Bonds may not come to stocks’ rescue any time soon.
Failing to do so would dash the hopes of millions of retirees and near-retirees who own diversified stock-bond portfolios. They put bonds in their portfolios on the theory that a fixed-income allocation would cushion stock market losses. But try telling that to the markets this year.
As of Jun. 13, the day the S&P 500 dropped below the 20% loss threshold that satisfies the definition of a bear market, long-term U.S. Treasurys had lost even more: The Vanguard Long-Term Treasury ETF
produced a 22.1% loss over the same period.
This is just the opposite of what happened in other recent bear markets. In the early-2020 bear market that accompanied the onset of the Covid-19 pandemic, for example, when the S&P 500
fell 33.9%, Vanguard’s Long-Term Treasury ETF gained 13.4%. And during the bear market accompanying the Global Financial Crisis from 2007 to 2009, during which the S&P 500 fell 56.8%, the Vanguard Long-Term Treasury Fund
gained 19.9%. (VGLT didn’t exist back then.)
Hope springs eternal, however, and many believe that—any day now—bonds will finally kick into gear and live up to this historical reputation. New research in effect says “not so fast:” An analysis of the last century shows that bonds could very well continue to lose ground in an extended stock bear market.
The research, from the Portfolio Solutions Group at AQR Capital Management, is entitled “The Stock/Bond Correlation.” Their study analyzed stock and bond data back to 1900, focusing on the degree to which the two asset classes have been correlated with each other. They found that, more often than not over the last 120 years, stocks and bonds were positively, not negatively, correlated.
Take a look at the accompanying chart, which I produced using a similar methodology as employed in this AQR study. It plots the trailing 10-year stock-bond correlation back to 1881, and notice that the majority of the time the correlation has been positive—71% of the time, in fact. With the exception of recent months, this century represents the longest sustained period in which the correlation was negative.
Only if you were to believe that this century represents the new normal would you want to bet that bonds and stocks will remain negatively correlated in coming years. Otherwise, such a bet rests on a shaky historical foundation.
The roles of inflations and recessions
Is that all we can say about the stock-bond correlation—that it on average is positive but fluctuates widely?
The AQR researchers believe we can say more than that. They found that a number of factors can explain fluctuations in the stock-bond correlation. They are:
Inflation. Unexpected changes in inflation will tend to cause the correlation to increase. We’ve seen this in recent months, of course, as rising inflation has caused both stocks and bonds to decline.
GDP growth. Unexpected changes in the economy’s growth rate will tend to cause the stock-bond correlation to decrease. That’s because bonds typically perform well during recessions and stocks poorly; the opposite is the case if the economy grows faster than expected.
The inflation-growth correlation. This third factor focuses on the relationship of the first two. While inflation and growth historically have often been positively correlated—with inflation rising as the economy heats up—this has not always been the case. And, the researchers found, when the two move in opposite directions—such as during the stagflation era of the 1970s—the stock-bond correlation tends to increase.
What insight do we get into the future of the stock-bond correlation by analyzing these three factors? The first and third point to a rising correlation, at least for the near term, while the second possibly to a declining one. On balance that suggests that the path of least resistance will be for the correlation to increase over the near term.
That’s why I conclude from this research that investors may be on shaky ground expecting bonds to perform well in an extended bear market.
What about the longer term? As the proverb says, it’s difficult to make predictions, especially about the future. A sustained period in which the stock-bond correlation remains positive, according to the AQR researchers, “would probably require a rise in longer-term inflation uncertainty accompanied by further supply-driving inflation shocks and or monetary policy errors.” They classify the likelihood of such a sustained period of positive correlation, while perhaps higher than in previous years, as still relatively remote—a “tail risk.”
If it does enter such a period, however, we will need to look to assets other than bonds to obtain the portfolio diversification we were otherwise hoping to gain. One such asset class, which the researchers specifically mention, is commodities. Another could be real estate.
Mark Hulbert is a regular contributor to MarketWatch. His Hulbert Ratings tracks investment newsletters that pay a flat fee to be audited. He can be reached at firstname.lastname@example.org.