Dynamic International Diversification Strategies

Abstract

This paper offers an ex-ante optimization-based comparison among several performance measures over a range of horizons. Apart from its ex-ante nature, the novel features of such a comparison are that it is based on an aggregated performance measure, which allows for an equal start for each of its individual component performance measures and the fact that the comparison is based both on static and dynamic strategies. That is strategies, which incorporate rebalancing decisions up to a given horizon. Dynamic International Diversification Strategies (DIDS) are constructed via ex-ante maximization of aggregate performance measure over different horizons. The example DIDS, considered in the paper, represent a sequence of portfolio decisions on eight country indices (G-7 countries and Switzerland) up to a predetermined horizon from 1 to 52 weeks. Our results suggest that Information ratio based ex-ante optimization criterion dominates the others for horizons longer then 9 weeks. For shorter horizons no measure clearly dominates the others. The relative theoretical contribution of this paper is threefold. First, we use projections-based SDF approach to generate expected returns over different horizons, which to the best of our knowledge is a novel approach. Second, we propose a way to aggregate several admissible performance measures and to determine current international diversification decisions in an attempt to ex-ante optimize an aggregate performance measure. We also motivate the use of decision weights in order to determine which measure outperforms the others in ex-ante terms. An implementation of such a comparison is the main empirical contribution of the paper. Third, we systematically investigate the role of the time horizon on the ex-post realized returns of average static and dynamic international diversification strategies.