Measuring COVID-19 Effects with Sector Index Event Studies

Summary

The document asks whether CAPM-style abnormal-return analysis can be applied to sector indices, then warns that a comparison during the COVID-19 period alone may confuse the event with longstanding sector performance. Technology had outperformed and energy had underperformed before the pandemic, so estimated alpha for those sectors during 2020 could reflect earlier patterns rather than a new shock.

To isolate a change associated with the event, the answer proposes adding a post-February indicator and a slope interaction with market returns to the analysis. The indicator captures a shift in the intercept, while the interaction allows the relationship between market returns and sector returns to differ after the event. The response is conceptual: it provides no dataset, fitted estimates, statistical tests, or detailed event-window choices. Results would depend on model specification and the selected comparison period, and the suggested design does not by itself establish that COVID-19 caused any observed change.

Key ideas

• Sector indices can be analyzed with CAPM-style abnormal-return methods.
• Preexisting sector trends can make event-period alpha misleading.
• A post-event indicator can represent an intercept shift.
• An interaction between market returns and the post-event indicator can represent a changed market sensitivity.
• The suggested model needs empirical testing and does not by itself prove causation.

Tags

# Event study using sector indices


# Event study using sector indices












Analyzing Covid-19's impact on different sectors I would like to use sector indices. Can you use CAPM or similar to calculate abnormal returns of indices or does it only work with stock prices?

## Answer by demully (score 1)

https://quant.stackexchange.com/a/58901

Sure. Your CAPM will tell you there's significant positive alpha in Tech and significant negative in Energy! Which makes total economic sense...

Except do the same for the last 5 years excluding 2020; and you might/will get largely the same result. Tech boomed; and Energy sucked for years before Covid ;-)

So if you're serious about measuring the Covid effect, put in a "dummy" (1 if after February this year) and a "slope dummy" variable into your analysis (ie market returns if post-Feb, 0 if before). These will tell you how Covid has changed the Tech-loving, Energy-hating behaviour of the markets in the Covid-period versus similar before ;-)

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.