Have you ever stopped to ask why you’re supposed to allocate a good chunk of their portfolios to fixed income BIV, -0.02% ?
Probably not, since most retirees take it as obvious that bonds have a low correlation to stocks. If that were the case, of course, a stock-bond portfolio would be a lot more conservative than a stock-only portfolio.
In fact, however, the correlation between stocks and bonds has varied widely over the last century. Sometimes that correlation has been low or even negative, as indeed is the conventional wisdom’s expectation. But at other times the correlation has been strongly positive, and retirees will be in for a rude awakening if that turns out to be the case during the next stock market downturn.
Let me start by reviewing the data. Take a look at the accompanying chart, which plots the correlation coefficient of the trailing five years’ monthly returns of the S&P 500 SPX, +0.18% and Intermediate-Term U.S. Treasurys. (This coefficient ranges from plus 1.0, which would mean the assets are perfectly correlated, to minus 1.0, which would indicate a perfectly inverse correlation; a zero coefficient would means the two have no correlation whatsoever.) Since the 1920s, this trailing five-year correlation coefficient has been as high as 0.57 (in the late 1990s) and as low as -0.51 (in the early 2000s).
This wide variability comes as a surprise to most retirement financial planners too. They typically base their asset allocation recommendations on the implicit assumption that the stock-bond correlation in the future will be what it has averaged over the last century. Since the correlation coefficient for the entire period since 1926 has been barely higher than zero—just 0.06, in fact—it is little wonder that most financial planners argue that retirees can greatly reduce their risk by allocating a large portfolio percentage to bonds.
Depending on whether the actual stock-bond correlation is higher or lower than the long-term average, however, you either will be better off than financial planners assume, or worse. During equity bull markets, of course, you’d be better off if the stock-bond correlation were positive, since that would mean that bonds were generally rising along with equities. During stock bear markets, in contrast, you’d want a negative correlation.
A close examination of the chart shows that these hoped-for correlations happen some but not all of the time. For example, stock-bond correlations were generally positive during the stock bull market of the 1990s, just as they were negative during the financial crisis in 2008 and 2009. So far, so good. But correlations were positive during the Great Depression and negative during the bull market since 2009; in both cases these caused stock-bond portfolios to have lower risk-adjusted returns than expected.
Many researchers over the years have tried to come up with economic models that forecast when stock-bond correlations will be positive and when negative. Perhaps the biggest factor they discovered is inflation. That’s because both stocks and bonds tend to be hurt simultaneously when inflation rises—causing correlations to rise. In contrast, correlations tend to be low or negative when inflation is low—as it has been over the last decade.
Perhaps the most important investment implication of these results is that retirees should reduce their expectations for what they will gain by shifting a sizable portion of their equity portfolios into bonds. Especially if the next stock downturn is accompanied by a worsening of inflation, there’s a good chance that the bond portion of your portfolio will decline alongside your stocks.
That would be adding insult to injury.
For more information, including descriptions of the Hulbert Sentiment Indices, go to The Hulbert Financial Digest or email mark@hulbertratings.com.
