<p>Technical trading rules have considerably been investigated empirically to determine if they are useful in predicting security prices based on historical data. Rarely have these rules been scrutinized theoretically for their efficiency. In this paper, we propose a performance measure, termed efficiency index (EI), to measure the amount of efficiency of three popular rules in an Arrow-Debreu framework. To investigate the feasibility of using the ex-ante ranking of the rules based on EIs as a predictor of their ex-post ranking based on actual returns (ARs), we use two approaches to carry out our research. With one, we use simulation to implement the three rules and determine their EIs given that the market is governed by a two-factor model estimated using in-sample data. With the other, we implement the three rules and determine their ARs using out-of-sample data. The above done, we rank the EIs and ARs of the three rules in decreasing order of magnitude and then determine if the two rankings are the same. Finally, using hypothesis tests, we show that our ranking results suggest some worthiness of EI as a viable predictor of the performance of these technical rules.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

An efficiency index as a predictor of the performance of some popular technical trading rules

  • Ka Po Kung

摘要

Technical trading rules have considerably been investigated empirically to determine if they are useful in predicting security prices based on historical data. Rarely have these rules been scrutinized theoretically for their efficiency. In this paper, we propose a performance measure, termed efficiency index (EI), to measure the amount of efficiency of three popular rules in an Arrow-Debreu framework. To investigate the feasibility of using the ex-ante ranking of the rules based on EIs as a predictor of their ex-post ranking based on actual returns (ARs), we use two approaches to carry out our research. With one, we use simulation to implement the three rules and determine their EIs given that the market is governed by a two-factor model estimated using in-sample data. With the other, we implement the three rules and determine their ARs using out-of-sample data. The above done, we rank the EIs and ARs of the three rules in decreasing order of magnitude and then determine if the two rankings are the same. Finally, using hypothesis tests, we show that our ranking results suggest some worthiness of EI as a viable predictor of the performance of these technical rules.