Soviet-era Polish cinematography is often a source of seemingly absurd catchphrases repeated for generations. “How much sugar is in your sugar” is a classic one from the quirky professor in the 1973 comedy Man-Woman Wanted. When we target particular factors within our equity exposures, I increasingly find myself taking on the role of the professor as I try to answer the question “How much factor is in my factor?”. It might seem like an odd question but we can answer this by relying on simple factor definitions and a holistic approach to combining factors. It’s only once we know what our true exposures are, that we can consider how we avoid any unintended secondary exposures that have the potential to sour the overall outcome.
The Polish catchphrase on sugar sounds like a question I hear time and time again in the context of FBI. How much 'value' in my value exposure? How much 'quality' in my quality index? And can a single-factor index actually be a multi-factor strategy in disguise?
So…how much 'factor' is in your factor exposure?
Investors interested in factor investing have a big task on their hands – to navigate a vast and often confusing universe of strategies, all claiming to provide you with ‘true’ factor exposure. This often leaves investors confused and debating a variety of questions:
How do I define my factor? How much factor exposure do I need? And if I adopt a multi-factor exposure, will my value stocks get along with my quality names and my low volatility securities or will they only dilute the exposures I already have?
What selection criterion you use for a given factor, for example value, will ultimately depend on your investment beliefs. Yet many investors expect these selection criteria to be broadly correlated. One could measure the portfolio’s style biases using the traditional ‘style skyline’ where each bar represents the significance of the particular bias – the higher the bar, the stronger the bias – relative to the benchmark where securities are weighted according to their market capitalisation.
A potential investor might expect the value index to look like the chart below, with the height of each ‘value’ bar broadly similar, representing a similar bias to book-to-price, dividend yield or EBITDA to enterprise value, and a negligible bias to other factors.
In reality, it is a bit more complicated than that. As soon as you select ‘value’ stocks for your portfolio, these names will inevitably bring along a whole range of incidental exposures in terms of their profitability, historical volatility, market-cap or momentum. As a result, the profile above can be challenging to achieve without complex optimisation techniques. Instead, a value index may actually look more like this:
As we can see, there is a positive bias to value but it is far from uniform across different valuation ratios. This might reflect choices that were made in designing the value index and, as we highlighted in the recent Diversified Thinking piece, these choices may have implications for the index performance too.
For the actual value index shown above, the index provider seems to favour book-to-price, the ratio used by Eugene Fama and Kenneth French in their seminal paper on equity risk factors. The value index also give a negative bias to profitability, one of the ‘quality’ characteristics. On top of that, it moves away from the largest stocks in the market-cap benchmark and adds a small size bias to your exposure as well.
This might not necessarily be the outcome you were looking for. With a starting point of looking to add a value factor element to your portfolio, the index ends up leaving you with a whole bunch of other exposures you might not have initially considered. And this sort of pattern is not unique to a value factor. Similar behaviour can be observed across other factors as well.
Then comes the question of concentration. How much factor exposure do I actually need? Should I consider a more broadly diversified mix of value stocks or opt for a smaller selection? The latter could potentially increase the value bias in my portfolio but it will again alter the incidental biases elsewhere and leave more stock-specific risk.
Finally, in the same way that you can’t expect the FTSE 100 to stay the same, investing in factors is hardly a ‘set-and-forget’ type of strategy. The index may rebalance regularly to meet its objective but its ‘strength’ in terms of the factor bias (intensity) and its interaction with other factors is likely to evolve over time. For example, currently a broadly diversified value index doesn’t seem to exhibit a momentum bias, with the momentum score close to zero. However, over last 10 years it could have provided you with a significant negative or a significant positive momentum, depending on the stage of the market cycle.
That’s why any factor exposure needs to be carefully managed and monitored over time.
A multi-factor approach is a potential way to control for your overall factor exposure by looking at it in a more integrated and holistic way. This allows you to keep in mind the problem of factor interaction but manage it by adjusting the individual factor weights. I will be covering some different solutions to this approach in a future post.
The professor never found a definitive answer to his sugar conspiracy theory. Luckily in the world of FBI we do have the tools to monitor and manage how much value your value strategy actually gives. So before you jump on the ‘factor’ bandwagon, make sure it takes you to the destination you envisaged for your portfolio.