Thursday, August 5, 2010

The Power of Mean Reversion

Index Universe (hat tip Abnormnal Returns) details why now may be the time to allocate to U.S. equities... reversion to the mean.

There have only been four decade-long periods where U.S. equities have delivered negative returns, which were the 10 years ending in 1937, 1938, 1939 and 2008 (2009 was not included in the study). In each case, the subsequent 10-year period was strongly positive, with equities delivering (on average) an 11 percent compound annual return. Reversion to the mean. It’s simple, but it works.
Rather than look at nominal returns, I took at look at rolling ten year total returns of the S&P 500 index in real terms and then took a look at what the relationship was to ten year forward total real returns (using year end levels).

While the past is not a perfect predictor of the future, there has not been one period in which the ten year real return of the S&P 500 was negative, then followed by another decade of negative ten year real returns going back to data from the 1880's (i.e. no marker in the bottom left quadrant).

Source: Irrational Exuberance


  1. Never say never. The overvaluation in the year 2000 (that's ten years ago!) hit new records, and has not, even now, been fully corrected. So this just might be the first time we get back-to-back decades with negative returns. Or at least a rather unrewarding point near the bottom of the upper left quadrant.

  2. Jake,
    I have a total mental block when it comes to scatter graphs....I am limited, I know!

    I will have a post up later you may like, if you know what grip tape is.......

  3. Ok, think powell and peralta, santa monica....

  4. I agree with Jim. Try this on the Nikkei index.

  5. the nikkei was a different beast. whereas s&p returns never crossed 20% annualized 10 year performance in real terms, the nikkei jumped past 30% (i.e. it was EXTREMELY extended whereas the S&P just got VERY extended).

    the appropriate comparison for the nikkei would be the nasdaq rather than the s&p

  6. Like the graph and generally agree with the concept of mean reversion but can you tell me how many, if any periods of Japanese style deflation ( something which seems to be moving from possibility to probability)did the US economy experience during the period covered ?