A reality check on technical trading rule profits in the u,s, futures markets, страница 14

In a recent study, Roberts (2005) also found that trading rules formulated by genetic programming produced statistically significant returns only for 2 of 24 futures markets traded in the U.S. exchanges over 1980–2000. In addition, the results of this article are largely consistent with those in Sullivan et al. (1999, 2003), Olson (2004), and Neely, Weller, and Ulrich (2009) who documented that technical trading profits disappeared or decreased after the mid1980s in the stock market and foreign exchange market, respectively.

There are three possible explanations for the disappearance of technical trading profits in the 1985–2004 period: (1) data snooping biases (or selection bias) in previous studies, (2) structural changes in futures markets, and (3) the inherently self-destructive nature of technical trading strategies. First, there exists a possibility that Lukac et al.’s (1988) successful finding might result from examination of a relatively short and profitable sample period by chance. As noted previously, data snooping problems can occur by searching for profitable in- and out-of-sample periods, trading systems, and trading model assumptions, as well as profitable trading rules.

As another explanation, Kidd and Brorsen (2004) report that returns to managed futures funds and commodity trading advisors (CTAs), which predominantly use technical analysis, declined dramatically in the 1990s. The decrease in technical trading profits could have been caused by structural changes in markets, such as reduced price volatility and increased kurtosis of daily price returns occurring while markets are closed. Since technical trading strategies make profits by the process of a market shifting to a new equilibrium, there may be fewer opportunities for profitable trading if prices are not as volatile.

Finally, forecasting methods are likely to be self-destructive (Malkiel, 2003; Schwert, 2003; Timmermann & Granger, 2004). New forecasting models may produce economic profits when first introduced. Once these models become popular in the industry, however, their information is likely to be impounded in prices, and thus their initial profitability may continue to deteriorate. This may explain why CTA returns in the 1990s decreased with the growing popularity of technical analysis. Schwert (2003) also finds that a wide variety of market anomalies in the stock market, such as the size effect and value effect, tend to have disappeared after the academic papers that made them famous were published.

Findings in this article come with a caveat. Unlike the approach adopted in most academic studies, technical traders in practice tend to periodically utilize back testing in which trading strategies are optimized over different amounts of historical data and then applied to the real markets. Technical traders would seek more profits by either adapting their trading strategies over time or developing more sophisticated trading strategies (Menkhoff & Taylor, 2007; Neely et al., 2009). To approximate the profitability of technical trading strategies more accurately, future research should increase efforts to effectively incorporate market practitioners’ approach into the trading model.

Lastly, these findings and conclusions contribute to the ongoing debate about the usefulness of technical trading strategies between market participants and academics and between competing theories (the efficient markets hypothesis vs. behavioral models). The results of this article suggest that unlike the behavioral finance view, technical traders are not likely to survive in the long run. While cognitive misperceptions (or psychological biases) may plague many individual market participants as widespread evidence indicates (see Barberis & Thaler, 2003), such misperceptions may not be an important determinant of aggregate market behavior (Merton, 1987). In other words, a relatively small number of “smart money” professional investors and traders are not affected or less affected by cognitive biases and thereby: (1) earn profits from the numerous “uninformed” individual investors and traders and (2) force market prices to efficient levels.