Negative beta stocks: The case of Zoom

One of the questions that comes up every time I teach the Capital Asset Pricing Model (CAPM) in a basic finance course is whether there are any negative beta stocks, and if so what would be their expected return. My standard answer has been that negative beta stocks are a theoretical possibility but possibly non existent in practice. Every time I have found a negative beta in practice, there was either a data error or the sample size was too small for the negative beta to be statistically significant. I would also often joke that a bankruptcy law firm would possibly have a negative beta, but fortunately or unfortunately, such firms are typically not listed. (The answer to the second part of the question is easier, if the beta is negative, the expected return is less than the risk free because it hedges the risk of the risk of the portfolio and one is willing to pay for this hedging benefit).

But now there is an interesting real life case of a negative beta stock: Zoom Video Communications, Inc. Not only is this a large company by market capitalization, but it is also a familiar company with so many online classes taking place on Zoom. During the Covid-19 pandemic, a plausible argument has been going round why Zoom should have a negative beta. The argument is that if the pandemic rages, the economy collapses while Zoom soars, and if the pandemic retreats, the economy recovers, and people go back to face to face meetings, and the Zoom boom is over.

Interestingly, the data supports this nice theory:

• The US declared Covid-19 to be a national emergency on March 13, 2020 a couple of days after the World Health Organization (WHO) declared it to be a global pandemic. From March 13, 2020 till August 21, 2020, Zoom has a beta of -0.39, and it is statistically significant (t-statistic is -2.09, and p-value is 3.97%). All betas in this post are relative to the S&P 500 index.
• This was a dramatic change from the pre-pandemic days. In the period from its IPO in April 2019 till the end of 2019, Zoom had a beta of 1.81 (with a t-statistic in excess of 5 and a p-value less than 0.01%). (The numbers are similar if we extend the end date till the Italian lockdown on February 23, 2020). In a few weeks, the high beta stock became a negative beta stock.
• More interestingly, Zoom appears to be going back to a high positive beta stock once again. I used July 1, 2020 as a break point and obtained the following results for two sub-periods since the pandemic began.
• March 13, 2020 to June 30, 2020: beta is -0.43, t-statistic is -2.21 and p-value is 2.9%.
• July 1, 2020 to August 21, 2020: beta is 2.66, t-statistic is 2.41 and p-value is 3.7%.

A better example of beta changing dramatically (going from around two to negative and then back to around two) within a few months without any change in the business mix of the company would be hard to find.

Negative betas may be a once in a 100-year event (the last global pandemic of comparable severity was in 1918), but the Zoom example illustrates the importance of estimating betas more carefully using shrinkage estimators and Bayesian methods as I explained in detail in a blog post ten years ago.

Posted at 5:37 pm IST on Sun, 23 Aug 2020         permanent link

Categories: CAPM

Comments