Blog - Investing Notes

July 17, 2006 - Doom and the Stock Market

Is the proximity of apocalyptic doom good or bad for stocks? With geopolitical concerns in the forefront, we investigate in this entry the relationship between global doom and the U.S. stock market. There are not many direct quantitative measures of doom. One is the "Doomsday Clock" judgment, revised occasionally via the Bulletin of Atomic Scientists, which expresses the level of doom in terms of "minutes to midnight" for the human race as a metaphor for the danger of nuclear armageddon. Using the historical record for the Doomsday Clock since inception and corresponding monthly data for the Dow Jones Industrial Average (DJIA), we find that:

The following chart depicts the DJIA on a log scale and doom in units of "minutes to midnight." We reverse the minutes-to-midnight scale so that doom is high (low) when the doom graph is high (low). Doom is most imminent in the 1950s and the 1980s, and most distant in the early 1990s. The average level of monthly doom over the entire period 1947-present is 7.7 minutes to midnight, close to the current value of 7 minutes to midnight. The standard deviation of monthly doom is 4.2, so doom is volatile on the scale chosen.

There is no obvious relationship between the level of doom and stock market performance. The market does well in the 1950s and 1980s when doom is high. Both doom and the stock market stagnated in the 1960s and 1970s. The market did well in the 1990s both while doom was falling sharply and then rising. Could we be entering another period of stagnate doom and poor stock returns? Or, will an escalation in doom spur the stock market to new highs?

The Pearson correlation between these two series is a modest 0.27, suggesting weakly that higher minutes to midnight (less doom) is good for the stock market. However, the doom sample (18 indications) is so small that the correlation is not meaningful.

In summary, there is no clear connection between proximity of doom for humanity and stock market performance. Other measures of doom may yield a relationship.

For further reading, see our blog entries of 5/19/06 and 9/26/05 on randomness.