How many investments do you need in your portfolio to be diversified? Can you reasonably buy plain-vanilla core stock and bond index funds and call it a day?
Many investors and their financial advisors appear to be examining those questions, as the torrent of asset inflows into ultradiversified total equity and bond index funds suggest.
One way to assess whether your portfolio is diversified is to examine how the pieces of your portfolio work together. If everything is humming along at the same time, that might feel good while it's happening, but you're courting risk because you're not diversified. Wall Street Journal columnist Jason Zweig once observed, "The very moment when something is the most painful to own, it's most likely to be a future bargain."
Of course, it's not a given that whatever performance patterns investments have exhibited in the past will hold up in the years ahead. The financial crisis is an instructive and not-too-distant example. Amid Armageddon in the global equity markets, categories such as corporate bonds, municipal bonds, and commodities, which had been reliable diversifiers in years prior, lost value at the same time stocks did. Only one major asset type, Treasury bonds, managed to hold its ground.
Does that blow a hole in the value of extrapolating from past returns in an effort to put together a well diversified portfolio? Not necessarily. For one thing, even though categories like municipal and corporate bonds couldn't escape losses in the financial crisis, they most certainly held up better than stocks. (The S&P 500 lost 37% in 2008, versus losses of 2% and 8% for the average municipal-bond and corporate bond funds, respectively.) There's also the not-small fact that barring new innovations in crystal ball technology, past performance patterns and a healthy dose of common sense are the best we've got when it comes to assembling well-diversified portfolios.
With the goal of examining performance correlations among various market segments over the short and long haul, I turned to Morningstar's Direct software, which is geared toward institutional investors.
The most recent run of the data--building on research conducted by my colleague Karen Wallace two years ago--shows that bonds have continued to be a decent diversifier for equity risk. Despite headlines touting an increasing correlation between stock and bond market performance, the correlation between the S&P 500 and the Bloomberg Barclays Aggregate has actually decreased over the past year relative to longer time periods.
There were a few interesting developments since our last look at correlations, however. A handful of investment types that in the past had exhibited a tight correlation with the S&P 500 appeared to be less correlated with the blue-chip index over the past year. On the short list: U.S. small caps (as represented by the Russell 2000 Index), foreign stocks, and high-yield bonds. Whether those investment types are likely to deliver diversification in the future remains an open question, however. Meanwhile, other categories that investors sometimes look to in order to garner diversification, especially those in the alternatives categories, have been a mixed bag on the correlations front.
Digging into the Data
Before going any further, let's take a closer look at how we measure correlations. Morningstar uses a statistic called "correlation coefficient" to measure an investment's performance correlation with another. If two investments are perfectly correlated, their correlation coefficient would be 1. If they're inversely correlated (for example, one investment gains in value over a given time frame and the other consistently declines), the correlation coefficient would be -1. If there's no correlation, the correlation coefficient would be 0. If your goal is to assemble a diversified portfolio, you'd ideally seek out investments with an inverse correlation to one another. Yet as noted above, correlations won't necessarily persist in the future; they can and will shift around based on what's happening in the stock and bond markets.
As we've done in previous assessments of correlations, I assessed correlations among a grab-bag of different indexes and mutual fund categories. Because indexes offer a pure play on an asset class, without the noise that can come along with mutual fund categories and their managers' active bets, indexes were my first choice in assembling these correlation matrixes. When I couldn't find an index with a sufficiently long track record, I used mutual fund categories and even an individual ETF (SPDR Gold Shares).
Morningstar Direct, which I used to assemble the data, creates what are called "correlation matrixes" to allow users to see performance relationships among asset classes. You can see correlation matrixes for the 1-, 3-, 5-, 10-, and 15-year periods here.
The matrixes can be a little tricky to make sense of at first, but the basic idea is that numbers along the rows in the horizontal axis correspond to the investment type with the same number along the columns in the vertical axis. To see the correlation between two investment indexes or categories, focus on the square where they intersect on the grid.
On the 10-year chart, for example, the first entry on the list is the S&P 500, so for each investment type listed below it, go to the first column to find the correlation of that investment type to the S&P 500. Over the past 10 years, the Russell 2000 index (Line 2, just below the S&P 500) has a high correlation of 0.91 with the S&P 500. The MSCI EAFE Index (Line 6) and long-short equity category (Line 8) have exhibited high correlations with stocks of 0.89 and 0.95, respectively.
Meanwhile, the Bloomberg Barclays 20+ Year Treasury Index and the ICE Bank of American Merrill Lynch High Yield Index (Rows 4 and 5, respectively) have exhibited the lowest correlation with the S&P 500 over the past decade: -0.30. Of the various "alternative" categories listed, only managed futures has exhibited a negative correlation with the S&P 500 over the past decade, though many funds in the various alternative investment groups, including managed futures, haven't been around for the whole decade.
Has Anything Changed?
Looking at correlations over long time periods is instructive because it encompasses more market environments and observations. But it can also be useful to examine correlations more recently, and compare them to past history.
One phenomenon I wanted to examine was whether correlations between the equity and bond markets had increased; earlier this year, the financial news media were filled with headlines suggesting a tighter performance link. However, our correlations data didn't bear that out. Turning to the 1-year correlations matrix, for example, and comparing that to the 10-year matrix, you can see that the correlation between the S&P 500 and the Bloomberg Barclays Aggregate Index has actually declined: It was 0.06 10 years ago and it was -0.14 for the 1-year period through May.
Another piece of data that jumped out at me was the declining correlation between high-yield bonds and the S&P 500. That's surprising given that many investors, me included, usually expect high-yield bond performance to fall somewhere between stocks and bonds. Recently, however, high-yield bond performance has been decidedly more bondlike. Over the past 1-, 3-, and 5-year periods, in fact, the high-yield index has had an even lower correlation with U.S. large caps than long-term Treasuries! Comparing the trailing 3-year correlations with the 3-year correlations from 2 years ago also demonstrates that high-yield bond performance is becoming more closely correlated with the U.S. Treasury market.
One possible explanation is simply that the yield differential (or "spread") between high-yield bonds and Treasury bonds remains quite low, and high-yield bond yields are low in absolute terms, too. Having less of a yield advantage could heighten high-yield bonds' sensitivity to rate changes. However, I'd hesitate to take it to the bank that high-yield bonds would continue to diversify U.S. equity exposure. After all, both junk bonds and equities are highly sensitive to the strength of the U.S. economy. While low-quality bonds typically lose less than stocks in a market decline sparked by economic concerns, there's no intuitive reason to expect the two investment types' performance wouldn't be directionally similar.
I was also interested in the fact that real estate funds' correlations with U.S. large caps had dropped significantly over the past year: The correlation coefficient between the S&P 500 and real estate funds is 0.76 over the past decade, but it's just 0.15 over the 1-year period. The explanation could lie with interest-rate increases, which have led to REIT losses over the past year even as the S&P 500 has notched solid gains.
Looking at correlations can be fascinating, but it's also important to remember how ephemeral they can be and to apply a healthy dose of common sense. For example, I'm inclined to take the declining correlations between U.S. stocks and junk bonds and foreign stocks with a grain of salt: When U.S. stocks are going down, those asset classes have rarely provided much solace.
One key element that's apparent in the near-term correlations data, however, is that a rising interest-rate environment has the potential to shake up correlations that had persisted during the declining-yield environment of the past several decades. High-quality U.S. bonds have been the most reliable diversifier for equities over the past several decades--and they have continued to be in 2018, headlines to the contrary. But if U.S. equities fall primarily because investors are concerned about rising interest rates, it's not a given that high-quality bonds, which are rate-sensitive themselves, will provide ballast. Instead, high-quality bonds' ability to diversify will be most useful in environments when investors are concerned about economic weakness. It's also worth noting that bonds haven't always been a reliable diversifier for equities, even though that relationship has persisted for the better part of the past 30 years. (This PIMCO research delves into that topic.)
Meanwhile, it's hard to get terribly excited about alternative investment types as a reliable means of diversifying equity risk, based on these data. While the managed futures category has been negatively correlated with the S&P 500 over the past decade, for example, more recently the correlation has increased to the point that it's even higher than the correlation between the Russell 2000 and the S&P 500. Categories like gold bullion (as represented by SPDR Gold Shares (GLD) in the matrix) appear to provide some diversification benefit relative to equities, but only make sense in very small increments due to their extreme volatility.
Christine Benz does not own shares in any of the securities mentioned above. Find out about Morningstar's editorial policies.