Thu, 3 Nov 2016
For most investors, it pays to maintain a static allocation to factors and be broadly diversified, says Morningstar's Alex Bryan.
Christine Benz: Hi, I'm Christine Benz for Morningstar.com. Could good timing be the key to successful factor investing? Joining me to discuss some research on that topic is Alex Bryan. He is director of passive strategies research in North America for Morningstar.
Alex, thank you so much for being here.
Alex Bryan: Thank you for having me.
Benz: Let's start by talking about what factor investing is before we get into your research on the timing of various factors. What do we mean when we talk about factor investing?
Bryan: Factor investing is basically targeting securities with characteristics that have historically been associated with better performance in one way or another. These include characteristics like low valuations, small market capitalization, high profitability, and strong momentum or strong recent performance. Each of those characteristics has helped explain stock returns over the long term, so the stocks that have stronger characteristics along each of those dimensions have historically offered higher returns. So, factor investing tries to take advantage of this historic relationship under the assumption that there is a real effect here that would lead you to expect these characteristics to predict future performance going forward.
Benz: OK. So, is past performance from these various factors predictive of what investors should expect from them going forward? The fact that a certain factor has led to outperformance, should investors take that to the bank and assume that it will hold up going forward?
Bryan: Well, I think it's important to understand why each of those factors has worked in the past and to use that to set your expectations for whether or not you think it's going to work going forward. I think each of those factors that we discussed, there is some economic merit as to why the effect has persisted in the past.
Now, some research that's been put up by Research Affiliates has suggested that it may not be a good idea to simply take that return premium from the past and to extrapolate that going forward. So, in the same way that we wouldn't set our return expectations for the stock market based on what's happened over the last 20, 30 years, we should also be a little bit more skeptical when we try to form a forward-looking view of how we think these factors should do going forward.
Rising valuations can often play a big part of returns in the historical data. So, if you look at the stock market, for example, from 1980 to the last few years, the valuations for most stocks have expanded quite a bit. Valuations cannot expand indefinitely. So, most people who are trying to set expectations going forward will strip out the effects of changing valuations.
The same can be said of factor investing. If factors are becoming more expensive, so for example, if stocks that are highly profitable have outperformed because they became more expensive relative to stocks that had poor profitability, well, that valuation expansion is probably not going to continue indefinitely. And so, Research Affiliates tried to disentangle how much of the return was coming from this valuation expansion or contraction, and how much of the return was coming from a true structural, persistent effect.
Benz: So, the basic Research Affiliates thesis was that, yes factors can be predictive, but only if you buy them at the right price, essentially?
Bryan: That's their argument, yes. They say that when factors become cheaper, that's the time to buy, and when they become more expensive, that's the time to maybe hold off on that. Valuations matter.
Benz: So, in ETFInvestor you took a look at this issue using your own research and using indexes for which there are ETFs to see if you could discern some similar findings or opposite findings. So, let's talk about, Alex, what you did, and the conclusions that you drew from your own research.
Bryan: Sure. So, I looked at several different factor strategies that investors can access through index ETFs. And I looked at the valuations of those indexes compared to stocks that had opposite characteristics or relative to the market.
Benz: And how did you measure valuation?
Bryan: I looked at price/book and price/earnings, which are two pretty common metrics. And then I tried to see, OK, could that valuation spread between, for example, value stocks and growth stocks predict how value stocks would perform relative to growth stocks over the next five years after looking at that valuation.
What I found was that in fact for things like value and size, the valuation spreads did matter. The cheaper value stocks became, the better they tended to do relative to growth stocks. Same was true of small-cap stocks relative to big stocks. So, I looked at the data from 1987 to this year, so a pretty good dataset, slightly different than what Research Affiliates did. But I found that the evidence was weaker for some other factors. So, for example, momentum--there wasn't a clear relationship between valuations and future performance, which is not surprising given the high turnover of the strategy. Also, for the low-volatility effect, I did not find a significant relationship; neither did Research Affiliates, incidentally.
Benz: Right. So, there has been a lot of interest in that low-volatility strategy in particular. What conclusions would you draw from the research that you did and what Research Affiliates did on that topic? How should investors think about approaching that area?
Bryan: So, Research Affiliates argued that a lot of the return benefit from owning low-volatility stocks has come from valuation expansion. And if you look at the valuations of these stocks relative to how they have traded over the last few decades, they are expensive compared to where they have been in the past.
So, Research Affiliates says the reason that there is not this apparent relationship between valuations and performance is that these stocks have become gradually more expensive and that mean-reversion has not yet taken place. Now they caution that that's a risk going forward.
I would say, yes, investors shouldn't simply extrapolate the past returns going forward, but if you are a more risk-averse investor, I think there is still a pretty compelling case to owning these stocks because even after taking out this valuation expansion, these stocks still did better on a risk-adjusted basis, and I think that's the important point here. If you are risk-averse, the strategy can still make sense, but perhaps you shouldn't just look at the last 30, 40 years and extrapolate that going forward.
Benz: So, the broad conclusion that you drew after crunching your own numbers and doing your own research is that because the correlation with some of the factors and future and performance isn't all that tight that investors shouldn't try to get too fancy in terms of timing their purchases of various factors, that maybe you are better off bundling the factors together?
Bryan: That's right. So, valuations do matter. There is a relationship, at least for low-turnover factor strategies like value, like size. But the relationships are weak and they are influenced by extreme events like the tech bubble. If you take the tech bubble out of the back-tested period, the relationships become a lot weaker. So, in most situations where valuations are not extreme, it doesn't make sense to time or if you do time, the benefit is likely to be relatively modest and I think that benefit is outweighed by the cost because you are giving up some diversification benefits when you do try to time your different factors and you incur some transaction costs with taxes and trading expenses and so on.
So, I think for most investors it really makes sense to figure out which factors you believe in over the long term and to maintain a static allocation to those factors and be broadly diversified. None of these factor strategies is going to pay off all the time, but I think by being broadly diversified across the different factor strategies, you can improve your likelihood of sticking with that plan and succeeding over the long term and I think that's a better plan than trying to move in and out based on valuations.
Benz: OK. Alex, I know this has been a hotly debated topic in your area of the world. Thank you so much for being here to share your research with us.
Bryan: Thank you for having me.
Benz: Thanks for watching. I'm Christine Benz for Morningstar.com.