Correlation and Causation

Jul 2014

If we told you there was a tight correlation between U.S. GDP and the kangaroo population in Australia, you’d probably come to the same conclusion.  However, at times specious correlations are not so obvious, because there seems to be an argument to be made that a cause-and-effect relationship exists—even if it doesn’t. F or example, just because the S&P 500 gained (or declined) more often than not when a given economic condition existed in the past, it is risky to draw the conclusion that the economic condition will again be followed by similar moves in the S&P 500 in the future.  Perhaps, for example, it’s not the specific economic condition that led to stock market moves, but rather some other variable that accompanied the condition in question in the past, but won’t necessarily do so now or in the future.  Another example of misleading coincidences would be the seeming similarity of stock price patterns.  For instance, after the sharp stock market downturn in October of 1987, Alan Abelson, the former editor of Barron’s and perennial pessimist, delighted in comparing a chart of stock prices in 1987 to a similar chart of 1929 stock prices.  Visually, the correlation appeared close.  Until it didn’t.  The aftermath of 1929 was the Great Depression, yet 1987 represented just a  temporary interruption to a long period of economic growth.  Forgetful of history, other pundits tried to compare stock prices in 2008 with those in 1929.  That didn’t work either.  For a while the correlation may have seemed close, but stock prices after 2008 bore no resemblance to stock prices after 1929.  Understanding when correlation does result from causation (and when it doesn’t) is one thing that separates successful and unsuccessful investors.