Q&A with Harin de Silva, Ph.D.
361 Capital: What sort of factors should advisors consider when building portfolios? And how much exposure to those factors should advisors be thinking about when implementing?
Harin de Silva: You can get factor exposure through a traditional active manager or through a factor-based ETF but either way, you should know what you have. That’s the first step. The second part is asking what you should own. That’s a harder question because there are two schools of thought.
Should you own an equal proportion, from a risk standpoint, of each factor? There are five or six general factors to think about (e.g., Value, Momentum, Low Vol, Size, Quality). To build a diversified portfolio, you’ll want to own all of these factors in some proportion which you’ll then need to manage and rebalance to keep your exposure constant.
The other way to do it is to vary the exposure over time. What we’ve observed is that investor behavior tends to move in waves. Value will go out of favor for a long time, then momentum will lose favor for a while. There may be a benefit to adjusting your factor exposure based on what’s in vogue.
Both approaches have merit. Maintaining constant exposure requires less turnover. Turnover is higher when you’re managing the size of your factor exposure and it may not be appropriate in a taxable account. The method you use is a function of your philosophy and what tools you have available to execute the strategy.
361 Capital: So factor exposure could be dynamic, but that’s really a timing decision. What about going through economic cycles. Certain factors seem to outperform in different phases, right?
Harin de Silva: Right. Value tends to do well in the early stages but poorly in the late stages of the economic cycle. So, if you’re adjusting your value exposure in a portfolio, you want to own the cheaper stocks when the market’s accelerating. In the mature stage or in a recovery, you want more emphasis on quality and on growth.
361 Capital: You’ve been managing factors in your portfolios for years now so you’ve seen some momentum in factors. Most investors are familiar with price momentum, but factor momentum may be a new concept. Can you explain this concept?
Harin de Silva: Factors tend to have momentum in a one- to three-year timeframe. After three years, factors tend to mean revert. This is a really interesting time to look at value factors, because value generally hasn’t worked well for the last three to five years. Value, though, historically exhibits strong performance over the longer run. If I were running a factor allocation portfolio right now, I’d say that value looks attractive from that standpoint.
361 Capital: Let’s talk about measuring factor exposure. Is there a standardized methodology for calculating exposure, or are there different techniques to gauge the strength of a factor?
Harin de Silva: It depends on the factor. The factor most people are aware of is their small-cap exposure. That’s very easy to measure. You can see the weighted average market cap for all mutual funds and ETFs. Low volatility is also easy because you can look at any ETF’s or mutual fund’s beta.
Momentum, I think, is much harder. It’s not easy to get a momentum score for every ETF and mutual fund. There isn’t a lot of agreement on how to measure it, either. The same with quality. Should quality be Return on Equity? Return on Assets? And the same thing, to some degree, with value because you can look at P/E or price-to-book, price-to-cash or price-to-cash flow. Most people, understandably, use multiple measures.
361 Capital: Let’s go back to factor momentum. How do your long/short equity strategies capitalize on factor momentum?
Harin de Silva: The first thing we do is to forecast which factors are going to be in favor, using a combination of factor momentum and mean reversion. That tells us, at a given point in time, how important each factor is. Then, we use that to score all the stocks in the universe. When we build the portfolio, we use an optimization routine to make sure that the highly rated stocks go in the long portfolio and the stocks that are undesirable from a factor standpoint go in the short portfolio.
361 Capital: Have your thoughts on factor momentum changed over time? Did you start with some assumptions only to find that the markets told you otherwise, or have the markets confirmed your initial notions?
Harin de Silva: We primarily used a factor momentum model when we started. Through ongoing research, we realized the downside of the model. If a factor has been performing, there’s a tendency to load up on it. There’s no governing mechanism to stop the loading. Now, I’m a big believer in factor momentum, but it needs to be combined with a mean reversion component. What we should be doing if something’s working better and better is to actually start reducing our exposure to it.
We also learned that it’s really important to make sure we have good exposure to all of the individual characteristics that determine a particular factor. We don’t, for example, define value simply as price-to-book, because cash flow and dividends and all these other things are important as well.
361 Capital: Have you found the risk of mean reversion varies by factor?
Harin de Silva: We haven’t. In fact, if you look at the mean reversion component and measure it statistically, you’ll find it’s the same in the U.S. as in the rest of the world. It’s the same in Japan, it’s the same in emerging markets and it’s the same in global small caps. I’ve come to believe it’s due to investor behavior rather than economic cycle. Why? Because the economic cycle in Japan has been very different than that of the U.S., yet you see the same relative contribution of momentum and mean reversion in both countries.
361 Capital: It seems we’re getting finer and finer factor slices defining stock returns. Fama and French started out with a three-factor model, which then became four- and five-factor models. With that in mind, do you think alpha still exists, or can we lay returns all at the feet of factor exposure?
Harin de Silva: I don’t think it’s all factor exposure. Think of the active manager who finds a stock with financials that look good from a valuation standpoint. But then, upon digging further into the numbers, he or she discovers the company’s payables are rising because the firm’s factoring purchases. It’s that kind of analysis which brings value to the table. I think there’s always going to be a role for this kind of traditional active management.
Still, advisors and clients need to think carefully about what they’re paying for. You shouldn’t pay for an active manager just because he or she has a value bias. You can get that cheaply. But if he or she adds something useful on top of a simple bias, that’s worth paying for.
That said, what active quants can provide is the capacity to adjust factor exposure. And among these, someone who’s managing portfolios using stocks as building blocks can do it more effectively than someone who’s using factor ETFs.
361 Capital: So, do you see factor-based investing moving towards finer and finer slices?
Harin de Silva: I actually see it going the other way. Eventually, you’re going to see more and more people just buying multi-factor ETFs. People will say, “I’m going to compare this multi-factor ETF with that multi-factor ETF, and the one I buy is going to be my core factor allocation. Then I’m going to hire active managers or long/short equity managers around that.”
361 Capital: Okay. But for the quants shopping for a better way to explain stock returns, is there room for more factors?
Harin de Silva: I can see people launching factors based on positive versus negative news on a company. We don’t really have agreement yet on an accounting complexity factor, but I think that will also come into being at some point.
361 Capital: Ostensibly, those could then be incorporated into multi-factor portfolios. I mean, if those factors are robust over time, then the universe of factors being harvested by these multi-factor portfolios is going to grow.
Harin de Silva: Right. With an ETF, there’s a tendency toward “buy-and-forget.” But you shouldn’t just forget. Suppose you have a three-factor ETF now, and two years hence, a five-factor ETF is launched. Your clients really need to go from three factors to five, but that’s hard to do without generating a taxable event. Factor exposure in a traditional mutual fund format is more malleable.
361 Capital: You talked about the pendulum swinging towards multi-factors. Meantime, there are new factors that may be included in a multi-factor portfolio. Can we get too many factors?
Harin de Silva: There’s a fair amount of research that has to be done to convince everyone that a factor is new and different. Having said that, I think the challenge with some of this new data is that it doesn’t go back that far. Circling back to the accounting complexity factor idea, accounting data goes back to 1920, so if you have a debate on its validity, you can resolve it with a simulation and performing a bunch of tests. But if you’re trying to use news analytics or using a machine learning algorithm to read the news on a company, that data set doesn’t go back that far. That factor becomes harder to test. We’ll have to see how that sorts itself out going forward.
361 Capital: So wrap this up by telling us your thoughts on the future of factor-based investing. You’ve told us you expect a move towards broad, multi-factor portfolios. Is there anything else that we should be looking at?
Harin de Silva: I think in five years you’ll see much clearer labeling of equity and even non-equity products in terms of their tilts. In my fantasy, it would be like a nutritional label, telling you “This comes with these tilts.” That will make it easier for people to assemble portfolios. It’s not clear to me how we get there because we don’t have standards yet. But if we can reach that point, much like Morningstar helped with style boxes, it would make things easier for the investor and the RIA communities.