Against international diversification

Given that I’m writing this the day after the UK left the European Union, a post on the perils of investing overseas seems appropriate. What better way to stick it to those straight-banana loving mainlanders than to take all our investments and move them back to our glorious island?

Thankfully I won’t be opining on the outcome of the dreaded B-word, as 1) this is an investing blog, not a politics blog, and 2) providing a fully-informed, nuanced opinion on Brexit falls outside my circle of competence. I have a feeling that if the British public had stuck within their circles of competence, we might not be in this position now.

So let’s stick to investing.

The arguments against international diversification

In the previous post, we saw that, in theory, investing in the FTSE 100 doesn’t provide much in the way of diversification. Despite this, “home country bias” still affects UK investors, with us Brits having over 25% of our portfolios in domestic equities:

Source: Vanguard

While not a catastrophic overweight for UK investors, it’s certainly a sizeable bet on the domestic market.

Interestingly, there are a couple of evidence-based reasons that people might (and do) use to justify not being globally diversified. This sounds counter-intuitive, given how important diversification is, but the arguments can sound convincing at first glance:  

1. 1. Thanks to economies becoming increasingly inter-dependent and the world becoming much more connected than in the past, correlations between countries’ stock markets are increasing. Countries’ stock markets are behaving more similarly.
   2. Correlations between different markets correlate to 1 in a selloff. It doesn’t matter which market you’re invested in – if the market crashes, all markets crash together.

Let’s look at the evidence for each of these in turn.

Rising correlations

Taking point 1 first, data from Morningstar shows that the correlation after the mid-1990s between US and ex-US markets has increased sharply:

Source: Morningstar

In the words of the author, “if the linkage between the U.S. and other global markets gets looser, we’d see a more compelling risk-reduction benefit from diversification. And if trade wars persist and major countries pursue isolationist policies, it’s conceivable that correlations between the U.S. and other global markets could trend lower. But it seems unlikely that we’d see correlations return to the same low levels as in the past. Capital and information flow across geographic boundaries more easily now than ever before.”

The findings are corroborated from research conducted by Capital Group, who also show the rise in global correlations:

Source: Capital Group

So it seems point 1 has some validity to it.

When markets crash, they all crash together

Looking at the second point (that correlations tend towards one in a selloff), AQR show in their paper ‘International Diversification Works Eventually’, that when a local market crashes, the rest of the market tends to crash too. The table below shows the performance of each country’s local portfolio performance versus the global portfolio in the worst months for the local portfolios (January 1950–December 2008):

Source: AQR

For the United Kingdom, the worst month was October 1987, when the local return was down 26.3%. During that same month, had a UK investor held a global portfolio, it would have been down 24.6%. Across all the countries, the average worst monthly local return was –27%, whereas the global portfolios produced average monthly returns of –17.2% in the worst periods

Although the global portfolios did better than their local counterparts, they were all still down.

Further evidence comes from this Financial Analysts Journal article by Sebastien Page and Robert Panariello, who note that “during market crises, diversification across risk assets almost completely disappears. Moreover, diversification seems to work remarkably well when investors do not need it—during market rallies. This undesirable asymmetry is pervasive across markets.”

The left pane of the chart below shows the correlations of most risky assets during a ‘left-tail’ event – i.e. a selloff – versus the correlations in market rallies.

Source: Financial Analysts Journal

The chart shows how correlations of risky assets approach 1 in a selloff (left-hand pane), but have lower correlations (i.e. are better diversifiers) during market rallies (right-hand pane).

This is a heavily researched area that’s backed up by plenty of academic studies*, so it seems like the second point is also valid – correlations do tend towards 1 in a selloff.

Conclusion

It’s true that cross-country correlations are high, and also that correlations are likely to converge to 1 during a crash.

But do these two points mean that we shouldn’t diversify internationally? Should you just be putting all your cash into the FTSE?

The next series of posts looks at the other side of the coin, and examines the evidence for a globally diversified portfolio.

* For those seriously interested in seeing how correlations rise to 1 in a selloff, also see See Erb et al. (1994, 1995); Longin and Solnik (1995, 2001); Karolyi and Stulz (1996); de Santis and Gerard (1997); Bekaert et al. (1998); Ang and Bekaert (2002); Ang and Chen (2002); Chua, Kritzman, and Page (2009); Leibowitz and Bova (2009); Garcia-Feijóo, Jensen, and Johnson (2012).