Abstract:

This paper reviews the evidence on the profitability of mechanical trading rules based technical analysis, using selected tests to challenge the belief that successive price movements are random. This paper reports on empirical tests that mechanical trading rules can generate abnormal returns above that of buy and hold strategy for FKLI, FCPO, Nikkei Futures and Hang Seng Futures over different periods. The test methods used are extensions of those used by Brock, Lakonishok, and LeBaron (1992). This study extends to include a newly constructed mechanical trading model, BBZ, using standard deviation bands. The basic principles in this trading model are moving averages and standard deviations as advocated by technical analysis. The prices and daily returns do not result in normal Gaussian bell shape curves, suggesting that they are not random. This paper proposes that BBZ tries to capture large price movements which happen beyond 1 standard deviation from the average.  The mechanical buy signal is above + 1 standard deviation and the mechanical sell signal is below –1 standard deviation. Using daily data from 1995 to 2004 for FKLI, the net profit (after transaction costs) for BBZ is 688 index points, which is better than buy and hold strategy (loss of 90 index points).  Similarly, positive returns are found for daily data sample of 250 trading days in 2004 for FCPO (267), Hang Seng futures (1,910) and Nikkei futures (1,339).