Testing the Predictive Power of Technical Analysis’ Trading Rules Using Modified Box-Tiao Time Series Models

Abstract

In the battery of existing statistical techniques for the testing of the predictive power of trading rules of technical analysis and, in extension, for the testing of the hypothesis of efficiency in financial markets, this work proposes one more for inclusion. The proposed technique is a modification of the Box-Tiao stochastic models for the assessment of the impact of external events on time series. The potential advantages of the proposed methodology over the existing ones (e.g. simple t -tests, bootstrapping techniques) are discussed. Application of the proposed statistical technique to the returns of the FT30 Index of the London Stock Exchange, in conjunction with the moving average technical trading rule, shows good agreement with the empirical findings of other methods.