Effects of time horizon and asset condition on the profitability of technical trading rules

Abstract

In much of the literature, the debate over technical trading strategies has centered around the question of whether an actively managed portfolio, controlled by a technical indicator, can outperform a passively managed portfolio. Typically, the time horizon is considered to be years. Additionally, the trader is assumed to use a technical trading strategy that is independent of asset conditions. These assumptions may not correspond well with reality. Traders often have much shorter time horizons and may switch between rebalancing or trading strategies on the basis of perceived shifts in market condition. This paper presents a study of the profitability of technical trading rules as a function of asset state or condition. Several common technical trading strategies were run on 296 stocks over a 15 year period. Strategies were run with 1 month rolling time horizons, significantly shorter than those used in similar studies in the literature. Stocks were segmented based on volatility and volume, which allowed for the examination of a strategy’s performance in different asset conditions. Several strategies were demonstrated to have consistently better risk-to-reward ratios under specific asset conditions and short time horizons. This finding helps to explain why some practitioners implement technical trading strategies.

Appendices

Appendix A: Mean and standard deviation of returns

Stock conditions are divided into the four categories listed below. For brevity, they are listed in the following tables in an abbreviated format. Each corresponding H or L represents a high or low for a given category. The categories are ordered as they appear below. For example, H—L—L—H represents a high volatility, low volume stock in a low volatility, high volume trading period. The numbers represent mean price return,standard deviation of price return. Green indicates that average return was greater than 1 % and orange means the average return was less than 0.

• 1. Volatility of the Stock

• 2. Volatility of the Trading Period

• 3. Volume Level of the Stock

• 4. Volume Level of the Trading Period

Fig. 6

Mean, standard deviation 1997-2004

Fig. 7

Mean, standard deviation 2005-2012

Appendix B: Pseudo sharpe ratio

The pseudo Sharpe Ratio is the mean return divided by the standard deviation of the returns. This can be thought of as a reward-risk Ratio. The higher the value, the better the reward is for the risk. This value can be raised by having a higher average return or a lower standard deviation of return. The boxes that are green have a larger pseudo Sharpe Ratio than the buy and hold strategy. This indicates that under these stock conditions, those strategies return more reward for the same amount of risk. It should be noted that this is not a statistical test and is not used to justify any particular investment, instead it is used provide a retaliative measure of risk between two strategies. Additionally, the pseudo Sharpe Ratio for all strategies are less than 1 because of the wide variance seen across stocks.

Fig. 8

Pseudo sharpe ratio 1997-2004

Fig. 9

Pseudo sharpe ratio 2005-2012

Appendix C: Tukey’s test groupings

The numbers in the cells represents groups of strategies whose returns are not statistically different. The lower the number, the higher the strategy’s returns are for a given asset condition. For example in Fig. 10, the buy and hold strategy returned a value 1, 2 when the stock condition was L—L—H—L. This means the strategy is not statistically different from strategies in both group 1 and group 2. Green indicates that the strategy , using a 95 % confidence interval, has a statistically higher return than the buy and hold strategy.

Fig. 10

Tukey’s test groupings 1997-2004

Fig. 11

Tukey’s test groupings 2005-2012