Trading under Crisis: A Study of Stock Pricing in the COVID Markets

Abstract

Over the course of 2020, the COVID-19 pandemic significantly impacted the stock market and resulted in extreme volatility. These dramatic moves exposed the fact that the stock market's prices failed to reflect the appropriate value of many companies, and in a more general sense, they seemed to show evidence that relationships and correlations between stocks changed. This thesis seeks to find a pricing model that can predict the fair price of stocks in the United States market despite the volatility. In particular, we will use a linear regression reversion model to predict the stocks' relative moves against an index and then trade against oversized or undersized moves. We hope to test two different hypotheses using this technique. First, we want to test whether this type of model can trade stocks profitably in situations before and after the pandemic's effects were felt. In addition, we also want to check whether these models change significantly over time in terms of performance or coefficients, which would be evidence of changing relationships among different stocks. In the process, the trades that the various linear regression models execute can shed light on the market's efficiency in the pandemic period, as well as the amount of noise in different stocks' returns resulting from idiosyncratic issues. The analysis resulting from this research will ultimately provide a sense of market conditions during the pandemic and illustrate whether this unprecedented period in the markets is actually as different as it seems.