During the COVID-19 crisis, municipal bond ETFs insulated the underlying bond market from the era's fire sale-type selling pressure, but that came with a cost: A large deviation between muni ETF prices and the funds' net asset values.
That's according to Yu He, author of a July 9 paper entitled "Stress-Tested: Municipal Bond ETFs During Market Turmoil," which he presented during the 15th Annual Municipal Finance Conference, held at the
In his paper, He, who in August is set to join James Madison University as an assistant professor of finance, looked at why muni bond ETFs experienced large and persistent deviations from NAV during the COVID-19 market turmoil.
In previewing results revealed in his paper for audience members attending a July 21 session on muni market functioning, which was also live-streamed, He said one thing the results showed was that arbitrage by authorized participants to close the gap between muni ETF prices and NAVs "was limited when it was needed most."
In his paper, He said while ETFs "are designed to track NAVs through in-kind creations and redemptions executed by APs, muni ETFs saw some of the most severe and prolonged pricing dislocations across all ETF categories in March 2020."
Periods of financial distress typically spur increased demand for liquidity, with investors looking to sell their most liquid assets first. In contrast to individual muni bonds traded in fragmented over-the-counter markets with limited liquidity, muni ETFs "provide an exchange-traded mechanism for accessing liquidity," the paper said.
"Consequently, ETF shares experienced substantial selling pressure during the COVID turmoil, leading to sharp declines in market prices," the paper said, adding that, meanwhile, NAVs didn't decline as sharply, "resulting in large and persistent deviations between ETF prices and their reported NAVs."
While that liquidity-driven selling pressure should have presented arbitrage opportunities, the NAV deviations persisted despite significant price gaps, which suggests "that APs did not engage in arbitrage to the extent expected," the paper said.
To examine the drivers of ETF pricing dislocations, He, who earned his Ph.D. in finance from the University of Nebraska-Lincoln, looked at trading volume in ETF-held bonds prior to and during the COVID-19 crisis.
"If arbitrage by APs were functioning effectively, I would expect to observe increased trading activity in ETF-held bonds during the crisis, as APs redeem ETF shares and offload the underlying bonds into the market," he said in the paper.
While the results showed that trading volume in ETF-held bonds did increase during the crisis period, the increase was concentrated in smaller trades, those with a par value below $100,000.
The concentration of trading volume growth in smaller trades indicates "that selling pressure in ETF-held bonds did not primarily flow through institutional-sized transactions," the paper said.
"Instead, activity was fragmented, likely reflecting a combination of retail selling and smaller-lot transactions related to ETF portfolio adjustments," the paper said.
Though he wasn't ruling out the possibility of arbitrage activity occurring in smaller trades, "the lack of growth in large trades indicates that ETF redemptions, if they occurred, were not accompanied by block-level execution in the underlying bonds," He's paper said.
"This further supports the view that selling pressure remained concentrated in the ETF secondary market rather than being transmitted to the underlying bond market through the arbitrage mechanism," the paper said.
To gauge whether fund level pressure "left a footprint" in the underlying market, He constructed a "bond-day measure of Fund Pressure" that aggregated return-adjusted net outflows from all funds holding a given muni bond, weighted by their pre-crisis holdings.
Mutual funds address redemptions by selling underlying assets, so outflows transmit pressure directly to portfolio bonds, the paper said. ETFs by contrast don't sell bonds to meet investor sales of ETF shares.
"Primary-market redemptions occur only when APs voluntarily exchange ETF shares for the underlying basket; therefore, ETF pressure reaches the bond market only if APs actively redeem and unwind baskets," the paper said. "Limited AP arbitrage implies little or no transmission."
Ultimately, He concluded in his paper that while ETFs protected the underlying market from fire sales, they also permitted pricing dislocations to continue, exposing investors to those pricing dislocations.
"The results highlight how structural friction can affect ETF pricing in fragmented and illiquid markets such as municipal bonds," the paper said.
