Mindful investing involves buying a moderately diversified set of low-cost stock index funds and holding them for the long term. In past posts, I’ve provided many reasons why this investing style makes sense. But I’ve been purposefully vague about what qualifies as “moderately diversified”. In this post, I’ll explore some of the nuances and difficulties involved with stock diversification.
Super-Simple Stock Portfolios
Let’s first consider super-simple stock portfolios such as holding a single U.S. all-stock-market fund. Or even simpler, you could hold a single S&P 500 index fund, which focuses on just U.S. large-cap stocks. In my last post, I compared these two super-simple stock portfolios to several more diversified portfolios over two example periods. In both periods the super-simple portfolios competed quite well against more diversified stock portfolios in terms of total annualized returns.
Further, I’m not the only one who’s noticed that less diversified stock portfolios have often achieved better historical returns and sometimes lower volatility. America’s favorite investing uncle, Warren Buffett, has said for years that most individual investors should simply buy a low-cost S&P 500 index fund and hold it for the long term. Similarly, the creator of index funds, Jack Bogle, suggested holding a combination of stocks and bonds, but for the stock portion, he advocated simply holding the equivalent of a U.S. all-stock-market index fund.
To further illustrate the merits of super-simple stock portfolios, I plugged in the U.S. large-cap and U.S. all-market stock portfolios into Portfolio Visualizer’s asset allocation tool and compared them to one flavor of a more diversified stock portfolio consisting of:
- 20% U.S. Large-cap Stocks
- 20% U.S. Mid-cap Value Stocks
- 20% U.S. Small-cap Value Stocks
- 20% Developed Market Stocks (no U.S. stocks included)
- 20% Emerging Market Stocks¹.
I compared the successive 30-year annualized returns for these three portfolios starting with 1972 to 2002 and ending with August 1989 to August 2019. This graph shows 30-year annualized returns of the more “diversified portfolio” minus the annualized returns of the U.S. all-stock and U.S. large-cap portfolios. A negative value on the vertical axis indicates that the diversified portfolio had worse performance than the super-simple portfolios, and a positive value indicates the diverse portfolio had better performance.
The super-simple portfolios outperformed the diversified portfolio in 8 out of 18 (44%) of the 30-year periods examined, nearly half the time. This might spur you to devise a different diversified portfolio that performs better against the super-simple portfolios. But in my last post, I showed that trying to optimize a portfolio through this sort of “back-testing” won’t ensure future superior performance, because the future never copies the back-test periods exactly.
For this reason, stock diversification is not about picking the “best” portfolio for all possible futures. Instead, diversifying your stock holdings helps guard against more generic uncertainties about the future. We don’t know what’s going to happen next, which means we want a portfolio with a reasonable chance of weathering future stock market storms that may blow in from any direction. So, we’re better off focusing less on the specific components used to diversify and more on the general level of diversification in any given portfolio.
S&P 500 Diversification
Let’s start by examining the general level of diversification in the simplest stock portfolio consisting only of the S&P 500. We can use this to define one end of the stock diversification spectrum. You might assume that an S&P 500 portfolio is extremely undiversified. But the 500 companies included in this index have many different characteristics.
Size – The “large-cap” S&P 500 index contains other sizes of stocks as shown in this chart.
Only 34% of the capitalization comes from large-cap stocks and about 10% comes from mid-cap stocks. That’s not a huge amount of size diversification, but it’s not zero either.
Sector – The S&P 500 also encompasses many different business sectors as shown in this May 2018 pie chart.
Geography – All the companies in the S&P 500 are domiciled in the U.S., but because most of them operate globally, that tells you very little about the geographic exposure of the index. This chart shows the geographic sources of aggregate revenue in the S&P 500.
And this chart shows the geographic sources of aggregate sales in the S&P 500.
It turns out that almost half of the sales within this U.S. index originate in other countries.
Style – Probably the most popular style differentiator for stocks is growth versus value. These style categories can be defined in different ways, but the ratio of price-to-book (P/B) value is one common measure, with value stocks generally having lower P/B values. Here are the P/B values for Vanguard’s S&P 500 index fund as compared to the S&P 500 Value and Growth funds.
- S&P 500 – 3.1
- S&P 500 Value – 2.2
- S&P 500 Growth – 5.2
These P/B values suggest that the S&P 500 index fund is a blend of value and growth companies with a tilt toward value.
U.S. All-Market Diversification
Expanding your portfolio from just the S&P 500 and into the entire U.S. stock market produces slightly better diversification in these same categories (using the same data sources as above):
- Size – An all-market portfolio will expand size exposures to 6.5% for small-cap and 17% for mid-cap (almost 25% combined).
- Sector – The all-market portfolio covers the same sectors as the S&P 500 but with slightly more balance. For example, technology stocks represent 26% of the S&P 500 but only 23% of the entire U.S. stock market.
- Geography – Similarly, the geographic revenue sources are slightly more balanced using a U.S. all-market fund with 35% of revenue coming from non-U.S. sources, as compared to 29% for the S&P 500.
- Style – The P/B of the U.S. all-market fund VTI is 3.0, nearly identical to the S&P 500 fund VOO.
So, both these super-simple portfolios already contain way more diversification than you might have originally assumed. Nonetheless, we can easily increase the general level of diversification beyond a super-simple portfolio without trying to pick the “best” variation on the theme.
More Geographic Diversification
One could argue for more diversification in various additional style categories, but many of these categories fall into the realm of so-called factor investing. I’ve argued that factor investing is not particularly mindful because it relies on the long-term continuation of some pretty fragile and short-lived historical stock patterns. That is, factor diversifiers of the past won’t necessarily diversify much in the future.
What about the potential to diversify more in terms of size² and geography? I hope to address size diversification more in a future post. But for now, I’ll simply note that the nearly 25% exposure to small and mid-caps provided by an all-market fund already provides a fair bit of size diversification. That leaves the question of more geographic diversification, which I’ve suggested in the past is an easy way to further diversify a mindful investment portfolio.
What’s so special about geographic diversification? Buffett and Bogle would say there’s nothing special about it at all because they both have consistently advised individual investors to stick with U.S. stocks. However, look in the rearview mirror at the 20th-century world that Buffett and Bogle invested in for most of their lives. Now compare that to the future world we can see coming at us through the windshield, which will probably resemble this chart from Visual Capitalist.
Right now the Gross Domestic Product (GDP) of China and India combined is a little over 1.5 times the U.S. GDP. But in about a decade China and India combined will have 3.5 times the GDP of the U.S. and will represent more than a third of the global GDP. If you focus only on U.S. stocks over the next decade, you’ll be putting all your eggs in an ever-dwindling sub-basket of the global economy.
I often warn that it’s impossible to predict the future. These predicted GDP trends likely don’t anticipate everything that might happen over the next decade. Perhaps American technology and know-how, combined with a wave of immigration reform will produce an incredible new surge of economic productivity in the U.S. But it seems much more reasonable to assume that China and India’s future absolute GDP growth will outpace U.S. growth rather than the opposite.
Potential Outcomes of Geographic Diversification
What could you gain from greater geographic diversification in your portfolio? Since we can’t predict the future, one of the few ways to answer this question is to use the admittedly limited tool of back-testing. Here’s a chart like the first one in this post, but using successive 10-year periods from 1995 to present because reliable emerging market data don’t go back any further than this. The difference between U.S. stock all-market annualized returns and both developed markets (without the U.S. included) and emerging markets are shown.
The results are very reminiscent of the see-saw performance in my opening example, which is pretty typical when comparing any asset classes. Periods of outperformance by one asset evolve into periods of underperformance and vice versa.³
A geographically diversified portfolio would hold a combination of these assets. Accordingly, let’s add a test portfolio consisting of 60% U.S. all-market, 25% developed markets (without the U.S.), and 15% emerging markets. The world’s stock market capitalization is around 40% U.S., 35% developed markets (without the U.S.), and 25% emerging markets. So, the test portfolio is still somewhat U.S.-centric but not extremely so. This graph adds our geographically diversified test portfolio to the comparisons with the super-simple U.S. all-market portfolio.
The annualized returns for the geographically diverse portfolio diverge much less from the U.S. all-market portfolio with no more than a few percent of outperformance or underperformance in any given 10-year period. And in case you’re wondering, here are the annualized returns for these portfolios over the entire period from 1995 to August 2019:
- U.S. all-market – 9.81%
- Developed markets (without the U.S.) – 4.63%
- Emerging markets – 5.79%
- Geographically diverse – 8.45%
We can’t know in advance that more geographic diversification will necessarily give us better returns. We can only say that moderate geographic diversification will often diffuse the impacts from potential future market events, and as a result, such diversification may smooth out the variations in our returns over the years.
Conclusions
I’ve presented the geographically diverse 60%/25%/15% portfolio before, but not because there is anything optimal or magical about this particular diversification scheme. It merely represents one reasonable example of further diversifying a super-simple portfolio.
In future posts, I hope to develop, back-test, and compare some additional stock diversification schemes. I’m particularly interested to see how a portfolio more focused on size diversification stacks up. But because of the typical see-saw performance of most stock types, my additional back-testing won’t suddenly reveal a portfolio that will perform wildly better or worse than today’s example. Rather, these exercises simply reinforce the idea that the specific combinations of stock types used to diversify are relatively unimportant, exactly because we can’t predict the future.
Some moderate additional diversification beyond a super-simple and U.S.-centric stock portfolio is a reasonable way to address fundamental uncertainty about the future. As I’ve noted before, diversification often means that you’ll sacrifice some returns as compared to the best performing asset in any given period. But moderate stock diversification of almost any kind relieves you of the impossible task of trying to pick the best-performing stock types in advance.
1 – Portfolio Visualizer emerging market data only go back to 1995. So, before this time, the portfolio is composed of 25% each of the other asset classes.
2 – Defined in a certain way, size is another type of investing factor. For example, some fairly narrow statistical tests have shown that small caps outperformed large caps over long historical periods. But I’m focusing here on the fact that the performance of different stock size categories will wax and wane over time as we saw with my opening example.
3- This graph may seem to suggest that foreign markets “are due” for a period of outperformance soon, but that’s a cognitive trap known as the Gambler’s Fallacy at work. We can’t know for sure how long the current underperformance of foreign stocks will last.
