Skip to content

Stock-Bond Correlations, Macroeconomic Regimes and Monetary Policy: summary and key takeaways

The paper tests whether inflation, output and monetary policy regimes explain the sign and size of stock-bond correlations in 10 developed markets. Negative correlations show up mainly when policy is accommodating and inflation is low. Correlations are positive whenever policy is restrictive.

Baele, L.; Holle, F.v. · 2017 · SSRN Electronic JournalSummary updated 20 November 2026
Published2017Working paper
Asset focusStocks+BondsStock-bond correlation
TypeResearchStock-Bond Correlation
Reading time4 minthis summary

Stock-Bond Correlations, Macroeconomic Regimes and Monetary Policy at a glance

The question
Do macroeconomic regimes and the monetary policy stance explain why stock-bond correlations switch between positive and negative?
Dataset & period
10 developed markets (US, Canada, Japan, UK, Germany, France, Netherlands, Belgium, Italy, Spain). Daily total equity index returns and 10-year benchmark government bond returns in local currency, with the longest series (US) starting in June 1961 and all samples ending in December 2013. Monthly data extend the shorter series. Correlations come from DCC-MIDAS (daily data) or plain DCC (monthly data).
Method
The authors first fit regime-switching models to inflation, the industrial-production output gap and a monetary policy rule, each estimated separately from asset returns. They then regress Fisher-transformed conditional stock-bond correlations on regime dummies, with no intercept, so each coefficient is an average correlation in that regime. Panel regressions force the same coefficients across countries and use either global or local regimes. Fit is judged by adjusted R2, mean absolute difference and sign hit rate.

Key findings of Stock-Bond Correlations, Macroeconomic Regimes and Monetary Policy

  • Inflation regimes alone fit poorly. Correlations are lowest and mostly negative but insignificant in the low inflation state, and positive and significant in intermediate and high inflation. The switch to low inflation came years before the correlation drop around 2000.
  • Output-gap regimes (recession versus expansion) explain very little of the variation in correlations.
  • In the global-regime panel, the average correlation is -11.5% when policy is accommodating and 28.1% when it is restrictive.
  • Combining output, inflation and policy regimes, low inflation with accommodating policy gives about -28.6% in recessions and -26.5% in expansions. Restrictive policy in low inflation and low growth gives 34.3% versus -28.6% under accommodating policy, a gap of nearly 63% points.
  • Adding policy interactions raises R2 from 26.3% to 59% and the sign hit ratio from 76.7% to 88.9%. States tied to negative correlations cover about 35% of the sample.
  • Global regimes explain correlations better than local ones. Local regimes add little once global regimes are included.

Robustness, as reported in the paper

  • Policy regimes re-estimated using data only until 2008 gave very similar accommodating/restrictive classifications, and results were robust to this choice.
  • Using US regimes as a stand-in for global regimes gives qualitatively similar results but a lower R2 (44% versus 59%).
  • An epilogue compares the model with actual US correlations from January 2014 to March 2017; the regression predicted the average US level over that period.
  • Results on the lack of a procyclical inflation regime hold with consumption growth, quarterly data, added lags and GDP-weighted global series.

Caveats and limits

  • The fitted model misses the most extreme positive correlations of the early 1990s and the most extreme negative ones since 2000, which the authors tie to flights to safety and sentiment.
  • Italy and Spain fit poorly because sovereign default premia drove bond returns during the euro debt crisis; Japan also fits worse.
  • Policy regimes after 2008 are hard to identify at the zero lower bound, so the paper assumes accommodating policy from then on in the restricted sample.
  • The regime dummies are in-sample classifications with no trading costs or portfolio test (inferred).

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

Why it matters

The paper ties the diversification value of bonds against equities to the policy backdrop and the inflation level, not to growth alone. A practitioner can read the results as a map of which macro states went with negative or positive correlations in this sample. It also argues that the global state matters more than local conditions.

Questions about this paper

When are stock-bond correlations negative according to the paper?

Mainly when monetary policy is accommodating and inflation is low to intermediate. In the global panel the average is -11.5% under accommodating policy.

What happens to the correlation when policy is restrictive?

It is positive in every inflation and growth regime. The global panel average is 28.1% in the restrictive state.

Do inflation or growth regimes alone explain the correlation?

No. Each has weak explanatory power, and the monetary policy stance adds the most once it is combined with them.

Where can I read the original paper?

The original is at its publisher: https://doi.org/10.2139/ssrn.3075816. 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.