This blog was first featured on the Wells Fargo Asset Management blog on March 11th, 2020.
As the world becomes increasingly concerned with the economic fallout related to the COVID-19 virus, global equity markets have sharply retreated. The chart below plots worldwide Google search activity for “coronavirus” relative to the performance of the MSCI World Index. Not surprisingly, the index performance is negatively correlated with the rise in searches for coronavirus, with the recent spike in searches inversely mirroring the steep drop in the MSCI World Index.
Since the market began pulling back on February 19, low volatility has been the best-performing factor by a wide margin, as seen in Figure 2, with quality a distant second.
Figure 3 plots the performance of five groups of securities based on their predicted beta*, with quintile one (Q1) representing low-risk securities and quintile five (Q5) representing high-risk securities. As the market sharply fell, the spread between low- and high-risk securities was roughly 10%.
Beta relative to the market is generally a good proxy for risk. However, the team at Analytic Investors further decomposes beta into three sources:
- 1. Risk that comes from country
- 2. Risk that comes from industry
- 3. Risk that comes from the stock itself, independent of country and industry
Each measure offers a glimpse into how investors are pricing risk in the current market. As shown in Figure 4, companies and industries with low sensitivity to the market have offered a potential safe haven since mid-February. However, while low country beta performed well at the onset of the outbreak, once it became clear that the coronavirus would be a worldwide phenomenon, allocating to lower-risk countries did not offer strong downside protection.
Another interesting observation is how risk models are adapting to the current market environment. Figure 5 plots the Barra risk forecast of each industry at two points in time. The horizontal axis plots today’s risk forecast while the vertical axis plots the forecast as of the beginning of the year. Barra’s model uses backward-looking measures of risk, such as country, book to price, and capitalization, and shows little reaction to the changing risk environment, despite the Chicago Board Options Exchange Volatility Index increasing by almost 350% since the beginning of the year.
Figure 6 plots the average risk of each industry using two risk models at a single point in time. The horizontal axis again represents Barra’s risk forecast today. However, the vertical axis plots Analytic’s proprietary risk forecast. This forward-looking model incorporates measures such as changes in implied volatility; news sentiment; credit default swap spreads; and environmental, social, and governance score. Whereas the risk forecast in the Barra model is about the same as at the start of the year, the Analytic forecast has significantly increased in the current environment, especially in higher-risk industries.
History shows us that having a diversified view of risk, both in the measures used as well as the tools employed, may give a more complete picture. This is even more critical in market environments where the sources of risk are rapidly evolving and have not been seen in prior periods.
Read our recent blog post, The Downside of Beta >