Machine Learning from the COVID-19 Pandemic About the Value of the NYSE Floor in Market Closing Time

Abstract

The spread of the COVID-19 forced the New York Stock Exchange (NYSE) to temporarily shut down its trading floor between March 23 and May 25, 2020. Using a machine learning approach, we investigate the effects of this Covid-19 closure on market quality during the market closing time, 3:50–4:00 pm. Analyzing NYSE- and Nasdaq-listed stocks in the Russell 3000 index for the period of February to June2020, we find that the closure of the NYSE floor has limited impacts on market quality for the NYSE-listed stocks: Percentage quoted spread and spot volatility for the NYSE-listed stocks increase relative to the Nasdaq-listed stocks only in the first three weeks of the floor closure, and it has no impact on consolidated displayed depth for the whole shutdown period. Our findings suggest that the role of the NYSE floor in the market closing time can be replicated electronically.

Keywords: NYSE Floor Closure, Covid-19 Pandemic, Synthetic Control Method, Machine Learning

JEL Classification: G12, G14