Small Levels of Predictability and Large Economic Gains

Small-return predictability in the stock market has been widely documented in empirical studies, yet little has been written on its economic importance. This paper examines the issue through profitability on a trading strategy that utilizes small levels of predictability and analyzes the statistical distribution for returns achieved under such a strategy. Our results suggest that small-return predictability is economically significant in the sense that such trading strategies not only yield high returns but also are less risky under a fat tail distribution assumption. Quantitatively, we demonstrate that such a strategy could have doubled the market return for the period 1952 to 1998. We investigate reliability of our results through simulation and bootstrapping