International diversification: The evidence – Part 3

If you cast your mind back to this post (Against International Diversification), we saw some legitimate, evidence-based reasons which might support not diversifying internationally.

Specifically, we saw that cross-country correlations are high, and that correlations are likely to converge to 1 during a crash.

But are these arguments good enough to justify heavily overweighting our home country?

Let’s start by taking each of these in turn and examining the counterarguments.

Cross-country correlations are high – so what?

As a brief reminder, the idea behind this point was that countries’ stock markets are behaving more similarly. 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:

Source: Capital Group

However – one of the problems with looking at a single correlation number is that it can mask dispersions within the data. The chart above implies that because correlations between the US and ex-US are high, that markets have all been moving in the same direction.

But if we look at how the US vs ex-US has actually performed, we get a different story. From looking at the chart above, it looks like correlations rose to reach 0.8 in around the year 2000. We’d therefore expect both the US and non-US markets to perform similarly since 2000.

This is what actually happened:

Source: Vanguard

You can see that despite having a high correlation in all years since the year 2000, the US significantly underperformed from 2000 to the bottom of the 2008 crisis, then rebounded quickly to have a long stretch of outperformance to the end of 2018.

It did not perform in the same way as the rest of the world, despite having a high correlation.

Looking at stock market dispersion with an eye on the UK, it’s the same story – the UK market hasn’t performed in line with the global market due to rising correlations either. The chart below shows that the UK has had plenty of periods where it outperformed the rest of the world, and plenty of periods where it didn’t:

Source: Vanguard

Looking back longer term (all the way to 1919), the chart below shows the difference in rolling 10-year total returns for US and UK stocks. Whenever the line is above zero, US stocks outperformed their UK counterparts over the past 10 years. When the line is below the zero, the opposite occurred.

Source: CFA Institute

In the words of the author:

“While the average performance difference between the two markets over the last 100 years is zero, the 10-year return differentials can be extremely large. From June 1942 to June 1952, for example, US equities eclipsed UK stocks by 11% per year. For 10 years!… On the other hand, there are plenty of intervals when UK equities outperformed by a wide margin. During the high-inflation era from 1975 to 1985, the UK market beat the US market by 17% per year (!) thanks to its tilt toward energy and mining companies.

 That’s why the focus on correlation over dispersion is shortsighted. It ignores the material diversification benefits that can be achieved with assets that seem to move in tandem even though the steepness of the trend might vary considerably, giving rise to significant performance differences.”

Using some of my favourite quilt charts from the excellent Novel Investor, we’re able to zoom out and think about dispersion between individual countries.  

This is how varied annual returns can be, despite apparently high correlations:

Source: Novel Investor

And within Emerging Markets there’s also plenty of dispersion:

Source: Novel Investor

It doesn’t look like markets are all moving in the same direction to me…

So it seems pretty clear that markets aren’t all moving in the same direction, despite rising correlations. But let’s now assume for the sake of argument that the correlation argument is correct in what it’s implying, and that all markets are moving in the same direction.

US Investment Manager Bridgewater Associates argue that cross-country correlations are cyclical, so even if markets were moving together, they likely wouldn’t stay that way:

“The surge of globalization in the postwar era under US dominance, with rising trade and capital ties between countries globally, has led to unprecedented high correlations among the equity returns of different countries. In the past, there have been ebbs and flows in the pace of globalization including a period of rising trade tensions culminating in the world wars and of course we see rising anti-globalization sentiment resurging today.

Going forward, rising conflict around trade and globalization may increase divergences across countries. Additionally, China’s ascent as an important economic and financial center with divergent secular conditions from much of the developed world (e.g., more ability to stimulate in the event of a downturn) raises the likelihood of an increasingly multipolar and less correlated world. All of these forces raise the importance of diversification going forward.”

Source: Bridgewater Associates

They’re saying that correlations might be high now – but they might not always stay that way. And that’s a major reason to stay diversified between countries – it’s better to create a portfolio that doesn’t assume that the world will look the same in 50 years.

There are very few certainties in investing, but one of the few is that the world won’t look the same in 50 years. The whole point of creating a diversified portfolio is admitting that we don’t know what will happen, so we shouldn’t build one assuming correlations will stay high.

To summarise so far, there are two arguments why high correlations aren’t as important as they seem on the surface:

• There’s still huge dispersion in returns between countries
• Even if high correlations implied all markets were moving in the same direction, it’s unlikely that it would stay that way in the future

Markets correlate to 1 in a crash – so what?

The second evidence-based reason why people might eschew international diversification is that markets tend to all fall at the same time:

Source: AQR

The AQR paper that demonstrated the argument in the previous post also does a fantastic job of showing why this tendency for markets to crash at the same time isn’t a problem for a long-term investor.

The best place to start might be the chart from the paper below:

Source: AQR

This chart uses the same list of 22 countries from the table we’ve seen before, and plots the performance of the average worst return for the local markets (blue line), the average global return during the same period (pink line), and the average worst global return (orange line).

The first thing to notice is that all three lines start at roughly the same place. That means that over a 1-month time horizon, there’s not much benefit to holding a globally diversified portfolio vs one only holding a local index.

The second thing to notice is that the blue line declines steeply with time, and remains very negative for a very long period of time (10 years). This is evidence that the worst-case scenario for investors in a local index isn’t a short-term crash, but long-term sustained drawdown. While short-term crashes might be painful, this is the real risk.

The third thing to notice is the difference between the orange and the blue lines. Over longer horizons, the gap widens considerably, showing that the worst cases for the global portfolios are significantly better than the worst cases for the local portfolios. So in a worst-case scenario, the globally diversified portfolio performed far better than the local portfolio.

For example, at 60 months the blue line is at -57% which represents the average worst five-year loss for local portfolios across all countries. So if you invest in a single country, it’s likely that eventually you’ll experience a five year period where your real wealth is down nearly 60%. The average worst five-year global portfolio performance is about -38% (the orange line at 60 months). -38% is a world of difference from -57%, particularly considering the power of compounding. Over the long term, there’s a significant improvement in worst cases of the global portfolios vs the local portfolios.

The authors conclude by noting that:

“Investors whose planning horizon is measured in decades should not devote a great deal of anxiety to the risk of common short-term crashes. Instead, they should care more about long drawn out bear markets which can be significantly more damaging to their wealth.

While short-term common crashes can be painful, long-term returns are far more important to wealth creation and destruction. We show that over the long-term, markets do not have the same tendency to crash at the same time. This is not surprising as even though market panics can be important drivers of short-term returns, over the long-term, country specific economic performance dominates. Diversification protects investors against the adverse effects of holding concentrated positions in countries with poor long-term economic performance. Let us not diminish the benefits of this protection.”

Conclusion

• Although correlations might be high:
  1.  
  2. 1. High correlations don’t tell the whole story – there are still large dispersions in returns between countries. Heavily overweighting a local index leaves investors open to poor individual country returns.
     2. Even if high correlations implied that all markets were moving in the same direction, it’s unlikely that it would stay that way in the future. Building an internationally diversified portfolio means being prepared for the eventuality that markets will be different in the future than they are today.

• • • •  
• Although markets might crash at the same time, this only matters for investors with a short-term view (i.e. hopefully no-one). Market panics can be drivers of short-term crashes, but long-term investors should understand that country-specific economic performance is the driver of returns. International diversification protects investors against the adverse effects of holding concentrated positions in countries with poor long-term economic performance.
  1. 1.