reg FD reduced information in positive daily stock returns, but it did not reduce information in negative daily stock returns. The findings suggest firms leak more positive than negative information, and this asymmetric leakage leads to more positive information in returns during the quarter.

The second possible reason for lower information in negative returns on nondisclosure days is short-sale constraints. Investors may not only receive more positive information during the quarter, as we show here, but also be constrained from shorting stocks on the negative information they do have. We indeed find that when investors short sell more stocks during the quarter, they make negative returns more informative. However, this is the case only when short interests are very significant, and approximately 8.5 percent of outstanding stocks are shorted. Only then does the precision of information in negative stock returns during the quarter increase to a level similar to positive daily stock returns. These results show that when investors have negative information and can short the stock, negative returns are as informative as positive returns.

Investors may not short more stocks during the quarter because they do not get as much negative as positive information, or because constraints prevent them from acting on their negative information. To test the latter alternative, we follow prior literature and use institutional holding as a proxy for short-sale constraints, where higher institutional holding means more stocks are available for shorting and short constraints are lower. However, we do not find support for the short-sale constraints alternative with the institutional holdings proxy, as higher institutional holding does not increase the precision of information in negative returns. Still, we do not rule out short-sale constraints as a possible explanation. Short-sale constraints and costs are driven by many factors, which may not be reliably inferred from stock characteristics, and it is possible that we do not fully capture the short sale constraints and costs that investors face during the quarter. Anyway, our difference-in-differences test around Reg FD allows us that information leakage from firms has a substantial role in the positive slant in informativeness of returns during the quarter.

Our findings suggest that negative stock returns reverse more and contain less information than positive stock returns, mostly on nondisclosure days. Accounting rules and litigation risk compel managers to reveal their negative information in disclosures, and if managers withhold negative information, they do it when investors are less likely to independently find the information. Consistent with the hypothesis that investors receive less negative information, we find that on nondisclosure days negative stock returns reverse more and reflect the change in the fundamental value of the firm less precisely than positive returns. On news days, the difference between the informativeness of negative and positive returns decreases. Moreover, we demonstrate that asymmetric leakage of more positive than negative information from firms is behind the difference in the information in positive and negative returns during the quarter. Together, the findings suggest that stock returns on nondisclosure days contain more positive than negative information on the fundamental value of firms, and the information gap between positive and negative returns closes only on disclosure days.

This post comes to us from professors Eli Amir, Shai Levi, and Roy Zuckerman at Tel Aviv University’s Coller School of Management. It is based on their recent paper. “The Differential Informativeness of Positive and Negative Stock Returns,” available here.