Permanent and Temporary Components of Stock Prices
Read the paperopens doi.org in a new tab
What they found
Fama and French regressed multi-year stock returns on the previous multi-year return for U.S. portfolios from 1926 to 1985. Autocorrelations were near zero for short horizons but became strongly negative at three to five years, reaching around minus 0.25 to minus 0.4, and then faded at longer horizons, which is the signature of a slowly mean-reverting temporary component in prices. Up to 40% of the variance of three to five year returns was predictable from past returns, though the effect was much weaker after 1940.
What you can use
- Even the fathers of efficient markets documented meaningful long-horizon mean reversion in stock prices.
- Mean reversion at multi-year horizons can reflect changing expected returns (risk premia), not necessarily mispricing.
- The effect was much weaker after 1940, which is a warning about how unstable long-horizon predictability is.
Caveats
The post-1940 weakening and the small number of independent long-horizon observations make the result statistically fragile; the authors say so. Not a trading strategy.
Tags: mean-reversion, long-horizon, expected-returns
Summaries are our own reading of the paper, not the authors' words. Educational only, not advice. Discuss it in Book Club.