The world of ETF academia is a rarefied one. But the spats it hosts are no less heated than those in the better known field of macroeconomics.
This month, a paper by a quartet of academics from three business schools — Columbia, Chicago Booth and Pennsylvania — set off a row about the niche issue of the market liquidity of the baskets of securities that underpin corporate bond exchange traded funds.
A similar dispute had followed the publication last year of a paper by two other academics — from the University of Notre Dame and the Bank for International Settlements — on a related subject. Each paper came to different conclusions, and each provoked disagreement.
Such disputes are in some ways surprising. ETFs are supposed to be ultra-simple products — cheap, easy-to-use funds that invest in a broad range of shares, bonds or other assets, spreading risk for inexpert investors. But as with so much of modern finance, the reality is very different from the appearance.
This particularly applies to corporate bond ETFs, which have proved enormously popular — worth more than $1.2tn, compared with barely $10bn in 2009. They deviate from the original ETF principle quite significantly, and in complex ways that average investors are unlikely to grasp.
The complexity is underpinned by practical, but problematic, reasoning. Corporate bonds, especially high-yield ones, are far less liquid than equities. To mitigate the problem of a mismatch between an instant-access ETF and illiquid underlying investments, ETF designers typically use a small number of bonds as proxies for the whole portfolio.
But this alchemy is not without risk. In normal times, the price of the ETF should reflect the underlying value of the bonds. Investment banks aim to facilitate that both as so-called authorised participants, creating new ETF shares or dissolving them, or as mainstream market makers matching buyers and sellers
It doesn’t always work. This was particularly evident when large gaps emerged between the pricing of ETFs and the underlying value of the bonds back at the height of the Covid-19 pandemic in early 2020. This was due in part to the illiquidity of the underlying bonds. But the mechanisms designed to counter that may have magnified the distortion.
Another anomaly during this market stress was that investment-grade bonds, and investment-grade ETFs, often traded less smoothly than higher-risk, less liquid bonds and linked ETFs. Again the experts of academia and the market disagree on the reasons. But again the reasons matter less than the effect, which is that bond ETFs are not what they seem: in effect they are derivative products, reliant on complex structures that may be open to abuse.
The combination of structural latitude and thin liquidity suggests, in extremis, a potential risk akin to the way the Libor interest rate benchmark was manipulated: bank traders who collaborated to set Libor rates abused the mechanism, advantaging their own financial interests. When ETF structurers pick proxy bonds, they may drive down certain bond prices while protecting others.
There are two additional reasons to be concerned. First, the corporate bond market itself will be severely tested over the coming period as economic pressures in many countries weigh on companies’ balance sheets. Second, the market may be rattled by the unwinding of the vast economic policy experiment that was quantitative easing.
Some central banks, notably the US Federal Reserve, actually bought corporate bond ETFs as part of their QE. These were in tiny quantities (the Fed bought $14bn relative to a total balance sheet of $9tn), so resale caused negligible market impact. But the unwinding of the broader QE programme could yet harbour stability risks for corporate bonds too: disarray in markets rarely respects asset class boundaries.
At the same time, the old systemic safeguards have been weakened. For perfectly good reasons, tough bank capital rules were introduced after the 2008 financial crisis. But that has made capital-intensive activities, like large-scale market making, unattractive, further damaging liquidity. As a 2020 Brookings Institution paper highlighted, dealer positions in corporate bonds slumped from $300bn in 2006 to less than $50bn in 2019.
Even if the coming stresses do not cause markets to melt down, the corporate bond sector, and corporate bond ETFs in particular, will remain a worrying systemic vulnerability.
patrick.jenkins@ft.com