Value investing may have spent the last decade on the sidelines of the broader bull market, but Financials are an exception. This week, we pivot back to our spotlight on Wolfe Research’s sector-focused risk models and uncover some astounding factor premia that have been driving the Financials space.
As we’ve noted in our recent posts on the Wolfe US Energy and Wolfe US TMT models, sector models isolate a more concentrated universe and allow us to focus on the drivers of risk and return relative to the sector in question vs. a broader universe that may dilute important sector-specific effects.
We’ll see in today’s analysis that the sector models not only allow us to extract sector-specific factors, but they also provide a mechanism by which we can surface otherwise shrouded factor premia within sectors.
Before we dive into our analysis, let’s first review the methodology used for the Wolfe Research QES US Financials (“Wolfe US Financials”) risk model. Analogous to the other sector risk models, the Wolfe US Financials risk model is constructed using an estimation universe of US all-cap financials securities, as opposed to using a broad US universe, as is the case with traditional broad market risk models. In addition to factors that are common across other sector and broad market models, the financials-specific model includes several factors relevant for understanding the behavior of the financials sector. This covers fundamental factors, such as Re-investment Rate, Return on Assets, and Receivables Turnover, macro factors, such as sensitivity to Term Spreads, High Yield Credit Spreads, and Interest Rates, and granular industry factors, such as Life & Health Insurers, Asset Managers, and Mortgage REITs.
While the model’s estimation universe is the financials sector, the model is able to cover all US equities due to the time series regression methodology used to estimate the security sensitivities to the factors. Please see the risk model factsheet for more information on the construction and factors used in this model.
We’ll also be using the Wolfe Research QES US Broad (“Wolfe US Broad”) risk model as a comparison for the sector model. The estimation universe for the broad market risk model represents a US all-cap universe that covers all sectors.
Hunting for Factor Premia? Look No Further than Financials
Though broad market risk models are very powerful tools for characterizing the systematic factors that move the general market, they can sometimes have the adverse effect of obscuring systematic premia that exist in within specific segments of the market.
To see this effect in action, we can look at the 12 month returns of the factors from the Wolfe US Financials model compared to the Wolfe US Broad model.
12 Month Style Factor Returns - Wolfe US Broad Risk Model
12 Month Style Factor Returns - Wolfe US Financials Risk Model
In general, factors with absolute returns greater than ~3% on an annual basis can be thought of as having meaningful factor premia. If we look at the returns of the style factors in the Wolfe US Broad model, we see that there are only 7 factors with meaningful positive or negative premia. In contrast, the Wolfe US Financials model shows 11 factors with meaningful premia and the magnitude of returns for these factors are generally much higher. If you were trying to access factor premia, the financials space would have offered you phenomenal returns for 2020 and 2021.
One factor in particular sticks out as having a much higher premia in the financials model compared to the broad model. In the former, Book to Price has a 12-month return of over 18%, while in the latter, Book to Price only has a return just shy of 2.5%. Looking over a longer period of time back to January 2007, we see this disparity widens even further, with the factor’s return in the Wolfe US Broad model at -3.6% while the return in the Wolfe US Financials model is > 80%.
For a long/short market-neutral equity factor, a premia of 80% is extremely strong. This trend is even starker when put in the context of a strategy that otherwise underperformed in the broad market.
Industry Characteristics of Value in Financials
The above observations prompted us to look at a portfolio of securities representing the Book to Price factor within the Wolfe US Financials model to determine whether any other factor biases exist in this strategy. borrowing from our usual “HML” analysis, we constructed a market-neutral high-minus-low (HML) portfolio based on this factor using the Russell 3000 universe. The long side of the portfolio is equal-weighted across all stocks with Book to Price exposure greater than 1 and the short side is equal-weighted across all stocks with Book to Price exposure less than -1.
Of interest in this analysis is the industry representation in our Book to Price HML portfolio. Sometimes, HML portfolios can show heavy industry concentration, especially when breaking out the long and short sides of the portfolio separately. However, in the case of the Book to Price HML portfolio within the financials model, the industry exposure is relatively unbiased. With the exception of some periods where Mortgage REITs have extreme exposure, namely 2010-2011, there is not a single industry that consistently dominates the portfolio. This tells us that the Value strategy within financials worked across more granular industries for the most part. Interestingly, the Banks factor is the one industry that has gone in and out of favor in the portfolio.
Our analysis shows that though Value investing might have stopped working in the broad markets, it certainly did not stop working in financials. However, this conclusion may have been challenging to identify without the help of a sector risk model focused on helping us isolate market effects specific to sectors.
If you are interested in further understanding how the Wolfe US Financials model or any of the other Wolfe sector models can help you develop strategies that may have otherwise flown under the radar, please reach out to us for additional details.
US & Global Market Summary
US Market: 03/01/21 - 03/05/21
- The markets saw another volatile week, with tech names continuing to suffer as the Nasdaq ended down -2% while the S&P 500 was up almost 1%.
- Interest rates remained front and center, as 10Y Treasury yields landed at 1.56% on the week, stoking the rotation out of tech and into cyclicals.
- February saw more jobs created than expected in the US, as the Labor Department reported a 379,000 increase in non-farm payrolls in Feb (vs. forecast of +182,000), after adding 166,000 in January. The US economy has now regained 12.7 million of the 22.2 million jobs lost due to the pandemic.
- The Senate passed President Biden’s $1.9T relief plan on Saturday in a party-line vote after an all-night session.
Normalized Factor Returns: Axioma US Equity Risk Model (AXUS4-MH)
Methodology for normalized factor returns
- Value was again the week’s biggest winner, moving up +0.74 standard deviations and entering Overbought space at 1.1 SD above the mean.
- Market Sensitivity also continued its rebound, and now sits .01 standard deviation away from its historical mean.
- Size is also now in line with historical trend at +0.03 SD above the mean.
- Volatility saw a large downward move, with normalized return falling 0.54 standard deviations and now at -1.63 SD below the mean.
- Momentum fell by 0.64 standard deviations after losing its Overbought label last week and now site perfectly in line with historical trend at 0.
- Growth continued its rapid descent, falling over a full standard deviation to -1.27 SD below the mean. This factor recently peaked at+1.49 SD above the mean on 2/5.
- US Total Risk (using the Russell 3000 as proxy) decreased by 11bps.
Normalized Factor Returns: Axioma Worldwide Equity Risk Model (AXWW4-MH)
Methodology for normalized factor returns
- Earnings Yield, not Value, was the week’s winningest factor in the global model, moving up past the mean and now sitting +0.54 standard deviations above it.
- Value also enjoyed a strong upwards move, getting closer to Overbought territory at +0.88 SD above the mean.
- Volatility felt ongoing weakness, falling by a quarter of a standard deviation to -0.69 SD below the mean.
- Momentum fell by 0.64 standard deviations and is now at -0.43 SD below the mean.
- Growth collapsed by over a full standard deviation, and is now an Oversold factor at -1.02 SD below the mean.
- Global Risk (using the ACWI as proxy) also declined by 11bps.