Until fairly recently the conventional wisdom in the finance academic community was that security prices follow a random walk. Some influential papers have uncovered evidence of mean reversion particularly over longer horizons. Siegel (1998) has suggested that given this evidence: "the holding period becomes a crucial issue when the data reveal the mean reversion of the stock returns." In this paper, we explore the pattern of mean reversion in post-World War ll U.S. stock returns and find that it peaks in a 4-year cycle. Given this empirical regularity we show that a buy-and-hold investment strategy, which is appropriate under a random walk, is no longer optimal.
Available at: http://works.bepress.com/sjaggia/13/