John Bogle Had It Right: The Benefits of International Diversification are Limited

Abstract

Stock index research over long-term periods supports a U-shaped relationship of the total risk to the allocation to foreign stocks (adding foreign equities lowers risk at first then increases it). This study constructs mean–variance efficient frontiers using 28 country exchange-traded funds (ETFs) available to US investors over two periods: 1996–2018 (mostly developed countries) and 2010–2018 (many emerging markets) to determine the optimal global allocation. Over both periods, the US-only ETF had the highest Sharpe ratio. For 1996–2018, the empirical mean–variance “market” portfolio of 17 country ETFs had a 94.5% allocation to the United States, and the minimum-variance portfolio had 56.55% in the United States. Over the 2010–2018 period, the empirical tangent portfolio of 28 country ETFs had a 100% allocation to the United States, and the minimum-variance threshold for efficiency was 52.65%. For both periods, global portfolios based on GDP or market capitalization weights were highly inefficient. Home-country bias may not be a bias at all. Perhaps investors just rationally maximize the return-to-risk tradeoff.

TOPICS: Wealth management, exchange-traded funds and applications, global markets

Key Findings

• • The empirical “market portfolio” of 28 country ETFs over 2010–2018 shows a 100% allocation to the US.

• • The empirical “market portfolio” of 17 country ETFs over 1996–2018 shows a 94.5% allocation to the US.

• • Strong US home country bias mean-variance dominates cap- or GDP-weighting in building global index portfolios.