Bad correlation redux

Receive free Unhedged updates

We’ll send you a myFT Daily Digest email rounding up the latest Unhedged news every morning.

Good morning. Last week, Ethan wrote that private equity companies adding more leverage to their portfolios should make us worry less about a debt blow-up than about dwindling returns. Two headlines in the FT over the weekend offered support for this view. One, “Private equity firms pivot away from traditional buyouts,” sat atop a story about how higher rates are pushing leveraged buyout returns down and pushing the big houses towards credit, infrastructure, and insurance investing instead. Another, “Investors devour risky corporate debt in boost to LBO business” pointed out that investors were still clamouring for the high-yield debt that finances buyouts. It’s not a shortage of financing or liquidity that is killing appetite for buyouts. It’s a shortage of high expected returns. Email me: [email protected]. 

Bad correlation II: back to school 

On Friday I wrote about how positive stock-bond correlation has returned since the start of August. Most investors diversity their stock holdings with bonds, but when the two rise and fall together — as they did in much of 2021 and all of 2022 — the diversification benefit disappears. Rising stocks through much of 2023 restored negative correlation, but now stocks are back to falling right alongside bonds. As my daughter would say, this is cringe. 

Several readers wrote in to remind me of something that I learned back when I was studying for the CFA exam, and should have remembered. A portfolio does not need negatively correlated assets to reap diversification benefits; it just needs imperfectly correlated assets. It is enough that the assets are correlated but have different volatilities. Roger Aliaga-Diaz, Vanguard’s global head of portfolio construction, emailed to say that:

As long as the correlation is less than 1 (i.e. perfect correlation), there are portfolio diversification benefits accruing to the investor. In fact, per our calculation, the average stock-bond correlation in the US since 1926 through December 2022 was positive at 0.25 per cent. The diversification benefit of such correlation for instance in a 60/40 stock-bond portfolio over that period was to achieve a 1.8ppt reduction in the portfolio volatility. Of course, if the stock-bond correlation becomes more positive, then there might be less diversification benefits. But as long as the correlation is below 1, still makes sense to diversify your stock portfolio with bonds.

If you systematically reinvest returns in a portfolio with imperfectly correlated assets, you reap benefits even if the correlation is positive. 

The left hand side of Aliaga-Diaz’s chart below shows that over the last 96 years — a period during which the average stock bond correlation was +. 25 per cent — a 60/40 portfolio delivered lower volatility and .44 per cent in additional annual returns just from diversification.

This is an important point, and reminds us not to let positive correlation scare us out of the bond market. But it does not explain why owning a 60/40 portfolio felt so great during the two decades up to early 2020. The reason for that is that bonds were still in a rip-roaring bull market. As Jim Caron, a portfolio manager at Morgan Stanley investment management, told me last week, bonds went up when stocks were down, up, or sideways. That is what we are really going to miss in the next 20 years. After the bond bull market, Caron proposes that investors who are keen to avoid big drawdowns will have to start targeting portfolio volatility by actively adjusting equity allocations as the risk environment changes.

Our regular correspondent Paul O’Brien wrote to make a similar point, saying that “ironically, the higher the correlation between bonds and stocks, the more bonds you may want to own.”  

Others argue that, in a positive correlation world, you should own different assets, rather than just shifting the old portfolio mix. Three years into the pandemic inflation volatility storm, these kinds of arguments are familiar. Most of them, to simplify only slightly, call for higher exposure to commodities, real estate, gold and inflation-indexed bonds.

In a paper about the implications of positive stock-bond correlation from earlier this year, Alfie Brixton and other researchers from AQR take a more nuanced approach. They note that market neutral or trend-following equity strategies might be useful diversifiers, too (they also comment, drily, that private markets “may provide some cushion against short-term volatility due to their lack of mark-to-market pricing, but their diversification potential is limited because they inherit the same underlying economic exposures as their public market equivalents”; amen to that.)

I broadly agree with AQR on this. The problem for investors like me, who need to buy off-the-rack solutions, is that it is hard to find low-cost, transparent products that provide exposure to market neutral and momentum factors, or even commodities, without taking on hard-to-measure idiosyncratic risks, most importantly manager competence. There are interesting offerings from firms such as AQR and Research Affiliates, but it takes more knowledge than even many financial advisers possess to fit them into a portfolio correctly. For many of us, stock and bonds will remain the meat and potatoes.

A final point. More important than the AQR paper’s asset recommendations is what it says about what causes high stock-bond correlations. It is widely understood that positive stock-bond correlation is associated with higher inflation. AQR shows that the relative volatility of growth and inflation, and their correlation, not the levels of growth and inflation, that best explains the correlation. This point has often been missed in recent discussions, which assume that in a higher-inflation, higher-rate regime, bonds must always be a poor diversifier.

One phenomenon that has not driven the SBC [stock-bond correlation] is the secular downward trend in real rates and related richening of both stocks and bonds — the SBC remained negative even as both asset classes experienced this tailwind. It follows that a reversal in the trend — a return to rising yields and cheapening of both asset classes — would not necessarily produce a positive SBC, unless it were accompanied by (or a response to) a sustained rise in inflation uncertainty.

There is probably no such thing as very high yet stable inflation. But moderately higher but steady inflation might be possible, and in such a world, it will still pay to own bonds.  

One good read 

Making money betting against the FTC.