A reality check on technical trading rule profits in the u,s, futures markets, страница 13

Notes. MAC, Moving Average Crossover; EMC, Exponential Moving Average Crossover; MACD, Moving Average ConvergenceDivergence; CHL, Outside Price Channel; LSO, L-S-O Price Channel; MII, M-II Price Channel; RSI, Relative Strength Index; DRI, Directional Indicator; RNQ, Range Quotient; REF, Reference Deviation; DRM, Directional Movement; ALX, Alexander’s Filter Rule; PAR, Parabolic Time/Price; DRP, Directional Parabolic; AR₅₀ denotes annual mean net returns (%) after adjustment for transaction costs of $50. p_N,W and p_N,H denote White and Hansen’s nominal p-values, respectively, which are obtained from applying their testing procedures only to the best rule or a single rule, thereby ignoring the effect of data snooping. p_RC denotes White’s Reality Check p-value. p_l, p_c, and p_u denote Hansen’s Lower, Consistent, and Upper SPA p-values, respectively. Sample periods for the mark and treasury-bills are 1985–1998 and 1985–1996, respectively.

analyses as above were conducted and similar results to those under the mean net return criterion were found.⁸

Finally, Table V shows the performance of individual technical trading systems measured by the number of the best trading rules they produce. Overall, the Moving Average Crossover (MAC), the Relative Strength Index (RSI), the Directional Indicator (DRI), Alexander’s Filter Rule (ALX), and the Directional Parabolic (DRP) systems performed the best across the performance criteria and sample periods. In particular, for each sample period the RSI system generates the best rule across various contracts, and the DRP system generates statistically significant profits for the Eurodollar even after adjustment for transaction costs and data snooping biases. This result implies that future studies on technical analysis should include the RSI and the DRP systems in analysis as well as the popular MAC and ALX systems.

⁸ The full results either with lower transaction costs or under the Sharp ratio criterion are available from the authors on request.

TABLE V

Frequency of the Highest Mean Net Return by Technical Trading System, 1985–2004

Number of Futures Markets Contracts with the Highest Mean Net Return Sample

Period MAC EMCMACDCHL LSO MII RSI DRI REF RNQDRMALX PAR DRP

1985–2004 3 0                         0              0 0                         1 3 1 0                      0                     0 3 0 6

1985–1994 2 0                         1              0 1                         1 6 0 0                      0                     0 1 1 4

1995–2004 2 1                         1              1 0                         0 7 0 0                      0                     1 3 0 1

Notes. MAC, Moving Average Crossover; EMC, Exponential Moving Average Crossover; MACD, Moving Average ConvergenceDivergence; CHL, Outside Price Channel; LSO, L-S-0 Price Channel; MII, M-II Price Channel; RSI, Relative Strength Index; DRI, Directional Indicator; RNQ, Range Quotient; REF, Reference Deviation; DRM, Directional Movement; ALX, Alexander’s Filter Rule; PAR, Parabolic Time/Price; DRP, Directional Parabolic.

CONCLUSIONS

The empirical evidence in this article indicates that technical trading rules in general have not been profitable in the U.S. futures markets over the 1985–2004 period after correcting for transaction costs and data snooping biases. In particular, the successful performance of Lukac et al.’s (1988) trading rules over the 1978–1984 period does not persist in the subsequent l985–2004 period. It suggests that the performance of technical trading strategies have become unstable over time since technical traders would experience large losses in most futures markets if they followed typical technical rules dealt with in most early studies.