You almost certainly are exaggerating the likelihood that the stock market will, in coming months, suffer a crash as bad as that of 1987.
Contrarians are smiling.
It’s entirely understandable that investors are worried about a crash, given the many analysts who are declaring the stock market to be forming a bubble. Keith McCullough, the CEO of Hedgeye Risk Management — who my colleague Jonathan Burton interviewed earlier this week — says that “there are a lot of similarities to 1987 now: the market’s quick start in January, people in love with stocks. That’s a catalyst for the stock market to crash.”
In fact, though, the probability of a 1987-magnitude crash in the next several months is very tiny — just 0.33%.
We know what the average investor believes the probability of a crash to be because of a survey that Yale University’s Robert Shiller began conducting several decades ago. He distills his results into one number: the proportion of respondents who think there is a less-than-10% probability of a 1987-magnitude crash in the subsequent six months. According to the latest survey, that number is 33.9%.
That means 66.1% of investors believe that the risk is above 10%. And that’s what is plotted in the accompanying chart. Notice that there has been a distinct uptrend in this percentage in recent years. In 2015, its 24-month moving average stood at 64%, versus 74% today — down only slightly from its high from last year of 77%.
We know the actual probability of a crash because of a study that appeared several years ago in the Quarterly Journal of Economics titled “Institutional Investors and Stock Market Volatility.” The study was conducted by Xavier Gabaix, a finance professor at Harvard University, and three scientists at Boston University’s Center for Polymer Studies: H. Eugene Stanley, Parameswaran Gopikrishnan and Vasiliki Plerou. After analyzing decades of stock-market history in the U.S. and other countries, the authors devised a formula that predicts the average frequency of stock-market crashes.
I fed into the researchers’ formula a one-day decline of 22.6% occurring in the next six months. That decline is how much the Dow Jones Industrial Average DJIA fell on Black Monday in 1987. According to the formula, the probability of such a decline is 0.33%.
Why investors have become increasingly worried about a crash
A big reason for the secular uptrend in investors’ belief in the likelihood of a crash is that the stock market suffered two bear markets in a recent two-and-a-half-year period: February-March 2020 and January-October 2022. It’s rare for two bear markets to occur in such quick succession, and that has soured many investors’ long-term outlooks.
Consider the 38 bear markets since 1900 that appear on the calendar maintained by Ned Davis Research. In only three cases did two bear markets occur in such quick succession. Two of those three instances occurred during the Great Depression, and the third took place in the early 1960s. It had been more than 60 years since two successive bear markets delivered the one-two punch that investors suffered recently.
For insight into the psychological significance of that one-two punch, I turned to a 2015 study by Camelia Kuhnen of the University of North Carolina. Kuhnen is an expert in neuroeconomics, an interdisciplinary field that brings together areas as diverse as computational biology, neuroscience, psychology and mathematical economics.
In her study, “Asymmetric Learning from Financial Information,” she reports that there is a significant difference between how investors update their beliefs following losses than how they do that following gains: Losses generate disproportionately more pessimism than gains lead to optimism. This tendency, rooted deep in how our brains work, leads to what she calls a “pessimism bias.” Not only is this bias widespread following periods with large losses, like the one we’re in today, it dissipates slowly — even when the market’s subsequent performance is quite strong.
This in turn suggests that Shiller’s crash-confidence index could be a useful contrarian indicator. I tested for that possibility on monthly data since mid-2001, which is when the survey began to be conducted each month. Sure enough, when investors were more worried about a crash, the stock market on average performed better over the subsequent one-, three- and five-year periods.
And unlike other sentiment indicators, which have only short-term significance, Shiller’s crash-confidence index has its greatest explanatory power over a several-year horizon. So while that index tells us little about where the market is headed over the next several months, it does show an increased likelihood of a strong market over the next several years.
Mark Hulbert is a regular contributor to MarketWatch. His Hulbert Ratings tracks investment newsletters that pay a flat fee to be audited. He can be reached at email@example.com.