Assessing an Allocation to Global Stocks
You know your client’s portfolio is properly diversified when there is always a portion of it you hate. Right now, that hateful piece of the equity allocation is global stocks.
The performance has been particularly frustrating for investors who compare performance with a domestic benchmark, such as the S&P 500 Index. It’s tempting to criticize a strategy based on recent performance, but using global equities as part of a diversified portfolio still makes sense.
How Fast We Forget
It wasn’t that long ago that everyone seemed to hate US stocks and wanted to increase their allocation of global equities (particularly emerging markets). Investors who chased performance and made significant shifts in their allocations following the global equity outperformance have recently been burned.
Diversification doesn’t work when you deviate from the client’s long-term investment plan. That makes it all the more important to recognize that periods of over- and underperformance are the norm when investing in a globally diversified portfolio.
Different Decade, Different Winners
The dataset for the MSCI World ex-US Index goes back to 1970. If you break out annualized returns by decade, you can see that periods of outperformance and underperformance for global equities have always been part of the deal.
Equally important from a diversification standpoint is that these two asset classes experience different levels of volatility, which you can see in the following chart, which breaks out annualized standard deviation by decade.
Because the point of diversification is to combine assets with different levels of return and volatility, we would argue that global equity exposure has been working for long-term investors.
Measuring Historical Success
Building an investment portfolio to meet your client’s long-term goals is all about trying to earn the highest return for a given level of risk. That requires combining investments that zig with others that zag in order to achieve the desired levels of risk and return in your client’s portfolio.
One way to measure the risk–return trade-off is with the Sharpe ratio. The following table compares US and global stocks as well as a diversified portfolio of 75% US stocks and 25% global stocks.
The first thing to notice is that the diversified portfolio has the lowest standard deviation, which means that it is less volatile than investing in either US stocks or global stocks alone. Although the returns of the diversified portfolio are slightly less than those of investing in US stocks, the lower volatility results in a better (higher) Sharpe ratio.
The diversified portfolio’s higher Sharpe ratio means that it earns a higher risk-adjusted return; so, if the goal is to earn the highest level of return at a given level of risk, investors should prefer the diversified portfolio.
Details in Data
There are a few shortcomings to the Sharpe ratio analysis shown previously. First, the diversified portfolio increases the Sharpe ratio by only 5%. That’s hardly an awe-inspiring improvement. Furthermore, we can’t expect future returns and volatility to perfectly mimic the past. There is always a small margin of error.
Second, the sample used in the previous analysis pits global equities against an extremely impressive period of US equity returns. Reliable data on US stock returns go back to 1926, but we are capturing only the past 44 years because the global equity data doesn’t begin until 1970 (and the inclusion of emerging market data within this index doesn’t begin until 1990). Within these past 44 years are six of the seven longest bull markets in US history (see the following table).
If we compare the sample period with the entire dataset for US markets, we find that the average return is lower and average volatility is higher than in the 1970–present dataset.
We couldn’t speculate on how the historical performance comparison may have differed if we hadn’t picked up such a positive period for US stocks. We can, however, measure the diversification benefit of global equities by looking at periods in which US stocks had a negative monthly return.
Since 1970, there were 207 months in which US stocks were down. Although global stocks tended to fall too, they only had a correlation of 0.62 and provided higher average returns than the US stocks. Again, this is what diversification is all about: combining assets that don’t move in lockstep in order to have a less volatile portfolio.
Although recent performance may suggest otherwise, investing in global equities as part of a diversified portfolio has historically provided some modest benefit to long-term investors via a small improvement in risk-adjusted returns and superior performance during down markets in the United States.
Investors considering reducing their allocation to global equities because of recent underperformance may want to consider the perils of chasing performance. It is extremely improbable that someone could accurately predict the best-performing asset class year in and year out.
If the diversification argument for global equities isn’t strong enough for you, then perhaps you ought to consider the tendency of global stocks to outperform in the five years following an annual loss.
Global equities currently have a negative annual return, which history says could be a precursor to outperformance in the next five years. Although this doesn’t always hold true, global stocks have prevailed in eight of the last 11 such instances.
There are many uncertainties facing global equity investors at the present moment, but it is those uncertainties that provide for higher expected returns. As Warren Buffett famously said, “Be fearful when others are greedy, and be greedy when others are fearful.”
Over time, global investing has improved diversification for disciplined investors, and we believe allocations should be maintained despite recent underperformance and the troubling outlook.
If you liked this post, don’t forget to subscribe to the Enterprising Investor.
All posts are the opinion of the author. As such, they should not be construed as investment advice, nor do the opinions expressed necessarily reflect the views of CFA Institute or the author’s employer.
Image credit: ©iStockPhoto.com/Meriel Jane Waissman
16 thoughts on “Assessing an Allocation to Global Stocks”
Excellent article, Peter. It’s very, very important to diversify using asset classes that provide not only low correlation with what’s already in the portfolio but also high risk-adjusted returns on their own. Most people seem to get the first part of that (although not necessarily why), but often they forget to pay attention to the second part: after all, money stashed under the mattress has a correlation of zero with every other asset class, but it doesn’t help your portfolio maintain good returns, even on a risk-adjusted basis.
The rule is this: if the Sharpe ratio of any new asset class is greater than the Sharpe ratio of the existing portfolio multiplied by the correlation between the existing portfolio and the new asset, then adding the new asset will improve the risk-adjusted returns of the portfolio.
I use a simple graph to help visualize the diversification benefits of different asset classes, and it’s posted at http://www.slideshare.net/casebrad/asset-class-diversification-is-more-important-than-global-diversification. On the vertical axis I show historical risk-adjusted returns (Sharpe ratio), and on the horizontal axis I show correlation with the broad U.S. stock market–making the assumption that the foundation of your portfolio is your allocation to U.S. equities. What the graph shows is that, while diversification into non-U.S. stock markets has benefited investors through low correlations, diversification into the real estate asset class–even domestically, through U.S. REITs–has provided BOTH low correlation AND strong historical risk-adjusted returns.
Of course, as you point out, the recent past (my analysis covers nearly 25 years) has been especially good for U.S. stocks and especially bad for non-U.S. stocks, and the future may be different. And of course my analysis doesn’t mean you shouldn’t have BOTH asset class diversification AND global diversification within a given asset class–in general, it seems likely that investors would benefit from both. But my analysis does suggest that asset class diversification (adding REITs to your stocks) is more important than geographic diversification (adding non-U.S. stocks to your U.S. stocks).
Thanks for posting such a good article.
I never thought of much of what you wrote despite having casually invested for years. This will definitely change the way I look at my portfolio.
As I was reading through your article I noticed there was a flaw in the Sharpe Ratio calculations that you used for time periods of 1970-Present. In you table titled Annualized Total Return and Standard Deviation (1970-Present) in the section Measuring Historical Success you use a Risk-free rate of 0% when the Risk-free rate over this time period was closer to 5%, which would make these Sharpe Ratios closer to half what you present in your article. Also the difference between the United States and the Diversified portfolio shrinks by 0.02, leading to a Sharpe Ratio difference of 0.01. If you could please explain why you choose to do this it would be greatly appreciated.
Great catch, Nick. I’m embarrassed that I didn’t notice it myself, because I’m tremendously bothered when I see people failing to subtract the risk-free rate of return but calling what they did a “Sharpe ratio.”
I very much doubt that Peter meant to mislead anybody, but I’ve seen many, many marketing materials for low-return assets (especially private real estate, and also hedge funds) in which the same mistake is made, and in many of those cases I expect it’s very much intentional, because subtracting the risk-free rate from the average (NET total) return of a low-return asset tends to make it look like a very poor investment.
Thanks for pointing it out.
Thanks for commenting, and thank you for catching the error. You are absolutely correct about the missing risk free return. In developing the table, my spreadsheet equation missed a cell that would have subtracted 5.09% from the annualized returns of the three portfolios (US, Global, Diversified). As a result the Sharpe Ratios should be: US = 0.342, Global = 0.246; Diversified = 0.353.
This mistake, however, does not change any of my conclusions about the merits of a global equity allocation. The originally published numbers reflect a 4.54% improvement in Sharpe Ratio, whereas using the correct data shows a 3.27% improvement in Sharpe Ratio.
As I state within the article, the modest improvement in Sharpe Ratio is not enough to justify global diversification, but the argument is strengthened by: (1) different levels of return and volatility, (2) performance and correlations during periods of negative US stock performance, (3) the historical tendency of global stocks to outperform following periods of negative returns.
My apologies to everyone for the error and many thanks to Nick for catching it!
It’s funny to see how our friends across the ocean see things differently. As an Australian, the current talk in the industry is how Australian shares have been underperforming the past few years and how we should be increasing client allocations to unhedged overseas shares. Personally I think that (as usual) the prevailing market opinion is behind the game here – I was into unhedged international shares when our dollar was worth nearly $1.10US, and am now starting to think about hedging again now that the AUD/USD exchange rate seems closer to fair value.
I had a question on the final table, which shows World Ex U.S. returns, then the following five year annualized return. That very last row, for 2008, the five year annualized returns must include 2008 as the first year, correct? There is no way that the U.S. has only returned 2% annualized starting in 2009.
You are absolutely correct and it appears that I made an input mistake somewhere along the way. The return for the U.S. in the five years following 2008 was 19.13% (annualized), which increases the average annualized 5-year return for U.S. stocks to 12.55% from 11.03%.
And is the international data for that row also correct?
Wow, how embarrassing! Yes, I did have the incorrect World number as well. The annualized 5-year return is 12.49%, not -3.43%. That raises the average for the World ex-U.S. to 15.65% from 12.49%.
Thank you for the careful review and correcting comment!