In a joint paper with William Goetzmann and Dasol Kim, we found that nearby earthquakes affect people’s judgment of the probability of a 1929- or 1987-size stock-market crash. If there was a substantial earthquake centering within 30 miles (48 kilometers) within the previous 30 days, respondents’ assessment of the probability of a crash was significantly higher. That is the affect heuristic at work.

It might make more sense to expect a stock-market drop from a disease epidemic than from a recent earthquake, but maybe not a crash of the magnitude seen recently. If it were widely believed that a treatment could limit the intensity of the COVID-19 pandemic to a matter of months, or even that the pandemic would last a year or two, that would suggest that the stock-market risk is not so great for a long-term investor. One could buy, hold, and wait it out.

But a contagion of financial anxiety works differently than a contagion of disease. It is fueled in part by people noticing others’ lack of confidence, reflected in price declines, and others’ emotional reaction to the declines. A negative bubble in the stock market occurs when people see prices falling, and, trying to discover why, start amplifying stories that explain the decline. Then, prices fall on subsequent days, and again and again.