Why last year’s correlation shock is not the new normal
One of the many unusual developments in financial markets last year was the decoupling between German Bunds and other safe haven G7 government bonds, most notably US Treasuries. Since the inception of the euro, it’s been quite a rare event that Bunds and Treasuries move in opposite directions for sustained periods of time. This reflects both the interconnectedness of financial markets and the fact that these are substitute assets, at least most of the time. While the US and European economies do not move perfectly in tandem every quarter, it would nevertheless be surprising to see an economic depression in the Eurozone and a massive boom in the US, and vice versa.
Last year, however, we witnessed the largest adjustment to Germany’s fiscal stance in decades. In March 2025, the government announced a relaxation of fiscal rules that effectively gave them the ability to increase defence spending dramatically. Bunds reacted accordingly, with yields moving sharply higher, and because this was a Germany-specific headline, correlations broke down completely. The graph below shows rolling correlations between total returns of the 7-10 year US Treasury and 7-10 year German Bund using daily returns. The grey line shows the correlation using the last three months’ returns, while the green one uses the last six months’ returns. As can be seen, correlations fell to 25-year lows for both data sets, due to the unexpected German fiscal event. The translation into total returns was clear; 10-year USTs returned just over 8% in 2025, while 10-year Bunds delivered a slightly negative return in euros, or a slight positive return, if hedged into dollars.
Since then, correlations have normalised, with Bunds and UST moving more closely together, and returning to normal ranges. When a financial asset underperforms by a large margin in a given year, it is reasonable to think that the following year, the opposite might happen. In this case, however, there’s no reason to expect the underperformance last year will be reversed, unless Germany was to reassert its debt brake and cancel its defence spending plans. At the same time, we believe that it is unlikely that correlations will become as distorted again, and exposures to the USD curve are unlikely to outperform in the same way they did last year.
The relative benefit of owning US Treasuries, rather than similar Bunds, also provided a boost to US high yield and investment grade bonds, given how favourable the US curve ultimately proved to be. With correlations normalising, we expect differences in credit fundamentals and active credit selection will once again play a bigger role than underlying curves. Investors can now enjoy the higher yields on the Bund curve, now that it has reset higher.
It is, of course, very difficult to predict what the return for 10-year Bunds and 10-year Treasuries will be in 2026. However, we are more confident in saying that a correlation breakdown of the magnitude seen last year is highly unlikely to be repeated.
