Market collapse cause depressions.

Stock-market crashes are informative about the prospects for macroeconomic depressions. Long-term data for 30 countries reveal that, conditional on a crash, the probability of a minor depression is 31 percent and of a major depression is 10 percent. The largest depressions are particularly likely to be accompanied by crashes. We allow for flexible timing between crashes and depressions to compute the covariance between stock-returns and an asset-pricing factor, which depends on the decline of consumption during a depression. With a coefficient of relative risk aversion around 3.5, this covariance accounts for the observed average (levered) equity premium of 7 percent.

The Samuelson quote is remarkable because it is simultaneously extremely clever and extremely misleading. We find, for 30 countries with long-term data, that stock-market crashes (cumulated multi-year real returns of –25 percent or less) go along with minor depressions (multi-year declines of consumption or GDP by 10 percent or more) 31 percent of the time and major depressions (declines by 25 percent or more) 10 percent of the time. In reverse, minor depressions feature stock-market crashes 71 percent of the time, whereas major depressions feature these crashes 92 percent of the time. Thus, as the Samuelson quote suggests, stock-market crashes are far more frequent than depressions. Nevertheless, a stock-market crash provides good reason for concern about the macroeconomy—because the conditional 31 percent chance of a minor depression and 10 percent chance of a major depression are far above the typical probabilities. In reverse, the absence of a stock-market crash is reassuring in the sense that a depression is highly unlikely.