Skip to content

US Stock-Bond Correlation: What Are the Macroeconomic Drivers? (PGIM): summary and key takeaways

The paper splits US stock-bond correlation into three macro components: rate level and volatility, the link between growth and rates, and whether equity and bond risk premia move in step. It ties regime shifts since the 1950s to monetary and fiscal policy settings. It also simulates what a correlation shift would do to a stock-bond portfolio.

Shen, J.; Weisberger, N. · 2021 · SSRNSummary updated 20 November 2026
Published2021Working paper
Asset focusStocks+BondsStock-bond correlation
TypeResearchStock-Bond Correlation
Reading time4 minthis summary

US Stock-Bond Correlation: What Are the Macroeconomic Drivers? (PGIM) at a glance

The question
Which macroeconomic conditions and policy settings determine whether US stocks and bonds are positively or negatively correlated, and what could change the current regime?
Dataset & period
S&P 500 and 10y US Treasury monthly total returns, 1950 to 2020, correlation from a centered rolling 5y window. Regime tables also use 3m and 10y yields, CPI inflation, real consumption growth and equity and bond risk premia. Sample size: Not stated in the paper.
Method
The authors write stock and bond prices as discounted cash flows and derive three drivers of their covariance. These are the variance of the risk-free rate, the covariance of equity cash flows (proxied by growth) with rates, and how equity and bond risk premia covary. They compare these drivers across three historical correlation regimes and across non-overlapping 5y blocks. They then link the drivers to policy and run a normal-returns portfolio simulation.

Key findings of US Stock-Bond Correlation: What Are the Macroeconomic Drivers? (PGIM)

  • Monthly stock-bond correlation was -0.29 in the 2000-2020 window shown in Figure 2. The paper says it has been negative since 2000 and was persistently positive from the late 1960s to the late 1990s.
  • Rate levels line up with the regime. 3m yields averaged 6.4% in the positive regime, against 2.3% and 1.6% in the two negative regimes.
  • Yield volatility was two to three times higher in the positive regime than in the negative ones.
  • Growth and rates moved in opposite directions in the positive regime (correlation -0.33) and together in the recent negative regime (0.2). The 1950s-1960s negative regime did not fit this pattern, though its growth-rates correlation was near zero.
  • Across non-overlapping 5y blocks, the 3m yield level explains 55.3% of the variation in stock-bond correlation, yield volatility 24.9%, and growth-rates correlation 36.0%.
  • In a simulation with equal stock and bond Sharpe ratios, moving correlation from the recent negative level to zero lifts 60/40 volatility to 7.5%. Holding volatility fixed then needs about 8ppt less in stocks and cuts expected return by about 0.4%/y. At a correlation of 0.3, the stock weight falls to about 40%.

Robustness, as reported in the paper

  • Using non-overlapping returns at 1m, 6m or 12m frequency inside a 5y window gives qualitatively similar results to the monthly measure.
  • The paper avoids rolling overlapping long-horizon returns, arguing they give an illusory sample size and persistent autocorrelation. The appendix illustrates this with simulated data and with the 1987 sell-off.
  • The appendix shows that sample size drives wide confidence intervals for correlation estimates, which is why the paper uses monthly data with 5y windows.
  • The three-component analysis is univariate; the paper notes a multivariate approach is possible.

Caveats and limits

  • The regime evidence rests on three historical episodes, so the number of independent regimes is small (inferred).
  • The link between policy settings and correlation is argued qualitatively. The paper says no single factor determines the sign and the framework is no panacea.
  • One regime does not match the growth-rates prediction, and the paper flags its discussion of a recent shared downward trend in risk premia as speculative.
  • The portfolio simulation assumes jointly normal returns and constant expected returns and volatilities, and ignores costs (inferred).

Items marked (inferred) are our reading of the design, not statements by the authors.

Why it matters

The paper gives a checklist of macro conditions to watch when judging whether the stock-bond hedge could weaken. It also shows how much a correlation shift alone changes portfolio volatility and the equity weight a risk budget allows.

Questions about this paper

What drives the sign of US stock-bond correlation according to the paper?

Three components: how volatile and high interest rates are, whether growth and rates move together, and whether equity and bond risk premia move together. These in turn reflect monetary policy style, fiscal sustainability, policy independence and supply versus demand shocks.

Are low interest rates enough to keep the correlation negative?

No. The paper says continued low rates alone may not support a negative correlation, and it describes low-rate scenarios that still produce a positive one.

What does a positive correlation do to a stock-bond portfolio in the paper's simulation?

Portfolio volatility and the risk-adjusted return worsen even if expected returns barely change. Keeping volatility fixed would require a lower stock allocation.

Where can I read the original paper?

The original is at its publisher: https://doi.org/10.2139/ssrn.3855610. We link to the paper and do not host the PDF.

Follow new quant research

Download MacroFinance market data as CSV or Excel, or pull it by API, and test the ideas in these papers yourself. New summaries are added as we review them.

Create a free account
Summaries are written by MacroFinance from the published papers, in our own words, for information only. They are not investment advice and may omit detail. Papers are not hosted here: each page links to the original at its publisher.