While it would be fun to hold a Portfolio in which all of the asset classes were going up at the same time, it would be a nightmare if they were all moving down at once.

To protect against epic losses, advisors and their clients seek uncorrelated returns — asset classes that behave dissimilarly so that a portfolio’s ingredients don’t all move in the same direction at the same time.

But diversified portfolios should be built with the knowledge that including uncorrelated assets means always having to endure pain in part of the portfolio. That part will be going down or remaining flat, since it tends to move differently from the uncorrelated part that is going up.


There is a well-known expression — “No pain, no gain” — that is applied to training for sports, music, dance, you name it. It also applies to portfolios, but it goes like this: “No pain, no low correlation.”

The basic premise of diversification and asset allocation is to build a portfolio in which the ingredients have low correlation with one another. The general goal is to avoid a portfolio in which all the asset classes collapse at the same time, although a well-diversified, low-correlation portfolio will probably be experiencing declines in some of its ingredients at any moment in time.

In the long haul, however, the gainers should overwhelm the losers. Over the past nine decades, the large-cap U.S. equity class has produced positive returns 73% of the time and the U.S. fixed-income class 90% of the time.

For other asset classes, including such diversifier classes as real estate and commodities, we also observe a tendency toward positive returns. Since 1970, real estate (as measured by the DJ U.S. Select REIT Index) has produced positive annual returns 80% of the time, while commodities (S&P GSCI) have generated positive returns 71% of the years.

Low correlation is really about finding asset classes that have a tendency to have high positive returns, but which produce those positive returns at different times. That is where the performance smoothing comes from, which is one of the distinctive benefits of a low-correlation portfolio.


The past six years is an interesting period to examine. The meltdown of 2008 actually began in November 2007 and lasted through February 2009. The S&P 500 lost nearly 51% during that painful 16-month period.

This study reviews the past six years in two distinct segments: the three-year period from August 2009 to July 2012 and the three-year period from August 2012 to July 2015.

In the first three years, virtually all asset classes rebounded from the devastation of 2008. The overall correlation within a diversified 12-index model was 0.4. (Recall that 1 indicates perfect positive correlation, minus 1 indicates perfect negative correlation and zero indicates a random correlation.)

In a portfolio with 12 asset classes, there are 66 unique correlations among the 12 ingredients: large-cap U.S. stocks to mid-cap U.S. stocks, large-cap U.S. stocks to small-cap U.S. stocks, and so forth. During this period, 26 of the 66 intra-portfolio correlations (or 39%) were high, having a correlation over 0.7.

It happens that, for this period, all of the asset classes except cash had relatively impressive three-year annualized returns. Real estate led the way with an annual return of 29.99%. The U.S. mid-cap stock category was second, with a return of 16.06%.

During this timeframe, a portfolio that includes all 12 indexes in equal-weighted allocations produced a three-year annualized return of 10.75% from August 2009 to July 2012.


During prosperous times, we tend to ignore rising correlations because we are enjoying positive returns. But the pathway to achieving a lower-correlation portfolio will necessitate that some asset classes take their turn in the red. And that is exactly what happened from August 2012 to July 2015.

As shown in “The Confusion” chart below, the number of index pairs with a high correlation decreased markedly. Only five of the 66 intra-portfolio correlations were high. This represents a decline from 39% high correlations within the portfolio to 8%. The overall correlation in the 12-asset portfolio fell to 0.3, representing a reduction of 25% from the earlier three-year period.

This type of reduction is encouraging if the goal is to have a low-correlation portfolio, but a price had to be paid to achieve a minimal level of correlation.

The three-year performance of the natural resources index, for instance, fell from 9.02% during the first three-year period to minus 1.19% in the most recent period. Even worse, the commodities index three-year annualized return fell from 6.13% in 2009-2012 to minus 16.25% in 2012-2015. TIPS went from 10.29% in 2009-2012 to minus 1.31% in 2012-2015. Painful stuff.

But there is an upside to all this: The correlation between commodities and large-cap U.S. stocks was 0.73 in 2009-2012. That is an unusually high correlation between commodities and stocks. In the later three-year period, the correlation between large-cap U.S. stocks (S&P 500) and commodities (Deutsche Bank Liquid Commodity Optimum Yield TR Index) shrank to 0.33 — a much more typical correlation.


There was only one way to get back to a normal correlation level: Either the S&P 500 or the Deutsche Bank Commodity Index needed to experience negative returns. During the past three years, it was the commodity index that took the arrow.

But there is no way to know in advance which asset class is going to underperform. Consider the not-so-distant time period of 1998 to 2002. Large-cap U.S. stocks enjoyed brilliant returns in 1998 and 1999. But then, in 2000, this category had a return of minus 9.1%. In 2001, it lost 11.9%. Then, in 2002, the wheels really fell off, with a 22.1% loss. By the end of 2002, the S&P 500 had experienced a three-year loss of nearly 38%.

A low-correlation portfolio is designed to help us during such times, and a number of other asset classes offered some protection from the losses of large-cap U.S. stock in 2000-2002. Real estate had positive returns in 2000, 2001 and 2002. Natural resources had a positive return in 2000. Commodities had large positive returns in 2000 and 2002.

Some traditional diversifiers such as non-U.S. stocks offered no help at all; the MSCI EAFE Index and MSCI EM Index experienced losses in all three years as well.

But U.S. bonds were helpful, as were TIPS. Non-U.S. bonds helped out in 2002. Cash was helpful in all three years (ah, the good old days of yields in cash over 1%).


The point here is simple: A low-correlation portfolio needs to have ingredients that make you happy sometimes and sad sometimes. Back in 2000-2002, it was the S&P 500 that was creating a lot of angst and commodities were the hero. For the past three years, large-cap U.S. stocks have been on top and commodities have been lagging.

It’s crucial to always look beyond the present moment and see the longer horizon. You need “super-vision” to avoid becoming biased by recent events. Don’t chase current winners, and don’t dump temporary losers.

A benefit of recent negative returns among the diversifier asset classes such as commodities, real estate and natural resources is that those asset classes now have much lower correlation with large-cap U.S. stock — a core asset class for nearly all U.S.-based investors.

Of course, there is an irony here: While many investors say they want a low-correlation portfolio, they don’t want to actually experience a low-correlation portfolio. The experience of a low-correlation portfolio is the experience of owning asset classes that are making you happy while others are making you sad. When that’s happening, you have low correlation. Enjoy it. 

Craig L. Israelsen, a Financial Planning contributing writer in Springville, Utah, is an executive in residence in the personal financial planning program at the Woodbury School of Business at Utah Valley University. He is also the developer of the 7Twelve portfolio.

Read more: