US stock-bond correlation has been negative for much of the past 20y. However, regime change – related in part to fiscal and monetary policies and the broader macroeconomic landscape – can occur, with implications for portfolio performance and construction.¹

Indeed, US stocks and bonds fell sharply in tandem during the first three quarters of 2022, damaging the performance of balanced public-market portfolios. While simultaneous large declines in stock and bond prices are likely temporary, a positive stock-bond correlation regime may persist. In this paper, we consider the portfolio management challenges of a positive stock-bond correlation world – a world unfamiliar to many CIOs. 

To be sure, positive correlation impacts portfolio performance. Shifting to a positive stock-bond correlation world will cause a balanced portfolio of stocks and bonds to deliver more volatile performance with a wider set of potential long-term outcomes – including more extreme tail events and deeper max drawdowns.

But the virtues of diversified portfolio remain intact even in a positive stock-bond correlation world (so far as they are not perfectly correlated). When correlation is positive, a balanced portfolio of stocks and bonds remains optimal, with asset weights only slightly different than the optimal allocation when stock-bond correlation is negative. As such, being right or wrong about future correlation is only moderately beneficial or costly in terms of risk-adjusted returns (for all but the most extreme correlation assumptions).

In fact, when correlation is positive, the optimal allocation to stocks may be only slightly different relative to when correlation is negative. In some cases (perhaps for a 60/40 investor!), it is optimal to lean into stocks when correlation is positive, with the incremental return from increased stock ownership sufficient to offset the incremental risk from positive correlation. 

Given the modest cost of being wrong about future correlation, investors need not rush to change their correlation assumption and re-optimize their asset mix until a new correlation regime becomes established. A slow adjustment of the forward correlation view, or even assuming correlation to be zero, may be a reasonable way to balance these tradeoffs when correlation is uncertain.