Blog Posts — November 11, 2020

Active Risk Pops on the Heels of a Big Day for Factors

In our November 10 blog post “Pfizer Vaccine Announcement Puts Momentum in the ICU” we showed that on November 9, many factors across multiple geographies produced returns that were two or more standard deviations away from their long-term averages. No factor’s returns were as outsized as Medium-Term Momentum’s, where the eight-to-10 factor standard deviation negative returns were the factor’s worst ever in many regions. And in some regions, such as the US, the return on November 9 was twice the magnitude of the next-worst-day since at least 1999. We attributed the factor’s reversal of fortune to the resurgence of stocks that will likely benefit from the reopening of global economies, such as hotels, airlines and realty companies.

High magnitude returns, of course, lead to an increase in factor volatility that is likely to drive active risk higher. A reversal in returns will lead to changes in correlations between factors that will also drive changes in active risk, although the direction of that change will be determined by the sign of the portfolio’s active bets.

To get a sense of this potential impact, we calculated the active risk (using the AXUS4 short-horizon fundamental model) of several STOXX Axioma USA 900 factor portfolios and found that a few did indeed see a substantial jump in risk, as expected. The active risk for the STOXX AX USA 900 Momentum portfolio jumped 75 basis points from November 9 to 10. Similarly, active risk for the Value portfolio rose 45 basis points on the first day and another 20 the following day. Other portfolios in the same family (e.g. Quality, Low Risk and Multifactor) saw more modest increases.

Exhibit 1. Predicted Short-Horizon Active Risk, STOXX Axioma USA 900 Factor Portfolios

Source: Qontigo

A deeper dive into the correlation matrix[1] for the US model shows substantial surges in volatility for several factors, with the biggest increases in Dividend Yield, Growth, Market Sensitivity and especially Medium-Term Momentum, where volatility increased more than 45% proportionally, compared with its level at the end of October.

A few factor pairs also saw relatively large changes in correlation—although not as large in magnitude as when the market turned down and factors went haywire in February-March of this year. Most notably, Profitability (with its negative return) saw the biggest correlation changes, becoming more correlated with Momentum and Growth (which also had negative returns) and less correlated with Dividend Yield and Market Sensitivity (which moved in the opposite direction).

Exhibit 2. Changes in Short-Horizon Volatility and Correlation, October 30 to November 10, 2020

Source: Qontigo

As was the case earlier this year, Monday’s moves seemed to be largely industry-based. We noted in the blog post that the market seemed to react as if we would all be traveling and returning to offices soon. Some industries experienced big moves on November 9, and therefore also saw volatility go up along with substantial changes in correlations with other industries, which may also have driven changes in active risk.

Although we have focused on factor performance, because of changes in industry volatility, industry correlations and, of course, unintended factor bets, many portfolios beyond just those investing in factor-based strategies may have also seen a sudden and substantial change in active risk. If the market is right and we are now seeing a turnaround in fortunes for parts of the economy, these changes are unlikely to revert any time soon.


[1] This is the predicted correlation matrix from the risk model. See your Qontigo representative for more details.