Mark Zmijewski on The Temporary Effect of COVID-19 and Riskiness of Stock Prices

The Temporary Effect of COVID-19 on the Riskiness of Stock Prices

Did the pandemic change the long-run riskiness of stocks, or did it distort the way we measure it? The data point clearly to the latter.

The COVID-19 pandemic produced one of the most pronounced short-term disruptions in the modern history of U.S. equity markets. During the seven-week period spanning the weeks ending February 28, 2020 through April 9, 2020, the S&P 500 Index recorded weekly returns ranging from approximately -15.0% to +12.1%, including a -11.4% week ending February 28, 2020 and a +12.1% week ending April 9, 2020. These movements were sufficient to alter standard estimates of equity risk for many publicly traded companies. As the data discussed below demonstrate, however, much of the resulting change in measured risk was temporary rather than a permanent change.

This article addresses a specific question relevant to valuation, litigation, and financial analysis: did the pandemic change the long-run riskiness of stock prices, or did it instead introduce a temporary distortion into the statistical measures most commonly used to quantify that risk? The evidence indicates the latter for a broad set of companies. For analysts and decision-makers who rely on beta-based cost of capital estimates, distinguishing the two is essential.

Measuring the riskiness of a stock

Systematic risk, the component of a stock’s total risk that cannot be eliminated through diversification, is commonly measured by beta. Beta measures the statistical relation between a company’s historical stock returns and the returns of a broad market index, typically the S&P 500.

Because beta is derived from observed returns rather than expected returns, any unusually large return observation in the estimation window, whether driven by company-specific events or by a market-wide shock, can affect the estimate of beta. When the estimation window is short (for example, 104 weekly return observations, a common convention), individual extreme observations carry meaningful weight. This estimation property is central to understanding why the pandemic produced large but ultimately transitory changes in measured riskiness.

The onset of COVID-19 as a temporary market intervention

A temporary intervention in a time-series statistical relation occurs when an external force causes the estimated relation to change abruptly for a period of time, after which it reverts toward its prior relation. The COVID-19 pandemic was such an intervention. The shock was concentrated in a narrow seven-week window, during which weekly market returns were several standard deviations from their historical norms, and during which different sectors of the economy experienced sharply divergent return patterns based on their exposure to pandemic conditions.

The relevant analytical point is that the shock altered the joint distribution of company and market returns for a discrete period. Companies whose businesses were adversely affected by pandemic restrictions (travel, lodging, in-person services) experienced large negative returns concurrently with the sharp market declines, producing a temporary increase in measured beta. Companies whose businesses benefited from pandemic conditions (online retail, certain technology services) experienced relatively smaller declines or positive returns concurrently with market declines, producing a temporary decrease in measured beta. In neither case did the change necessarily reflect a permanent shift in the company’s long-run systematic risk.

Illustration: Amazon.com Inc. and Hyatt Corporation

The temporary nature of the COVID-19 effect on measured riskiness can be illustrated by comparing two companies in industries that were affected in opposite directions: Amazon.com Inc., in the online general-merchandise retail industry, and Hyatt Corporation, in the hotel industry. Both betas below are estimated using rolling 104-week (two-year) windows of overlapping weekly returns from January 2019 through December 2024.

During the onset of the pandemic (the weeks of February 28, 2020 through April 9, 2020), Amazon’s estimated beta declined from approximately 1.4 to approximately 0.8. Over the same weeks, Hyatt’s estimated beta rose from approximately 0.8 to approximately 1.6. These are large movements, on the order of a 40% decrease and a doubling, respectively, occurring within seven weeks.

Both estimates then remained at relatively the same level for approximately two years. This stability is an artifact of the 104-week rolling estimation window: once the pandemic-onset weeks were embedded in the estimation window, they continued to influence the estimate until they were no longer included. By mid-June 2022, when the pandemic-onset weeks fell outside the 104-week window, both betas began to revert toward their pre-pandemic levels. Amazon’s beta returned to approximately 1.4, essentially its pre-pandemic value. Hyatt’s beta declined to approximately 1.0, modestly above its pre-pandemic level of approximately 0.8, a residual difference that, according to the underlying analysis, is at least partly attributable to an increase in Hyatt’s financial leverage during the period rather than to a change in the riskiness of its underlying operations.

The reversion itself is the key evidence: the change was a temporary measurement effect, not a long-term change in the underlying riskiness of the stocks.

Conclusion

The COVID-19 pandemic produced large and economically significant changes in measured beta for many publicly traded companies. The evidence for representative companies in adversely and favorably affected industries, Hyatt Corporation and Amazon.com Inc., indicates that those changes were predominantly temporary. Once the pandemic-onset weeks exited the rolling estimation window, both companies’ betas reverted close to their pre-pandemic levels. For purposes of estimating a long-run, forward-looking measure of the riskiness of a stock, the temporary effect of COVID-19 on beta should be identified and, where appropriate, excluded.

About Professor Mark Zmijewski

Mark E. Zmijewski is the Charles T. Horngren Professor of Accounting Emeritus at the University of Chicago Booth School of Business. He is a Senior Consultant to Econic Partners and the principal of Z Consulting Group, LLC. He writes on valuation and financial economics at markzmijewski.com.