Before we jump in this week, I wanted to let you know that a recording of my presentation on Evaluating Alpha Signals at last month’s Wolfe Research Global Quantitative & Macro Investment Conference has been posted. Please don’t hesitate to reach out if you have any questions or comments. Now on to this week’s topic...

Portfolios were rocked this past week as the world heard of news that Pfizer and BioNTech have produced an effective COVID-19 vaccine that is now in the final stages of testing. While the Nov 9 announcement indicates that there’s a light at the end of the long coronavirus tunnel, it caused massive market rotations as investors moved out of the stay-at-home stocks that had previously been skyrocketing throughout the global quarantines. This rotation was mainly realized through the Momentum factor, which saw the highest daily drawdown in longer than a decade. The Momentum drop - close to 500 bps according to the Axioma US4 Medium Horizon risk model - was the largest daily drawdown in the factor’s recent history and translated to a 19-standard deviation event.

But perhaps even more interesting was the Momentum behavior in the subsequent days. Nov 11 saw a 5-standard deviation positive reversal in Momentum but by Nov 13, the factor was back on the downward trajectory with a 3.5-standard deviation down move.

Let’s take a look at the characteristics of Momentum to help us understand what drove these big moves and how this factor may be kicking up dust in portfolios.

Momentum’s DNA

Similar to analyses we’ve run in the past, we used the Russell 3000 universe and built a market-neutral high-minus-low (HML) Momentum portfolio. The long side of the portfolio is equal-weighted across all stocks with a Momentum exposure greater than 1 and the short side is equal-weighted across all stocks with a Momentum exposure less than -1.

A look at the sector breakdown of the portfolio highlights some of the intuition around what has been driving the Momentum factor this year and particularly during the Nov 9 drawdown.

During the March drawdown as COVID-19 lockdowns were instituted around the world, Pharma & Biotech names drove the factor’s long tilts. Unsurprisingly, traditionally low momentum names in the Energy sector drove the short tilts. The anti-COVID Retail sector was also in the top short exposures. However, by Nov 9, Retail had switched sides of the portfolio, going from one of the top short exposures to one of the top long exposures.

Interestingly, with the exception of Retail and Banks, the sector composition of Momentum was not drastically different in November compared to March. This points to an intuition around the COVID-related theme in Momentum, with the same types of stocks that carried the factor earlier in the year remaining consistent and likely contributing to the downturn in the factor last week.

The fundamental exposures of the Momentum portfolio also provide some insight. Of course, this portfolio has very high exposure to the Momentum factor, but we can see other factor tilts in the portfolio as well. Momentum stocks tend to be heavy growth names and anti-value, and this shows through in the Growth and Value factor exposures for the portfolio.

Though it’s intuitive that the portfolio would have these exposures, we can see that its exposure to Growth is the highest in recent history. Though Value is not at its lowest exposure ever within the portfolio, it is trending towards a low-point. The traditional high growth, low value Momentum stock is becoming even more extreme in those characteristics.

What Do The Correlations Say?

It’s no secret that Momentum tends to have a high positive correlation with Growth and a high negative correlation with Value. However, in the post-COVID era, these correlations are becoming even more extreme.

The correlation between Momentum and Growth factors, as estimated by the Axioma US4 Medium-Horizon risk model covariance matrix, is at the highest level that it has been in over a decade. Similarly, the Momentum and Value factor correlation is at an all-time low point. This has major implications for growth and value investors who have likely had to work harder than ever to manage extreme Momentum exposures in their portfolios.

Is Momentum Crowded?

Using the Wolfe Research QES US Broad risk model on our Momentum portfolio reveals that there is potential crowding happening within the factor. Wolfe’s risk models track several crowding measures, one of which is Short Interest. We can see in the chart below that the Short Interest exposure for our Momentum portfolio is at an all-time high and currently sits far above the historical average.

This level of short interest exposure is a likely driver of the unprecedented swings we saw last week in high momentum stocks. Breaking the Momentum portfolio into the long and short sides allows us to dig into this insight further.

The current highpoint of Short Interest exposure is driven by the long side of the portfolio. The long side has a current exposure of 0.23, compared to the short side which has an exposure of 0.10. The short side, which corresponds to low momentum stocks, has had negative exposure to the Short Interest factor since mid-2013 - a trend that reversed starting in late September 2020. This tells us that the sensitivity of high momentum stocks to short interest has been steadily increasing, especially in 2020, while the sensitivity of low momentum stocks has been steadily decreasing. The trend in the Short Interest exposure, particularly the spike in 2020, is shows that the Momentum factor may have been suffering from crowding. This was likely a major contributor to the crash in Momentum and an indication that the Short Interest factor exposure can help signal potential factor rotations.

What Can We Do About Momentum?

Hopefully the worst of the Momentum move is over, but if this last week has taught us anything, it’s that this factor needs to be watched closely. Investors may do well to look for ways to better manage the exposure to this factor while we ride out the Momentum wave. We’ve recently explored ways that factor-targeting baskets can be used to avoid undesirable exposures and protect portfolios against big market swings. If you’re interested in exploring these methods for your portfolios, please reach out to discuss further.

US Market: 11/09/20 - 11/13/20

• Monday’s big rally on the Pfizer and BioNTech news yielded a mixed week for US markets, as the Dow (+4%) and S&P (+2.2%, hitting a new record on Friday) surged while the Nasdaq (-0.6%) ended slightly lower.
• Cyclicals and “Reopen” industries such as Theaters, Advertising, and Travel saw big gains, while growthy Tech and “Stay at Home” names fell out of favor.
• The University of Michigan’s consumer sentiment index fell to 77 in early November from 81.8 in October, coming in lower than consensus estimates of 81.8. With COVID cases exploding nationwide, the risks of more stay at home orders and business restrictions remain very real.
• 10-year Treasury yields rose to +0.97% in the past week (a seven-month high), due to optimism around the perceived vaccine breakthrough and subsequent implications for an economic rebound.

Normalized Factor Returns: Axioma US Equity Risk Model (AXUS4-MH)

Methodology for normalized factor returns

• Value entered positive normalized space as out-of-favor stocks enjoyed big gains around the Pfizer vaccine trial announcement.
• Market Sensitivity and Volatility both saw strength given the week’s risk-on trend. Market Sensitivity crossed into positive normalized space at +0.14 SD above the mean.
• Profitability hit a recent peak of +0.16 SD above the mean on 11/10 and started to revert back down towards the mean.
• Earnings Yield saw a slight move down as it continued to revert after peaking at +1.53 SD above the mean on 10/8.
• Growth was a casualty of the week’s trading, as it fell by -0.35 SD and headed lower, towards Oversold space.
• As discussed in greater detail above, Momentum was by far the week’s biggest loser as it fell by over one full standard deviation in just five trading days.
• US Total Risk (using the Russell 3000 as proxy) declined by 22bps.

Normalized Factor Returns: Axioma Worldwide Equity Risk Model (AXWW4-MH)

Methodology for normalized factor returns

• Size factor was once again the week’s biggest winner globally, up +0.59 standard deviations.
• Similar to the US, Volatility and Market Sensitivity both enjoyed solid normalized gains due to the risk-on trade.
• Earnings Yield reversed course after recent weakness, crossing into positive normalized territory after hitting a recent trough of -0.21 SD below the mean on 11/6.
• Value rallied on a normalized basis, up +0.29 SD and inching closer to an Overbought label at +0.88 SD above the mean.
• The decline in Growth was even more pronounced worldwide vs. the US, as the factor feel by -0.82 standard deviations and approached Oversold space.
• The Momentum crash didn’t spare the global market, as the factor fell by -1.13 standard deviations and now sits squarely in Oversold territory.
• Global Risk (using the ACWI as proxy) decreased by 36bps.

Please reach out if you’d like to further discuss this week’s major factor moves, analyze your portfolio’s exposure to Momentum, and see how Omega Point can help you mitigate your risk from these factors.

Regards,
Alyx