What Do Quoted Spreads Tell Us About Machine Trading at Times of Market Stress? Evidence from Treasury and FX Markets during the COVID-19-Related Market Turmoil in March 2020

Abstract

We highlight four main results from our analysis. First, we find that quoted spreads did not merely rise to much higher than usual levels on average, but were also oscillating over a wider range; while at some points within the day spreads were substantially wider than on a typical day, at other times spreads were as narrow as on a typical day. This result suggests that market participants, likely including both dealers and PTFs, became less willing to replenish the order book fast enough to keep quoted bid-ask spreads consistently tight in these markets, which may have amplified the initial shock to liquidity, particularly for the 30-year bond. Second, using alternative volume-weighted quoted spread measures we put forward in this note, we also find evidence that some market participants employing machine trading strategies were nonetheless able to mitigate the increase in trading costs—by executing trades when bid-ask spreads were relatively tight. Third, we further document that liquidity conditions in Treasury futures markets deteriorated less than those in Treasury cash markets. Fourth, we find that quoted spreads in other liquid futures markets, such as those for major foreign exchange (FX) rates, increased notably more than most of the Treasury futures market (the exception being the 30-year Ultra Treasury bond futures contract). Our findings thus point to the systemic nature of the market strains in March, extending beyond differences in market structure and participation. Our findings also underscore the potential benefits from the use of sophisticated algorithmic execution strategies that can adapt rapidly to changing liquidity conditions.