At the start of the Covid pandemic, a ‘dash for cash’ saw even the safest and most highly liquid assets experience large price declines as liquidity deteriorated in financial markets, and the cost of trading increased.
Our recent analysis showed that price data from ETFs, such as NAV discounts and bid-ask spreads, could help OEFs to more accurately identify the cost of trading underlying assets. OEFs base their NAVs on the prices of the underlying assets they hold, and this can become harder to do accurately in stressed market conditions. So information provided by ETF price data could help OEF fund managers to calculate price adjustments more accurately, especially during a shock.
During the ‘dash for cash’, market intelligence suggests that ETF NAV discounts also signalled the future direction of NAV movements in OEFs, suggesting that OEF NAVs were slower to incorporate price information. The market intelligence indicates that ETF prices more closely reflected ‘real’ trading costs than reported spreads on underlying assets, which OEF NAVs were based on.
We used a regression model to test this hypothesis. We found that in normal market conditions, and consistent with efficient market pricing, ETF NAV discounts provide no information about future changes in OEF NAVs.
However, during the ‘dash for cash’ stress period, there was a correlation (Chart A). A 1% ETF NAV discount was associated with a 0.28% negative change in expected OEF NAV movements on the following day.
Chart A: ETF NAV discounts offered price information for corporate bond investors in March 2020
Expected change in OEF NAV following a 1% ETF NAV discount on the previous day
- Sources: Bank and Financial Conduct Authority survey responses, Bloomberg Finance L.P., Refinitiv Eikon from LSEG, Morningstar and Bank calculations.
- Results from a regression model, which controlled for other factors (including corporate bond prices). Various specifications were tested, and regressors have been checked for high correlations, including in specific time periods. The relationship between ETF NAV discounts and OEF NAVs during the ‘dash for cash’ holds across further relationships, including controlling for the proportion of highly liquid assets in the OEF, the inclusion of the VIX, the inclusion of additional ETF NAV discount lags, and is insensitive to whether the NAV is adjusted or unadjusted (ie whether or not we account for the ‘swing factor’). Errors are clustered-robust on the date and fund level.
Our regression model estimated the impact of the median sterling corporate bond ETF NAV discount on expected change in next day sterling corporate bond OEF NAVs. We looked at this relationship in normal market conditions, during the lead up to the ‘dash for cash’ and during the ‘dash for cash’. Our null hypothesis was that, when controlling for broader movements in sterling corporate bonds, the ETF NAV discount should provide no additional information about OEF NAVs.
Our results supported the argument that ETF NAV discounts offered price information for corporate bond investors during the ‘dash for cash’, and that OEF NAVs were slower to incorporate price information. In future stress periods, if investors anticipate that OEF NAVs are ‘stale’ and slow to adjust, this could trigger a first-mover advantage, which could contribute to the sort of market dysfunction seen in March 2020.
This post was prepared with the help of Owen Lock and James Semark.
This analysis was presented to the Financial Policy Committee in June 2021. This analysis was previously published in the July 2021 report on market-based finance.
Share your thoughts with us at BankOverground@bankofengland.co.uk.