Mike Rawson: I'm Mike Rawson, ETF analyst with Morningstar. We're talking today about factor investing, and joining me is Craig Lazzara, a senior director and product manager at S&P Indices.
Craig, thanks for joining me.
Craig Lazzara: Mike, thank you for having me.
Rawson: Craig, what is meant by this term factor investing? It's kind of a newer term that's been associated with some of the new ETFs that are out there. How is factor investing different than traditional index investing?
Lazzara: One might think of factor investing as a second- or third-generation of traditional indexing. If you think of cap-weighted indexes like the S&P 500, for example, as the first generation, factor investing is either the second or third, depending on how you count. But the thing that factor indexes have in common is that they try to embody a quality or a factor, obviously, hence the name, but a quality or an attribute that many stocks have and it's a quality or attribute with which excess returns are thought to be associated. For example, sometimes people think of low valuation as a factor of return; people have thought for years of beta as a factor of return, meaning an attribute with which certain behaviors are associated.
To the degree that it's possible to look at what an active manager does and subdivide that into an active part and a part that really comes from a certain exposure to a set of attributes that are more or less always there, it's appropriate and possible now to be able to dissect that active management process and say that a certain portion of what this active manager does is really just providing you with exposure to a factor, to a set of attributes that give certain return patterns. And those factors indexes can be used obviously either as benchmarks for that kind of manager or as the basis for a passive product.
Rawson: How should I be using these in my portfolio? Is this something where they would form the core of my portfolio, these factor ETFs, or would I go out and buy the S&P 500 as a core of my portfolio and then maybe supplement that with some satellite positions, some tilts using the factor ETFs?
Lazzara: I would never tell you not to use the S&P 500 for any purpose you're inclined to use it for, so certainly, you should do that. I think in terms of how factor indexes and product based on factor indexes should be viewed, it really depends on the factor you are talking about.
I'll give you an example. We launched about a year and two months ago, two new indexes and then soon thereafter there was product on these indexes. The indexes were called S&P 500 Low Volatility and S&P 500 High Beta. They are twins, basically; they came out on the same day. Methodologically they are very similar, except that the Low Volatility Index is designed to give you exposure to the least volatile members of the S&P 500 and you can think of low volatility as a factor in the sense we are mentioning. The High Beta index, as the name suggests, is designed to give you exposure to the stocks in the 500 with the higher systematic risk, and again people thought of beta as a factor for many, many years.
Now, the way in which those two indexes behave--they were designed to be different, and they are different. What High Beta will do is that in a strong market environment, it will do very well; it magnifies the returns of the market. In a weak environment it will do quite badly because it magnifies the returns in a negative direction, as well. People tend to look at that pattern of returns of High Beta and view it more as a trading vehicle or a tactical vehicle than a core position.
On the other hand, if you look at Low Volatility, it tends to give you participation in rising markets, not necessarily full participation, but certainly substantial participation in rising markets, and it tends to give you some protection in falling markets. So, your overall pattern of returns is muted relative to what it would be in a cap-weighted index, or the parent index from which these factor indexes were derived.
That being the case, a lot people will view a low-volatility approach as a potential core position because it gives them as investors, the possibility of participation in rising markets and at the same time, it shields them from some of the worst damage in falling markets.
Rawson: Craig, you've done a lot of work on this low-volatility theme, factor phenomena, that's observed out there in the market where low-volatility stocks tend to outperform, on a risk-adjusted basis, these high-volatile stocks. That theme holds a lot of appeal, particularly in this current economic environment where economies are really volatile, the stock markets seem to be more volatile than usual, times are uncertain, and we are in the lower-return type of world where interest rates are really low. You mentioned the launch of the S&P Low Volatility ETF, its symbol is SPLV. That was one of the most popular ETF launches of last year; it already has more than a $1 billion in assets. Can you talk a little bit more about how that particular index is constructed?
Lazzara: With the S&P 500 Low Vol Index, one of the things I think that has led to its appeal is that it's really very simple. And I say that in the context of warning that for anyone listening to these words, sometimes you get at the low-volatility indexes and what are some of that minimum-variance approaches, and they are very complicated mathematically.
We set out to try to give you a simple way to access low volatility in this index, and it's very simple. You start out with the 500 stocks in the S&P 500, you look back a year, and you measure the actual volatility of each stock in the 500, just using 252 approximately daily returns. The 100 stocks with the lowest volatility, the lowest standard deviation, become the constituents of the Low Vol Index, and those constituents are then weighted in inverse proportion to their volatility. So, if I have one stock with the volatility of 12% and another stock with the volatility of 6%, the 6% stock gets twice the weight as a 12% guy.
Now, it turns out that both of those steps, the selection of the least volatile stocks and the weighting by the inverse of volatility, both of those contribute to the overall return pattern. In other words, if you simply took the least volatile stocks and cap-weighted them or equal-weighted them, the result would not be as good as the result in fact that we got.
Rawson: Not to mention that you're starting with the S&P 500, which is generally a high-quality index to begin with?
Lazzara: Exactly. It is a large-cap index. We've actually done some work on the S&P 400 MidCap and the S&P 600 SmallCap indexes, which are obviously as the names suggest, farther down the capitalization range, and the low-volatility effect seems to be present there, as well. When you think of low volatility as a factor, as we said earlier, it's present in a big way.
Rawson: There are two ETFs out there which are kind of competitors to SPLV. You mentioned these ETFs try to provide a minimum variance, and they are a little bit more complicated mathematically in that they use an optimization program or risk optimizer. According to theory that should be a better approach. Why is it that you've chosen not to go with an optimization program?
Lazzara: Well, there are a couple of reasons. One was we had actually tried it very early on in the work we're doing; it didn't work as well. You got a better solution with the simple approach. The second reason which I think is more telling philosophically is that--let me step back and say if you speak to other indexers or certainly to people who are active managers of minimum-variance portfolios and ask them how they produce the pattern of returns that they produce, the description goes something like this, and I don't think I'm being unfair in offering these comments. Well, first of all, you need a really good risk model because you have to understand how various stocks interact with one other. The second thing you need is a really good optimization program because the risk model is too complicated if you were just to look at it and understand what to do. So, you need to be able to extract the value from the risk model. The third thing you need is a lot of experience in constraining optimization routines because optimizers are famous as error maximizers. If you have a mistake in an optimization program, no matter what you are trying to do, the optimizer will find the mistake and overweight it. Therefore, the practitioner of, say, a minimum-variance approach has to have a lot of experience at how to constrain his optimizer so that it doesn't give results that are just impractical or completely unintuitive.
So, you have to believe in lot of things. You have to believe in the risk model. You have to believe in the optimizer. You have to believe in their ability to marry the two. And I'm not implying that any of the things you have believe are actually untrue, but there are a lot of them. What we wanted to do is to structure something more simple, and what you have to believe to believe that our approach makes sense is simply that low volatility persists because if low volatility persists then if you look back a year as we do and say what were the least volatile stocks in the market in 2011, for example, it's fairly likely they are going to be among the least volatile stocks in 2012. It's not a perfect indicator, but it's not bad.
Rawson: Absolutely. Another relationship between low volatility is the relationship to value stocks. There tends to be an overweighting in maybe consumer staples, utilities, and maybe health care. But can you talk a little bit about how a low-volatility strategy might be related to a value strategy?
Lazzara: There is a certainly an overlap, but the low-volatility factor, if we can go back to that nomenclature, the low-volatility factor is not simply a proxy for the value factor. They really operate in different ways. And the simple way to see it is by looking at the results of let's say the S&P 500 Low Vol Index last year. If memory serves, it was up on the order of 14% for the calendar year 2011, the S&P 500 was up 2% or so in total-return terms, and S&P 500 Value was actually, I believe, down. It certainly greatly underperformed [S&P 500 Growth].
So, here you had a year in which S&P 500 Growth dramatically outperformed S&P 500 Value. So, the value factor was not working particularly well, and yet the low-volatility factor did very well. There are number of reasons for it. One of them is that where you find low volatility at a point in time can vary. We look back a year to at least the Low-Volatility index's constituents, and we rebalance it four times a year. So, if you were to look right now, as you observed, the biggest sectors in the Low Volatility index are utilities and consumer staples.
If you'd looked in 2002 or 2003, the biggest sector by far was financials because financial stocks had low volatility at that point. So, we would've had a big overweight in financials, all of which was gone by the middle of 2007, luckily, but because the location of low volatility in the market shifts, the Low Volatility index can chase it, if you will.
Value indexes tend to be much more static in their sector allocations. There is never going to be much in technologies; there always going to be a lot of utilities, consumer staples, and things like that. You don't have quite the same ability to pursue the low-volatility factor, and to the degree that, as I say, the factor exists independently of value, the bigger your exposure to it can be the better off we think you can be over time.
Rawson: Craig, thanks for those insights. Craig Lazzara with S&P Indexes. I'm Mike Rawson with Morningstar. For more on ETFs check out etf.morningstar.com.