Correlations Between Stocks and Bonds Are Not Stable

One of the big arguments for portfolio diversification is correlation. Correlation in simple terms is how similar, dissimilar or opposite the returns of different investments are. Correlations are rarely static. They ebb and flow. Over longer periods of time, correlations between different types of investments can move closer to or further away from each other.

Consider two broad asset class groups: large-cap stocks (defined as the S&P 500 index) and 10-year Treasury bonds. An analysis by investment firm PGIM found that the correlations varied significantly over the past 70 years (1950–2020). At the low end, monthly correlations for the five-year period ending in 2015 were –0.66. At the high end, correlations were 0.45 for the five-year period ending in 1995.

(Correlations, for those of you not familiar with them, range from –1.0 to 1.0. A value of –1.0 occurs when the returns of two investments move in opposite directions. A value of 1.0 occurs when their returns move in the same direction. A correlation of 0.0 implies that the returns are independent of one another.)

Authors Junying Shen and Noah Weisberger identified three key periods where stock-bond correlations were either negative or positive. Correlations were –0.16 between 1950 and 1965 and –0.29 between 2000 and 2020. Correlations were a positive 0.28 between 1965 and 2000.

The obvious question is why?

One reason given by Shen and Weisberger is the level and volatility of short-term bond yields. Yields on the three-month Treasury note were 6.4% on an annualized basis between 1965 and 2000. Volatility in those yields was 2.5% annualized. Both yields and yield volatility were lower during the negative correlation periods.

During the positive correlation period of 1965–2000, inflation was 4.3%. Inflation ran at 1.3% between 1950 and 1965 and at 1.8% between 2000 and 2020.

Another reason given was macroeconomic policy. Discretionary monetary policy—such as a large rate hike that slows economic growth but supports the dollar—has been associated with positive correlations. Likewise, expansionary fiscal policy can contribute to positive correlations if it raises interest rates and crowds out private investment.

In contrast, raising rates during growth periods to fend off inflation reflects a rules-based approach. When monetary policy reflects economic conditions, negative stock-bond correlations have been more likely to occur. The link with fiscal policy is tougher to define but the authors suggest if interest rates are above growth rates, then stock-bond correlations are more likely to be positive.

Shen and Weisberger did not give an explicit forecast for where they think correlations are going. Rather, they struck a middle ground of sorts by suggesting chief investment officers without strong macroeconomic policy views anticipate a stock-bond correlation of 0.0.

Even if correlations were to become positive again, diversification benefits will be realized. Those benefits will still be reduced as correlations rise above 0.0 but the magnitude of the reduction depends where on the scale of 0.0 to 1.0 those correlations are.

The concept of correlations helps to explain how diversification works. Changes in correlations also explain why the benefits of diversification are more apparent during some years and less apparent during others.

Attempting to optimize your portfolio based on prevailing correlations is nowhere as effective as merely having a mix of stocks, bonds and cash aligned with your personal goals. Adhering to simple strategies like bucket strategies (allocating to risky and safe assets based on the timing of when cash flows are needed) or a preset percentage mix of stocks, bonds and cash has a far greater positive impact on long-term wealth.