Did Corporate Bond ETFs Exacerbate the Panic?

Did Corporate Bond ETFs Exacerbate the Panic?

A recent academic study suggests yes.

Reviewed by: Sean Daly
Edited by: Sean Daly

Nearly $200 billion flowed into fixed income exchange-traded funds last year, and BlackRock Inc., the world’s largest ETF issuer, expects assets in these funds to triple to $5 trillion by 2030.  

While seemingly a good sign for the ETF industry, a new academic study has a warning: ETFs can take the liquidity out of the underlying corporate bond market during market panics.  

The report’s findings, seemingly at odds with prevailing views, has triggered a public debate about liquidity and ETF construction. 

The paper, titled “Steering a Ship in Illiquid Waters: Active Management of Passive Funds,” argues that ETFs, despite being seen largely as passive index trackers, do a lot of “active” managing behind the scenes because of the illiquidity of the underlying investment.  

This is a “trade-off between index tracking and liquidity transformation,” the authors wrote. 

To manage index deviations, ETFs depend on authorized participants to conduct arbitrage trades. These APs create and redeem shares in exchange for baskets of bonds the “creation basket” and “redemption basket,” respectively. These baskets can deviate substantially from the underlying index, and the APs are hard at work adjusting those baskets dynamically.  

To reduce the cost of all that trading, the baskets only include a fraction of the bonds represented by the underlying index. This use of “fractional baskets” is what the paper means by “liquidity transformation.”  

With that AP mechanism under the hood, the ETF purrs during normal conditions. The enhanced liquidity and apparent price discovery feel great to the investor. The chosen bonds for inclusion in the basket also benefit, as they are made more liquid in normal times.  

The problem with fractional baskets is during liquidity shocks. During a crisis, with investors selling hand over fist, redemptions submerge creations, and those bonds included in the basket enter a crush zone.  

The authors point to the wave of selling in March 2020, when those bonds represented in redemption baskets became heavily represented in APs’ inventory. They lose their most natural buyers, the market makers, as the APs become reluctant to purchase more of the same bonds. 

If the paper is correct, the ever-growing corporate bond ETF sector could, during panics, enact suffocating pockets of illiquidity upon the underlying fixed income market.  

Bond ETFs’ market share in the fund industry has grown from 14% to 24% since 2017, as more investors are blending bond ETFs with active strategies, reframing the traditional 60/40 portfolio and bond construction in the process.  

Other academics and quite a few practitioners, however, take the opposite view. 

In a paper titled “ETFs, Illiquid Assets, and Fire Sales,” the authors suggest that the APs act as a very effective buffer, mitigating the risk of fire sales as they incur mark-to-market losses on existing inventory or add to inventory.  

Whereas mutual funds may be forced to sell illiquid assets to satisfy large redemptions that could trigger a fire sale, bond ETFs can absorb large redemptions without transmitting the selling pressure to the more fragile bond market because of slow NAV adjustment. So the issue might be more about mutual funds. 

Michael John Lytle, CEO of Tabula Investment Management, suggests that the study may potentially have mis-attributed the trends that it saw in the data.  

“It is not that the APs were expressing a preference, but that the issuers were having to adjust their baskets in order to preserve the tracking error of their funds,” he noted. 

Broader Issues 

This discussion leads to a few bigger issues that were at work beyond ETFs. Dealer net positions in corporate bonds fell from $300 billion in 2006 to less than $50 billion in 2019. The pandemic exposed just how reduced the dealers have now become in their ability to offer market liquidity.  

Simultaneously, the ever-growing corporate bond market has seen a transition to greater ownership by mutual funds. Investment-grade corporate bond mutual funds ownership rose from $738 billion in 2008 to $2.159 trillion in 2019, just before the COVID-19 crisis.  

This development can add to the sell-off during a drawdown. Since the NAV of open-end mutual funds is set end-of-day, without addressing any price impact from that day’s investor redemption, investors feel the need to rush to the exit immediately to avoid bearing the illiquidity costs from high redemptions. This aspect of open-end mutual fund construction may have also impacted the March 2020 sell-off, perhaps more than the ETF fractional basket issue.  

A few solutions have been suggested to expand illiquidity in the corporate bond market. Nellie Liang, previously at the Brookings Institution, suggests redemption restrictions related to how quickly bonds can be sold, in order to better align the offer of liquidity by the mutual funds to the liquidity of the underlying assets, giving dealers more flexibility to provide market-making without jeopardizing the bank’s far better post-GFC safety regime, and dispersing deal-making to the hundreds of smaller shops could mitigate stresses during a panic.  

Given the large redemptions, the vulnerability of investment-grade bond mutual funds was the biggest surprise of March 2020. In the days before the PMCCF/SMCCF announcement on March 23, the spreads of investment-grade bonds and high-yield bonds both spiked, but the ratio of the investment-grade bond to high-yield bond spread was 25% higher than typical. This occurrence suggests some of the dislocation in investment grade was not related to rising credit risk.  

It’s unclear whether this anomaly came from ETFs, mutual fund selling or the types of sellers panicking, but the issue will continue to be an important one for investors to keep an eye on.  

Sean Daly writes on ETFs, biotech and wealth management. He was educated at Columbia University and has taught international finance, computing and financial risk management at Pace, Tulane and Princeton. Follow him on Twitter (X) via @Sean_Daly_.