Introduced three years ago, Qontigo’s proprietary ROOF Scores quantify and track the direction and strength of investor sentiment. Our Weekly ROOF Highlights1 provide timely quantified insights into the sentimental state of markets, i.e., are they trending bullish, bearish or neutral?
The value of such insights is self-evident. But what if this same methodology could be used by active investors to evaluate the bullishness or bearishness of their own portfolios vis-à-vis their benchmark? In other words, what if investors had a tool to ensure that the portfolios they construct are sentimentally aligned with both the manager’s intent and the benchmarks? A tool that by definition enhances the portfolio-construction process by exposing—and enabling the fine tuning—of factor exposures impacting that sentimental alignment?
Thanks to the efforts of Qontigo’s Applied Research team, this is now possible.
As a framework for our discussion, we will make use of the bottom-up factor-based methodology of the Sector ROOF Scores2 to analyze active portfolios against their benchmarks, by computing their active ROOF Scores. The advantage of this methodology is that it allows us to analyze any portfolio, not just the market portfolio. In fact, we use this same methodology to construct Risk-On and Risk-Off portfolios that seek to capture the market’s reward for being bullish or bearish3. The active ROOF Score described in this note is simply the quantification of the implied active sentiment of the portfolio versus its benchmark, i.e., is the portfolio implying a more bullish or bearish view?
To demonstrate this novel approach and to highlight the insights obtainable from this analysis, we used the “lightly constrained” long-only factor portfolios constructed by the Applied Research team to approximate a typical institutional approach to factor investing in the US market. These portfolios aim to maximize the exposure to the target factor, subject to the following constraints:
- Long-only and fully invested.
- Minimized exposures to non-target factors (+/- 0.2 standard deviations).
- Tight active-sector bounds to achieve near sector-neutrality (+/- 2%).
- Maximum tracking error to the STOXX ® USA 900 benchmark of 3%.
- Month-end rebalancing with no liquidity or turnover constraints.
The Sector ROOF Score methodology defines the characteristic of each of the 11 GICS sector portfolios—risk-tolerant or risk-averse—in terms of their exposures to eight fundamental style factors in the Axioma Short-Horizon Fundamental Risk Model (see table below). Of course, it is rare for a sector to exhibit all the characteristics that would define it as risk-tolerant or risk-averse, so we use a proportional allocation based on the overall balance of scores.
We extend this methodology into the active space by measuring the active exposure of a portfolio to these same eight style factors versus the portfolio’s benchmark, thus computing the portfolio’s active ROOF Score to determine if the manager is implementing a more bullish or bearish strategy than the benchmark consensus.
If the portfolio’s active ROOF Score is significantly higher (>20%) than the benchmark’s, it implies that the portfolio manager is more bullish than the average investor in the benchmark. Conversely, if a portfolio’s active ROOF Score is significantly lower (<-20%) than the benchmark, it implies that the portfolio manager is more bearish than the average investor in the benchmark. If the portfolio’s active ROOF Score is insignificantly different from the benchmark (+/-20%), this implies that the manager does not have a strong directional view, i.e., that the manager is neutral. The chart below depicts the three implied zones of active sentiment described above. We will use it to look at the implied confidence and directional bias of our sample active factor portfolios.
US Momentum Portfolio
The Momentum factor portfolio is an interesting starting point for our analysis because price momentum can be either bullish or bearish, depending on which types of stocks are currently in favor. Essentially, exposures to the Momentum factor alone do not tell us much about the investment thesis behind them, and performance simply quantifies its popularity with investors, i.e., are the stocks with positive momentum ones that are typically viewed as being bullish or bearish?
Looking at the active Momentum factor portfolio through the lens of the ROOF methodology provides much needed insight as to the directional bias driving price momentum at different times in our history, in this case the period immediately preceding the COVID crisis through the current period (see chart below).
Zone A in the chart above reflects a period when the active ROOF Score of the Momentum portfolio pointed to the implementation of an increasingly bullish strategy, whose active ROOF Score rose from being 30% more risk-tolerant than the benchmark in January 2020, to up to 60% more by March 2020, and averaged 50% more bullish than the benchmark during the five-month period. This implies that price momentum was being driven by the popular implementation of a confidently bullish bias, which caused the Momentum portfolio to deviate significantly from the benchmark portfolio during that period.
Zone B represents the period when the COVID-19 pandemic created a huge crisis of confidence in economic forecasts. This sudden surge in uncertainty drove investors back to the safety of the benchmark, and the active ROOF Score of our Momentum portfolio was driven by a lack of directional bias or at least a lack of consensus on what the directional bias should be, i.e., there may have been an equal number of investors implementing bullish and bearish strategies, but no dominant (popular) bias. We see that the active ROOF score of the Momentum portfolio during the rest of 2020 pointed to a continued lack of confidence among investors, despite unprecedented levels of monetary and fiscal stimulus. During that period, directional bias remained mostly neutral versus the benchmark.
Zone C captures the inflation tantrum of February 2021, when the lack of a policy response from the Fed seems to have convinced most investors that inflation was a bigger problem than Fed Chair Jerome Powell acknowledged and that the central bank may have been behind the curve on this issue. Investors felt confident enough about this bearish outlook to deviate substantially from the benchmark by implementing strategies that were, at times, some 50% more bearish than the benchmark going into the Q2 earnings season. The strong Q2 earnings results and repeated assurances by CEOs that inflationary pressures were manageable and would not affect profit margins helped decrease the confidence in bearish forecasts. Investors, in turn, unwound their bearish bets by the end of July 2021, favoring more neutral strategies.
Zone D, from August through November 2021, included negative news about inflation and supply-chain bottlenecks, balanced against a strong Q3 earnings season with over 80% of companies reporting stronger-than-expected results (according to FactSet), and CEOs reassuring investors that inflation was not impacting their profit margins. The Federal Reserve also reiterated its forecast of a temporary inflation picture, and announced that they would slowly unwind their asset-purchasing program, before contemplating raising interest rates in the second half of 2022. This was enough to make investors question their previously negative views and start to unwind their bearish bets, adopting a more neutral bias in the process.
A disciplined investment process entails the implementation of a portfolio that is the best representation of the manager’s investment thesis. This means that a manager implementing the Momentum factor portfolio described in our example above must have been confidently bullish in phase A, confidently bearish in phase C, and too uncertain in phases B and D to deviate from the risk appetite of the benchmark with any confidence. However, if their views did not support the portfolio’s expected market direction, they may have wanted to rebalance or hedge their bullish (bearish) exposures.
Using the same methodology, we analyzed four other factor portfolios: Growth, Value, Low Volatility, and a Multi-Factor (Value + Momentum + Profitability + Earnings Yield – Size – Volatility) portfolio. As per the ROOF-style factor classification, Value, Low Volatility, and Multi-Factor all suggest the implementation of active bearish strategies, i.e., they were more risk-averse than the risk appetite implied by holding the benchmark portfolio. The analysis served to confirm the consistency of this directional bias and that the portfolio-construction rules used to implement it remained consistent throughout our history, i.e., we observed no significant style drift. Keep in mind that unlike the Momentum portfolio, these portfolios explicitly comprise alpha factors that are used to define the ROOF scores, so their ROOF scores, especially for the Multi-Factor portfolio, should not be surprising. However, in all cases the degree of implied bullishness or bearishness varied substantially.
For example, the Low Volatility factor portfolio is always representative of a bearish strategy and the confidence level implied in its active exposures since January 2020 ranged from 30% to 80% more bearish than the benchmark, with an average of 60% for the period (see chart below).
Next, we looked at our Value factor portfolio. A strong Value tilt in our ROOF methodology is associated with a bearish sentiment, although it can sometimes simply reflect the forecast that cheaper stocks will outperform the benchmark. Because the ROOF methodology involves eight different factor tilts, a tilt on Value alone as a strategy to outperform the market is not strong enough to lift the overall active ROOF score into positive territory, if the other exposures are all bearish. The chart below shows the active ROOF Score for our Value portfolio since January 2020. It implies that the implementation of a strong Value strategy (average active exposure to Value factor was >1 StDev), consistently reflected a more bearish bias than the benchmark during the period.
The multi-factor portfolio mostly tilts towards styles that are more often found in bearish portfolios, according to the ROOF classification. The presence of a Momentum tilt can also lead to period when a bullish sentiment is present, as we saw earlier, meaning that, at times, we would expect the Momentum factor to ‘contradict’ the overall bearish bias implied by the other factor tilts in the portfolio. The same is true for the slight small-cap bias in this strategy. The chart below shows the active ROOF Score for our multi-factor portfolio since January 2020, and confirms that its implementation reflects a mostly bearish bias versus the benchmark, and highlights periods of ‘disagreement’ between some of the factors, i.e., Size and Momentum, which show up on the chart as periods when the bearish bias was weakening.
Finally, we looked at our Growth factor portfolio. A Growth tilt is defined as the implementation of a bullish bias in the ROOF classification. As such, we expect to see the active ROOF Score of our Growth portfolio to imply a consistently more bullish stance than the benchmark. The chart below confirms our assumption with an active ROOF Score for the Growth portfolio that implies the implementation of a consistently more bullish bias than the benchmark.
In this note, we described an extension of our ROOF Score methodology that can be applied to the analysis of active portfolios. The goal of this analysis is to quantify and track the implied sentiment bias of the portfolio, as defined by its portfolio construction and resulting active factor exposures. The benefit to portfolio managers is to ensure the alignment of their stated investment thesis with the portfolio’s implied directional bias. A manager with a bullish view should construct a portfolio with a strongly positive active ROOF Score, and vice-versa. If directional bias is part of your strategy, then the ROOF Score can ensure that bias also becomes part of your implementation process and keeps you in alignment with your stated goals. Alternately, if your implementation suggests an uncomfortable bias relative to the market, you may want to think about reigning in bets or hedging that bias. The latter is a perfect use case for our Risk-On / Risk-Off ROOF portfolios.