A New Global Portfolio Weighting Strategy Based on Cointegration Methods

Abstract

The author applies cointegration methods to separate desirable segmented equity markets from redundant cointegrated ones. The cointegration framework is robust to the intertemporal correlation instability that plagues modern portfolio theory (MPT)–based portfolios and identifies segmented market portfolios that consistently outperform both MPT and cointegrated counterparts over the 1995–2014 test period, based on 23 developed countries. The author proposes a new allocation strategy that more heavily weights countries with smaller likelihood ratios and that captures the degree of market segmentation for each country. This cointegration-based portfolio has fewer risks and performs better, particularly in declining market conditions, when diversification is most needed. This article provides a novel global portfolio management framework for country selection and allocation.