Economic Narratives Shape How Investors Perceive Risks

Narratives Inform Beliefs and Explain Predictive Changes in Market Volatility

Newspapers are important vehicles for the spread of ideas. When large shocks to the financial markets occur, historical references often play a role in news stories—and the way newspapers present market-related facts may affect how readers interpret and contextualize them. In a recent working paper, “Crash Narratives,” OFR Research Principal Dasol Kim and Yale economists William Goetzmann and Robert Shiller explore the role of economic narratives in shaping investor beliefs about stock market crashes.

Hypothesis

The authors theorize that investors use narratives to inform their beliefs and, in turn, their choices. Narratives often elicit causal relationships between a sequence of events connecting one state to another and things that happen in between. Narratives thus suggest an analogy for how and why things are happening and what the future outcome might be. Investors may be especially influenced by narratives related to rare events—such as stock market crashes—compared with narratives related to common (market)events. The psychology literature provides evidence that individuals are susceptible to a variety of behavioral biases when they evaluate low-probability events. Investors may seek a framework to contextualize rare events, and narratives in the media may offer that framework.

Methodology and Findings

The authors develop a measure of news narrativity specifically related to major stock market crashes. Detecting narrativity is generally complicated; it requires researchers to identify highly complicated relationships among unstructured data. However, recent advances in computational linguistics allow researchers to do just that. The authors use this advanced technology to extract narratives about stock market crashes—or crash narratives—that appeared in the news in the days following the 1987 and 1929 crashes. The authors also develop methodologies that allow them to estimate the prevalence of such narratives in news articles. The authors show that this measure of crash narrative prevalence correspond with major market-related events over the past 30 years, including the Long-Term Capital Management collapse, the September 11 terrorist attacks, the Great Financial Crisis, the European Debt Crisis, and the COVID-19 pandemic. This suggests that contemporary journalists broadly employ narratives that earlier journalists used following the 1987 crash, in spite of the differences in historical contexts. The crash narrative measure also has a significant relationship with market-based crash indicators in the post-1987 period. Manual inspection suggests that these patterns are unlikely to be driven by the prevalence of specific words alone, but rather by the sequencing and distribution of words representing narrative structure.

Crash Narratives That Appear the Previous Day Have a Positive and Significant Association with Crash Attention

The first set of tests provides suggestive evidence that the prevalence of crash narratives in the media explains a significant amount of predictive variation in investor attention to stock market crashes and market volatility. Crash narratives that appear in the financial press have a positive and significant association with public attention to stock market crashes, as indicated in internet search volumes. These results suggest the broad propagation among the general public of crash narratives from the financial press. They also suggest that crash narratives prime investor crash concerns and, in turn, market feedback—meaning that narratives have a positive and significant association with future market volatility, as measured by the Chicago Board Options Exchange Volatility Index (VIX).

Crash Narratives Directly Affect Investor Crash Belief

The authors next use investor survey data to “look under the hood” at the effects of narratives on investor beliefs. They find evidence that crash narratives directly affect investor crash beliefs. Investor crash belief probabilities are elevated following the appearance of crash narratives the previous day. The authors also find evidence that when investors respond to survey questions asking them to describe general stock market conditions, they repeat language found in newspaper articles related to crash narratives.

Journalists Are More Likely to Use Narratives During Periods of High Public Attention to the Stock Market

In the final set of tests, the authors examine periods when journalists are more likely to employ general narrativity. For these tests, the authors construct a pure measure of narrativity based on classical folktales published almost two centuries prior to the analysis. The authors find a strong correspondence between the crash narrative measure and the pure narrativity measure, affirming the interpretation that the crash narrative measure is indeed capturing narrativity rather than fundamental factors. The authors also show that journalists are more likely to employ general narratives during periods when the general public is paying more attention to the stock market. This is consistent with the view that journalists may use narratives to make the news more accessible and are therefore more likely to use narratives during periods when readership of the financial press is more diverse. Finally, the authors show that the pure narrativity measure has a similar effect on investor crash beliefs.

The role of the financial press as a mechanism for propagating economic narratives is potentially important. News articles make comparisons to past catastrophes to make salient the gravity of current events and focus public attention on a singular narrative about what the future may bring. The paper provides evidence that narratives about stock market crashes propagate broadly once they appear in news articles, and it shows strong effects on investor belief formation.