Investor attention on individual stocks can predict marketwide performance

A growing number of studies show that when investors pay close attention to individual stocks, it significantly impacts how they learn about and trade those stocks, which in turn drives stock price movements. In other words, what catches investors’ eyes directly influences their decisions and shapes the stock market. However, not all investors watch the same things, and that variability creates very different outcomes in the market. New research from the University of Notre Dame looks at how investor buzz sways the stock market, revealing that attention on individual stocks might actually predict the entire market’s short-term future performance, and noting an interesting split between retail and institutional investors. 

When everyday investors, known as retail investors, suddenly pay more attention to the market, returns over the next week tend to fall. But when professional investors such as mutual funds and hedge funds, known as institutional investors, start paying more attention, returns over the next week tend to rise — especially when big news is about to come out, according to Zhi Da, the Howard J. and Geraldine F. Korth Professor of Finance at Notre Dame’s Mendoza College of Business. Da’s research, “Market returns and a tale of two types of attention,” is forthcoming in Management Science. Da, with co-authors Jian Hua and Lin Peng from Baruch College and Tim Chih-Ching Hung from National Taiwan University, used Google’s daily search volume index to measure retail investor attention and Bloomberg’s “Daily Maximum Readership” score, which captures how often professional investors on Bloomberg are reading news about a particular stock, for institutional investor attention. Rather than looking at each stock separately, the researchers averaged the abnormal attention measures across all stocks, allowing them to create two daily, market-level attention indexes: Aggregate Retail Attention (ARA) and Aggregate Institutional Attention (AIA). They tested whether these market-level attention measures can predict future stock market returns by running regressions of market returns on ARA and AIA. “Investors have to notice a stock before they trade it, and trading is what moves prices,” Da said. “So it’s not surprising that attention can predict returns. What is surprising are two patterns we uncover.” First, the study shows that different types of investors behave distinctly. Retail investors and institutional investors show opposite effects. Retail attention predicts lower future returns, the researchers found. In other words, when everyone’s talking about a stock, it usually means it’s about to underperform. Popular stocks with lots of buzz often disappoint later because individual investors are late to the party and push prices too high initially. Once the hype dies down, prices fall back to earth, meaning lower returns for those who jumped on the bandwagon. On the other hand, the researchers saw that institutional attention often acts as an early signal for higher future returns. Institutional investors start digging into a stock before major news breaks. When their interest is piqued, it’s often a signal that some uncertainty is on the horizon. This requires investors to demand a higher return for holding that stock around the time when the actual news comes out. Second, the study reveals that looking only at market-level attention, such as searches for “Dow,” “S&P 500,” etc., doesn’t predict returns well. “The market is really just the sum of many individual stocks,” Da said. “We measure attention at the stock level and then combine all that data into a bigger picture. This bottom-up attention measure works much better than the top-down approach.” Da said identifying new factors that can forecast market ups and downs is valuable for two main reasons. “It helps us understand what actually pushes the market higher or lower, and there’s the obvious practical perk — better predictions mean better decisions when we’re deciding where to put our money.” Contact: Zhi Da, 574-631-0354, zda@nd.edu