A currency preferential approach to international equity investment

Abstract

We propose a Timex strategy for reducing the foreign exchange risk associated with international equity investment, pertaining to countries with currencies correctly or undervalued by the standard of PPP. The performance of Timex is examined from the perspectives of eight developed nations with long histories of free-floating currencies. Based on the data from 1986:Q1 to 2014:Q4, we find unambiguous evidence for the superior performance of Timex in the foreign exchange market. Compared with the passive diversification strategy and the Morgan Stanley Capital International (MSCI) World index, Timex offers higher total returns and risk-adjusted total returns when rebalanced every 6 or 12 months for investors based in all eight countries under study. When rebalanced at a 3-year interval, Timex outperforms the passive diversification and the MSCI World index for five and all eight countries, respectively.

Keywords:

• investment
• international
• foreign exchange
• purchasing power parity

JEL Classification:

• G11
• G15

Acknowledgement

The authors are grateful to the participants at the 2013 Global Finance Conference for comments and suggestions which have helped improve this article.

Notes

¹ We prefer equal weighting to value weighting to prevent the US stock market from dominating the portfolio because the US stock market makes up more than 50% of the total capitalization of the eight countries included in this study. Hence, a value-weighting scheme would result in the US market swamping the effects of other countries that have a smaller combined capitalization.

² The relative performance of our proposed strategy does not appear to be particularly sensitive to the choice of estimation period used to calculate deviation from PPP.

³ We conduct the Fisher-type ADF test (Choi, 2001) allowing for individual unit root process, and Levin, Lin and Chu (LLC) test (Levin et al., 2002) assuming common unit root for the US$ real exchange rates of seven currencies, estimated over a 5-year estimation period from 1986:Q1 to 2014:Q4. Both tests reject the null hypothesis of a unit root. The authors thank an anonymous referee for the suggestion of conducting these tests.

⁴ Our results on performance evaluation will be more conservative with comparisons made to this benchmark portfolio as Choueifaty and Coignard (2008) show that an equal-weighted index generally outperforms a value-weighted index.

⁵ The summary statistics for returns of benchmark portfolios are omitted from reporting but are available upon request from the corresponding author.

⁶ The authors thank an anonymous referee for the suggestion of conducting this robustness check.

⁷ A summary of average numbers of countries that Timex portfolios invest in is available upon request from the corresponding author.

⁸ All of the slopes are positive and the majority of them have values smaller than one. The detailed estimates for regression slopes are available upon request from the corresponding author.

⁹ The regression slopes for excess returns to Timex and Global portfolios are 0.89 and 0.83, respectively, from Japanese perspective; 0.71 and 0.67, respectively, from UK perspective; and 0.89 and 0.81, respectively, from US perspective.

¹⁰ The results for portfolio returns obtained in different market conditions are omitted from reporting but are available upon request from the corresponding author.

¹¹ Other factors that may be used for timing the currency market performance include, but not restricted to, forward rate discounts and interest rate differentials.