A new approach to a 60/40 portfolio

Let’s talk about the venerable 60/40 portfolio. For decades, this has been the conventional way to provide moderate-risk, long-term growth in a retirement portfolio.

The classic mix is 60% large-cap U.S. stocks and 40% U.S. bonds.

At the moment, I won’t mess with the 60% stock portion, but how about the 40% component? If you were building a 60/40 model from scratch, you might wonder about that 40% piece. Does it really have to be U.S. bonds, or is there another asset class that would work even better?

The chart “Potential Large-Cap Partners” shows the return and risk metrics for large-cap U.S. stocks and for 11 other major asset classes that could be considered as partners for large-cap U.S. stocks in a 60/40 portfolio. Included among the 11 are U.S. aggregate bonds (see "U.S. Bonds"), which is the typical 40% partner for large-cap U.S. stocks. In this analysis, the performance of large-cap U.S. stocks is represented by the S&P 500.

The 11 major asset classes are listed in the table in order of their 15-year correlation with large-cap U.S. stocks — from lowest to highest. The first potential partner for large-cap stocks is U.S. TIPS, with a 15-year correlation with large-cap stocks of 0.02; 15-year average annualized return of 5.3% from 2002-2016; a 15-year standard deviation of annual returns of 6.69%; and a 2017 YTD return of 1.94% through Oct. 31, 2017.

Recall that in most cases the desired correlation coefficient is zero (or close to it), indicating a completely random correlation. So, a correlation coefficient of 0.02 is a very low correlation.

The asset class with the next lowest correlation with large-cap U.S. stocks over the past 15 years was U.S. cash at minus 0.04.

Cash produced a 15-year return of 1.25% with very little volatility (the standard deviation was 1.64%). Next were non-U.S. bonds with a 15-year correlation of 0.06, then U.S. bonds with a 15-year correlation of minus 0.09, and finally commodities with a 15-year correlation of 0.32 and a 15-year return of 4.85%.

Unlike the fixed income asset classes (TIPS, cash, bonds, and non-U.S. bonds), commodities had a very large standard deviation of return at 21.68%. The commodities 2017 YTD return was 1.19% as of Oct. 31. These five asset classes represent the low-correlation potential partners for large-cap U.S. stock.

The remaining six asset classes in the chart all have higher correlation with large-cap U.S. stocks. Of the six, global real estate has the lowest 15-year correlation of 0.63, just a bit lower than natural resources with a 0.66 correlation.

All but one of the higher correlation asset classes outperformed large-cap U.S. stocks over the 15-year period 2002-2016. The exception was developed non-U.S. stocks (MSCI EAFE index). However, it is worth noting that as of Oct. 31, 2017, the year-to-date performance of the EAFE Index was 21.78% — well ahead of the 16.9% return of the S&P 500 Index. The 2017 YTD performance of emerging market stocks, however, was the clear winner at 32.64%.

The correlation of non-U.S. developed stocks and large-cap U.S. stocks over the past 15 years was 0.86 — only slightly lower than the 0.91 correlation between large-cap U.S. stocks and mid-cap U.S. stocks. This suggests that non-U.S. stocks — particularly in developed non-U.S. economies — are no longer a reliable diversifier for U.S. investors. In the 1970’s and 1980’s, developed non-U.S. stocks had much lower correlation with the U.S. equity market and, as such, were a useful diversifier for U.S. investors. However, the economic intertwining of the globe has synchronized markets to a high degree over the past 10 to 15 years.

WHICH IS BEST?
Now, on to the real issue: which of these 11 asset classes was the best partner when teamed with large-cap stocks over the past 15 years? The answer depends on your goal: do you want lower volatility or enhanced performance? A summary of the 15-year risk and return measurements for various 60/40 portfolio combinations is provided in the table “60/40 Combos.”

If your goal was to reduce volatility below that of the S&P 500, the five low-correlation asset classes (US TIPS, cash, non-U.S. bonds, U.S. bonds and commodities) achieved that. For example, a portfolio that consisted of 60% large-cap U.S. stocks and 40% TIPS had a 15-year standard deviation of return of 10.92%, compared with 18.16% for large-cap U.S. stocks by themselves. Yet, the 15-year return dropped by only 4 basis points — from 6.69% to 6.65%.

Recall from the previous table that the 15-year return of TIPS by itself was 5.3%, so one might think that adding TIPS as the 40% component in a 60/40 portfolio would have degraded performance by far more than 4 bps. The magic of lowering volatility while maintaining performance is in the low correlation of TIPS to large-cap U.S. stocks (which was 0.02 over this 15-year period). TIPS and large-cap U.S. stocks do not march to the same drummer — and that’s a good thing. Thus, the performance of the teammate to large-cap stock can have a lower return than large-cap U.S. stock, but if it also has a low correlation to U.S. stocks, the net result can be a 60/40 portfolio that has nearly same return as large-cap stock by itself — but with far less volatility.

If your objective was to maximize return, the high correlation ingredients accomplished that goal — with the exception of developed non-U.S. stocks. In fact, let’s examine the impact of blending developed non-U.S. stocks with large-cap U.S. stocks in a 60/40 portfolio a bit more deeply. You will notice that developed non-U.S. stocks (specifically the MSCI EAFE Index) had a 15-year return by themselves of 5.28% — almost identical to the 15-year return of TIPS.

However, developed non-U.S. stocks had a high correlation of 0.86 with large-cap U.S. stocks. Thus, unlike TIPS, combining the MSCI EAFE Index with the S&P 500 resulted in a 15-year return of 6.23% — or 42 basis points below the return of a 60% U.S. large stock/40% TIPS portfolio.

GOOD TEAMMATES
Here is a key takeaway message: When building portfolios, we need to know if the fund we are considering will be a good teammate to the other funds already in the portfolio, rather than its performance as a stand-alone fund.

Calculating the correlation of a fund with the other funds already in a portfolio is one way to assess that particular fund’s potential to be a good teammate. Understandably, correlation in the past is not a perfect predictor of correlation in the future, but it is a logical starting point to evaluate when building multiasset portfolios.

Note that the Barclays U.S. Treasury U.S. TIPS Index and the MSCI EAFE Index had a nearly identical 15-year return in 2002-2016. But, the TIPS index was a better teammate for large-cap U.S. stocks because of its lower correlation to the S&P 500. Blending low-correlation funds together is not a guarantee that performance will improve, but it almost always results in a reduction of volatility. And that alone makes it worth doing.

As a final note, as shown in the last row of the chart, if all 12 asset classes (large-cap U.S. stock through midcap U.S. stock) were blended together in equal portions of 8.33% and rebalanced annually, the 15-year standard deviation of return was 13.22% (27% lower than the S&P 500 Index by itself), the 15-year return was 7.6% (91 bps higher than the S&P 500 by itself), and the performance in 2008 was minus 25.5% (31% better than the 37% loss experienced by the S&P 500.

Moreover, the 12-asset portfolio (which is a 65% growth/35% fixed income model) outperformed the traditional 60/40 model by 124 bps over the 15-year period 2002-2016, albeit with somewhat higher volatility (13.22% standard deviation versus 10.43% standard deviation).

Now, to return to my core question, is there a better partner for U.S. large-cap stocks in a 60/40 portfolio? Depending on your investment objectives, there are several that you might find better. But, beyond that, consider more than two asset classes.