Dynamic Asset Allocation Using a Combined Criteria Decision System

In this paper we examine the predictability of asset returns by developing an approach that combines quantitative methods of forecasting, based on technical analysis. As an innovation we introduce a multiple criteria decision system making simultaneous use of trend indicators and other confirming indicators. By combining trend indicators with confirming indicators it is possible to build a superior technical trading strategy that captures a more comprehensive aspect of predictability in past prices. This study also proposes a test for weak form efficiency based on a combining approach. Previous approaches typically make inferences based on the empirical results of testing only one class of technical rules. Applying the combining criteria decision system the evidence suggests that the strategies proposed here have predictive ability on a data sample based on three European stocks Index Markets. Our results rejects the null hypothesis that the returns earned from applying trading rules are equal to those achieved from a naive buy and hold strategy, even after deducting transaction costs. Evidence also suggests that oscillators capture some aspect of predictability in past prices that moving averages do not detect.