An Update on the Stock-Bond Correlation

Still positive as inflation remains the dominant concern
 

By: Greg Obenshain

Investors tend to think of stocks and bonds as negatively correlated, with stocks doing well in positive growth environments and bonds doing well when growth slows or declines. And that was largely true for the last quarter century.

But earlier this year, the trailing three-year correlation between stocks and bonds turned positive for the first time since November of 2000, a development we wrote about at the time. Over the last eight months, the stock-bond correlation has been 0.61, and three-year trailing stock bond-correlations are now as high as they have been since last century.

Figure 1: 3Y Trailing Monthly Stock-Bond Correlations, 1929–2023

Source: Bloomberg. SBBI Ibbotson US Large Stocks and US LT Govt prior to 1989.

This shift has prompted a surge in research on the stock-bond correlation.

AQR published a study at the end of 2022 that showed that a key determinant of the stock-bond correlation is the relative dominance of growth uncertainty and inflation uncertainty. Stocks and bonds react to growth shocks in opposite ways: stocks go up and bonds go down. But stocks and bonds react in the same direction to inflation shocks: stocks go down and bonds usually go down more, in our opinion.

Now researchers at Robeco and the State of Wisconsin Investment Board have published another valuable paper on the stock-bond correlation (Molenaar et al). In it, they explore what characterizes periods in which the stock-bond correlation is above or below zero. Like AQR, they find that higher expected inflation and higher uncertainty around inflation drive higher stock-bond correlations.

One of the most eye-popping charts from the paper is recreated below. When the core inflation rate averages above 4%, the stock-bond correlation has been positive with few exceptions. Core inflation has averaged 4.5% for the past three years and is currently 4.7%.

Figure 2: 3Y Trailing Monthly Stock-Bond Correlation vs. Avg. Core Inflation Rate over Same Period

Source: Bloomberg. SBBI Ibbotson US Large Stocks and US LT Govt prior to 1989.

While both papers note that high levels of inflation are associated with high stock-bond correlation, both papers forcefully argue that it’s not so much the level of inflation that is the driver of the correlation but the uncertainty around inflation. Both papers come to this conclusion using different data: AQR uses the trailing volatility of inflation while Robeco and the State of Wisconsin use the divergence of inflation survey expectations. As it happens, inflation uncertainty is highly correlated with inflation, meaning uncertainty is higher when inflation is higher, but even when untangling this relationship through statistical means, inflation uncertainty remains important.

How might uncertainty around inflation lead to a more positive stock-bond correlation? One reason is simply that, to the extent uncertainty is associated with greater variance of inflation, the movement of inflation is more likely to be the dominant driver of change in both debt yields and equity discount rates. Higher yields drive debt prices down, and higher discount rates drive equity prices down.

But that ignores equity growth, which matters greatly to equity price. There is a second pathway that includes growth: inflation as a leading indicator of counter-cyclical monetary policy. When inflation is uncertain, the Fed and the market are more likely to be focused on inflation. Upward surprises lead to upward revisions of expected Fed policy rates, and downward surprises lead to downward revisions of expected Fed policy rates. Not only do these move yields and discount rates in the same direction, but they change growth expectations in a reinforcing way: either growth expectations fall in combination with expected higher policy rates and rising inflation, or growth expectations rise in combination with expected lower policy rates and lower inflation. Each of these pushes stock and bond prices in the same direction.

The theory is that, if inflation is the primary driver of the Fed reaction function, then stocks and bonds are likely to be more highly correlated. When inflation is not a concern, then growth becomes the primary driver of the Fed reaction function, and policy rate moves become driven by growth concerns. Adjustment to lower rate expectations comes with lower growth expectations. This drives stocks and bonds in opposite directions. In this context, it makes sense that high stock-bond correlations exist when inflation is the focus, which happens most often when inflation is high. By contrast, it makes sense that positive stock-bond correlations require quiescent inflation and happen most often when inflation is low and contained. It is also intriguing that Robeco and the State of Wisconsin find that their regressions become much more significant after 1952, in the era of independent central banks and counter-cyclical monetary policy.

In this context, the stock-bond correlation can be seen as an indicator of what is the dominant risk—inflation or growth—and how it is changing. This is perhaps why we have found it is an important leading indicator in our own work. It is important to note that the charts we have shown have been smoothed over three years. Over shorter time horizons, the stock-bond correlation is constantly shifting as expectations in the market update. Below we add the one-year stock-bond correlation to the chart we showed before.

Figure 3: 1Y and 3Y Trailing Monthly Stock-Bond Correlations, 1929–2023

Source: Bloomberg. SBBI Ibbotson US Large Stocks and US LT Govt prior to 1989.

Even in periods of high stock-bond correlations, stocks and bonds can be negatively correlated over shorter periods. In fact, over the first eight months of this year, stocks and bonds moved opposite one another in May, June, and July. This also helps explain how in the 1970s during a period of sustained positive stock-bond correlations, Treasurys still had positive returns in recessions, as we noted here. We believe a high and sustained stock-bond correlation is certainly a reason for investors to revisit their strategic allocations, but it does not mean Treasurys are to be avoided altogether. For the tactical asset allocator, the more important message comes from how the stock-bond correlation is changing.