>> NBER study on Recessions & Depressions

---- Abstract -----

Long-term data for 25 countries up to 2006 reveal 195 stock-market crashes (multi-year real returns of -25% or less) and 84 depressions (multi-year macroeconomic declines of 10% or more), with 58 of the cases matched by timing. The United States has two of the matched events--the Great Depression 1929-33 and the post-WWI years 1917-21, likely driven by the Great Influenza Epidemic. 45% of the matched cases are associated with war, and the two world wars are prominent. Conditional on a stock-market crash, the probability of a minor depression (macroeconomic decline of at least 10%) is 30% and of a major depression (at least 25%) is 11%. In a non-war environment, these probabilities are lower but still substantial--20% for a minor depression and 3% for a major depression. Thus, the stock-market crashes of 2008-09 in the United States and other countries provide ample reason for concern about depression. In reverse, the probability of a stock-market crash is 69%, conditional on a depression of 10% or more, and 91% for 25% or more. Thus, the largest depressions are particularly likely to be accompanied by stock-market crashes, and this finding applies equally to non-war and war events. We allow for flexible timing between stock-market crashes and depressions for the 58 matched cases to compute the covariance between stock returns and an asset-pricing factor, which depends on the proportionate decline of consumption during a depression. If we assume a coefficient of relative risk aversion around 3.5, this covariance is large enough to account in a familiar looking asset-pricing formula for the observed average (levered) equity premium of 7% per year. This finding complements previous analysis that were based on the probability and size distribution of macroeconomic disasters but did not consider explicitly the covariance between macroeconomic declines and stock returns.

Emphasis mine. Continue reading here