The Diversifier Investors Thought They Already Had

Daniel Li, PhD. Head of Index Research

In 2021, U.S.-listed managed futures ETFs held less than a billion dollars in combined assets. By May 2026, that number had reached roughly $7 billion, with the largest fund in the cohort tripling its asset base in roughly fourteen months. That kind of trajectory doesn’t happen because retail investors woke up one morning enthusiastic about commodity trading advisors. Something structural is pulling capital into a strategy that, for forty years, occupied a small, stable corner of institutional portfolios — typically around 3-5%.

The structural thing is that the bond side of the 60/40 stopped working, and nobody is sure when it will start again.

The diversifier that disappeared

In 2022 the Bloomberg Aggregate Bond Index lost roughly 13% while U.S. equities fell 19%. The 60/40 portfolio declined 17.5% — its worst calendar year since 1937 and its fourth-worst in two centuries. The decline itself wasn’t the most consequential part^[1]. What mattered was that the 12-month rolling correlation between stocks and bonds approached +1.0, peaking near 0.80 in mid-2024. The mechanism that had carried balanced portfolios for forty years stopped functioning at exactly the moment investors needed it.

The longer-term picture, traced through the trailing three-year stock-bond correlation across both the Bloomberg Aggregate and the U.S. Government Index, makes the regime shift clearer than any single year of returns does:

From 2001 through 2020, the correlation oscillated between -0.6 and 0.0, doing the diversifying work that the 60/40 was built on. From 2021 forward, it inverts, climbs above +0.6 by 2022, and only begins to fall in 2025. That’s the regime shift in one chart.

Defenders of the classic allocation point out, fairly, that this kind of positive correlation has happened before. Prior to 2000 it was the norm rather than the exception, and 2023’s snapback (the 60/40 returned 17.2%), combined with the gradual normalization of the correlation back toward 0.16 by late 2025, suggests the model is cyclical rather than broken. For an allocator sitting in 2026, though, with valuations elevated, inflation sticky, and central bank reaction functions still uncertain, eventually-correlated bonds don’t help with the year you’re actually in. The practical question is what to hold when bonds don’t diversify.

That is the gap retail flows are filling.