International journal of finance & economics vol:10 issue:1 pages:41-62
We adapt Brandt's (1999) nonparametric approach to determine the optimal portfolio choice of a risk averse foreign exchange investor who uses moving average trading signals as the information instrument for investment opportunities. Additionally, we assess the economic value of the estimated optimal trading rules based on the investor's preferences. The approach consists of a conditional generalized method of moments (GMM) applied to the conditional Euler optimality conditions. The method presents two main advantages: (i) it avoids ad hoe specifications of statistical models used to explain return predictability; and (ii) it implicitly incorporates all return moments in the investor's expected utility maximization problem. We apply the procedure to different moving average trading rules for the German mark-US dollar exchange rate for the period 1973-2001. We find that technical trading rules are partially recovered and that the estimated optimal trading rules represent a significant economic value for the investor. Copyright (C) 2005 John Wiley Sons, Ltd.