Knowledge is power, especially in the world of finance. Take investors, for instance, who typically gather and analyze information prior to trading. What information are they paying attention to, exactly? And how does the information help produce higher returns?

New research from Michigan State University, Rutgers Business School, Rice University and the University of Notre Dame is the first to provide a direct measurement of information processing by institutional investors across many stocks — showing that knowledge is power with a high ROI.

“We wanted to test the theory that stocks are riskier and earn higher returns when cross-learning occurs, like when peer firms have important announcements, or when there are important macroeconomic announcements,” Ryan Israelsen, assistant professor of finance and coauthor of the paper, explained. “Because these events often have important implications for other firms, investors know there will be a resolution of uncertainty and they require a return premium to hold the stocks as compensation for the risk.”

The research paper, “Information Consumption and Asset Pricing,” was published in the February 2021 issue of The Journal of Finance, the most widely cited academic journal on finance. Azi Ben-Rephael, associate professor of finance and economics, Bruce Carlin, professor of finance, and Zhi Da, professor of finance, coauthored the paper alongside Israelsen.

In this Q&A, Israelsen outlines the research findings and application in the business world.

Broad News: As mentioned in the paper, how information becomes incorporated into asset prices is one of the most fundamental questions in finance. How does your research help answer this question?

Israelsen: The challenge in finance lies in knowing what information investors are paying attention to — especially sophisticated investors. As such, traditional asset pricing models sort of ignore the role of information processing. They assume that information is both transmitted instantaneously and processed instantaneously by investors.

The measure that we used in this paper — expected information consumption — is a direct measure of news-reading activity on Bloomberg terminals by institutional investors. This measure is the first direct measure of information processing by institutional investors across many stocks.

Broad News: What’s the major takeaway from this research paper?

Israelsen: The rewards for owning risky stocks are realized on days when investors process information about the macro-economy and important news about stocks.

For example, we can tell from the past that every quarter, when Twitter announces its earnings, investors try to figure out what the implications are for other firms that are related to Twitter — its competitors, suppliers or other firms in its industry. Specifically, if we know that 90% of the time there is an earnings announcement for Twitter, a lot of investors are reading and looking for information about Facebook on Bloomberg, we can predict that the same thing will happen the next time Twitter announces earnings.

If I’m an investor and I know that Twitter’s earnings could be unexpectedly low, I might be hesitant to hold Facebook on those days because the market may learn bad news that has implications for Facebook’s business, too. Investors will be hesitant to hold Facebook on those risky days unless they are compensated with higher expected returns.

Broad News: Who can take action based on your findings? What further research is needed?

Israelsen: Investors already seem to be applying my research — at least, the principles that we describe. The fact that positive returns accrue on these days is evidence that investors already understand that when important news is revealed, many stocks may rise and fall. Those are particularly risky days to own stocks. Investors who do so are compensated with higher returns, at least on average. This is the type of information that hedge fund managers use when they form trading strategies.

In general, our results help explain why previous asset pricing models have failed. There has been a puzzle in the academic literature as to why traditional asset pricing models don’t do a good job at explaining stock returns. Our paper helps explain why this might be. Traditional models generally assume risk is relatively constant from day to day. We show that for most stocks, while there is little systematic risk on most days, most of the annual return accrues on a few days per year.

That said, we present results using the aggregate behavior of investors. We still don’t know exactly how a given individual investor helps facilitate the spillover of information across stocks. Future research, including what I’m currently working on, can use the type of data that will help us understand which specific investors play key roles in this process.