Q&A: Getting What You Pay for From Factors
Andrew Dougan, Director of Research and Analytics, at FTSE Russell
In many cases, factors are the dominant driver of portfolio performance. But how a portfolio is constructed – how those factors are included in it so they don’t have unintended effects on it – is vital. Few people have spent as much time considering this challenge as Andrew Dougan, Director of Research and Analytics, at FTSE Russell. As a contributor to many thought leadership and white papers for FTSE Russell, one of which appears in this digital module and is referenced below, Dougan is a leading expert on factor portfolio construction. Institutional Investor recently chatted with him about the paper you can find by clicking on the tab labeled “Paper.”
Academics have known for some time that for suitably diversified portfolios, factors drive risk and performance outcomes. In your paper, how have you gone about turning the very technical discussion that proves they work into one that is more easily understood?
We created two very simple portfolio construction methods. In one there was 50% of the stocks of the starting universe, and the other held all of the stocks. The only thing the two portfolios had in common is the same five factor exposures. What we show is that if you have two very different portfolios, that just happen to have exactly the same factor exposures, they perform almost identically. I think that is rather surprising to many people. We aren’t proving this is correct, as that’s already known – we’re just demonstrating it in a very clear way.
You also seek to demonstrate in the paper that a bottom-up approach is the most efficient way of constructing a factor portfolio. Why is that?
We’ve done a lot of research at FTSE Russell comparing top-down versus bottom-up approaches, and the big picture results show that bottom-up approach probably gives you a better tradeoff between exposures you’re looking for, factor exposures, and things around investability, and diversification.
In a bottom-up approach you consider all the factor characteristics simultaneously – and because you’re more flexible from a bottom up perspective, there’s more ways in which you can create your portfolio, because in the top-down approach you constrain yourself to first creating a quality portfolio, then creating a value portfolio, and so on, and then combining them by averaging. In the bottom-up approach there are many more ways you can construct your portfolio. So, it’s not surprising that you can find a portfolio that is more efficient in terms of the tradeoff between the exposures you’re looking for and things like capacity, diversification, turnover, etc.
The overriding thing is that a bottom-up approach allows you more degrees of freedom to create a more efficient portfolio.
How does FTSE Russell’s tilt-tilt methodology help ensure factor purity?
There is a particular type of approach to creating single factor indices called selection and weighting. You start with a factor – let’s say quality, for example – and you order your universe of stocks by that quality factor and then you say, “I’m going to pick the top 50% to give me exposure to quality.” And then you weight it according to some weightings scheme – it could be equal weighting, fundamental weighting – but the important thing to know is that the weighting scheme will introduce other factor exposures into the portfolio quite apart from quality. For example, if you take the top 50% by quality and equally weight them, you end up with a big exposure to size. So that portfolio that you’ve constructed, is actually not just a quality portfolio, it’s actually a quality, size, and lots of other factor exposures portfolio, because you’re not controlling for those other factor exposures. It’s not a pure factor portfolio. That can be particularly problematic if you were wanting a quality portfolio for downside protection during a financial crisis.
The tilting methodology allows you to create pure factor indices. More generally, tilting allows you to target all sorts of factor exposures. You give me a number of factor exposures – let’s say quality, low volatility, size, momentum, and value – and the exposure you want to them, and I can find you a tilt portfolio which will give you those exposures. In particular, I can find you a tilt portfolio which gives you a quality exposure, let’s say it’s one, and then zero exposure to all of the other factors. That’s a pure factor portfolio, and in the paper, we show that a pure factor portfolio behaves in a way that you expect during a financial crisis or during periods of turbulence.
This is all very important because if you’ve been sold an index based on some exposures, a quality index, you should get what you’re paying for, right? You shouldn’t be getting stuff that you don’t want, i.e. size exposures and uncontrolled exposure to other factors. Factor purity is about transparency – you want to be sure that the portfolio’s behavior is based on what you intended it to be based on, not unintended influences.