There has been a substantial interest in predictability of stock return in recent years.
For example, DeBondt and Thaler (1985, 1987) report that long-term past losers (three- to five-years) outperform long-term past winners over the subsequent three to five years. Jegadeesh (1990) and Lehmann (1990) find a short-term (one-week to one-month) return reverses.
By contrast, Jegadeesh and Titman (1993) document that over a medium-term (three- to twelve-months) horizon, firms with higher past returns continue to outperform firms with lower past returns over the same period.
In fact, investment strategies that exploit such momentum (reversal) by buying past winners (losers) and selling past losers (winners) are used by a large number of professional investors both in the U.S. and other equity markets.
You are required to perform a research to validate the effectiveness of momentum and contrarian strategies in US stock market by using 20-year testing periods starting 1996 - 2015. Specifically, you are required to perform three sets of analysis:
1) Test the effectiveness of a contrarian strategy over one-month horizon (following Panel A of Table I of Jegadeesh and Titman (1993), and having J=1-month and K=1-month).
2) Test the effectiveness of a momentum strategy over six-month horizon (following Panel A of Table I of Jegadeesh and Titman (1993), and having J=6-month and K=6-month).
3) Test the effectiveness of a contrarian strategy over three-year horizon (similar to Panel A of Table I of Jegadeesh and Titman (1993), and having J=3-year and K=3-year. But, winners and losers portfolios are formed only at the beginning of each year).
All that using SAS