Low Volatility vs High Dividend: Which One Is Better?

Shannon Zimmerman: Good morning everyone and welcome to our panel, "Low Volatility Versus High Dividend: Which One Is Better?"

I'm Shannon Zimmerman. I'm an associate director of fund analysis at Morningstar, and I'm very pleased to be here for this panel today, to moderate this panel with a group of experts who I had a chance to pre-interview in preparing for today's session, and I know that they are all very engaging and engaged with the topic that we are going to discuss today.

Let me give you a brief introduction for each of our panelists. Beginning on my far right, Raman Subramanian is an executive director at MSCI, and he's responsible for the creation of new methodologies, the improvement of existing methodologies as well as research in the strategic positioning of their index series covering various asset classes.

Next to Raman is Craig Lazzara, who is the head of index investing for the S&P Dow Jones Indices, and he was previously the product manager for the firm's U.S. equity and real estate indices, including the Case-Shiller Home Price Index.

Then finally Jeremy Schwartz, the director of research at WisdomTree, and he is responsible for equity index construction, the process there, and he oversees the research process across the entire WisdomTree equity family.

So with those introductions I want to begin with a little bit of a challenge to the premise embedded in the title of this panel, and ask each of our panelists to speak to that "versus" part of it. Is it the case that it's one or the other? I mean in terms of investment product marketing, these strategies are often written about and promoted in quite different ways, yet one has to assume given the typically low-volatility nature of dividend strategies that there is some level of correlation.

Jeremy, you want to go ahead and get started with that question.

Jeremy Schwartz: Sure, thanks.

Thanks everybody for coming to our panel today. I'd there are some similarities between high dividend strategies and low volatility strategies, but they are focused on different things. The low volatility obviously has a sort of the market on beta or volatility, and they are really focused on that return stream, whereas a dividend yield focus is really focusing on the valuation characteristics. So, it's more about, what price are you paying to access those securities. So, there is this difference between returns versus valuation focus.

<TRANSCRIPT>

But if you think about the trends in indexing, I think the concept of what they are doing at minimum variance, the low volatility from S&P, and what we're doing with dividend stream weighting speaks to a very powerful force overall and trend in the indexing community, which is that, if you look at how ETF investors have traditionally allocated, 95% of all the ETFs today are following one strategy, which is a market-cap-weighted indexing strategy from U.S. large, mid, small to foreign to even the bond indexes are all following these market-cap weighted indexes. And if we were here at the Morningstar Active Manager Conference, how many people would use one active manager for their entire portfolio? PIMCO, Bill Gross is very good at bonds, but would you use them for your small-cap emerging market equities, probably not.

But with ETFs, it's routine to follow one strategy of this market-cap weighted indexing, and I think that investors are eventually going to say, it's not the end-all be-all. We are going to want to diversify our indexing approaches, and indexing is just about what stocks you select and how do you weight them. And there are a lot of different ways to do that with different pros and cons in different market environments, and I think all these things can be packaged together in different ways to just help investors get a more diversified portfolio and perform better in different market environments.

And that's my big overall theme is that, at WisdomTree we think you should diversify just from strict cap weighting using an multi-indexing framework, and I think that speaks to what we're doing today on this panel.

Zimmerman: So, Craig what about that? Are these strategies in tension? Do they complement one another? Should investors think about having exposure to both in their portfolios, or is one maybe sufficient because of the degree of correlation that there is?

Craig Lazzara: Let me start, if I may, with a preamble as proprietor of the S&P 500, which has about $1.4 trillion indexed to it, no word of criticism of cap-weighted investing will pass my lips this morning.

That said, the way I would think about the question of the panel, low volatility versus dividend is, we view that as kind of second or maybe a third generation in the progression of indexing--the original generation being big cap-weighted indices such as the S&P 500, MSCI EAFE. The second generation being extensions and subdivisions of the first generation, so MSCI Emerging Markets in the international case, S&P 600 Small Cap and 400 Mid Cap, slices as to growth and value, which we've all done, or into sectors by various means. The third generation is what we would call strategy or sometimes factor indices, and that's where low volatility and dividend strategy and others as well sit.

What I like to tell people when we talk about low volatility, low volatility is not simply a proxy for value or a synonym for value or for yield. If you think, say, in a Fama-French context, Fama-French identified two factors, small size and cheap valuation, as important factors in addition to market beta. So, valuation kind of fits into the Fama-French framework, and I would argue that the correlation between high dividend strategies, depending on how you construct them, and classic value strategies is pretty high.

Volatility is a thing apart. There is, I would argue, a volatility factor independent of high dividend, and you see that both in the correlation of low vol strategies and value strategies, which are not particularly high for equity strategies, and also if you look at things like the sector diversification--low vol looks very different than traditional value.

So, to answer the question, is there a difference? Yes, there is a meaningful difference, because the low volatility strategies or minimum variance strategies have quite large exposures to this low volatility factor, and value typically does not.

Zimmerman: So, I want to come back to that in just a minute and talk about low volatility as a factor, maybe the next canonized factor, so to speak. But Raman, could you speak to question, too, that's embedded in the title of our discussion today? Is it an either/or or a versus proposition, or is there significant overlap between the strategies that investors should be aware of?

Raman Subramanian: The way we have seen any of this, as Craig and Jeremy said, that if you look at the evolution of indexing, it all started with market-cap weighted indices, because we know that markets in the long run can be efficient. In that context, if some investor believes in market efficiencies, what they are going to do is that they are going to allocate to a broad market-cap weighted index, whether it can be a U.S. broad market index or a global broad market index.

Now, as we know that markets can go through cycles of inefficiencies or in-equilibrium, that's where factors will come into play. So, sometimes you will find that a valuation-dependent regime will happen, and low valuation stocks will tend to outperform. Then there can be regimes where lower volatility stocks will try to outperform.

Now, the question is, can you market-time that? As an investor, can you figure out what is the regime we are in? And the way we have seen in the institutional space, the broad institution--and we're talking about pension plans, global pension plans--that's the best practice, which tickles down to the smaller retailers. What they have seen is that, it's very difficult to do the market-timing. So what they do is that they form core strategic allocations to the various risk premium factors, that includes valuation-driven strategies like a value strategy, they are looking to more dividend strategies, they can also go into lower volatility strategy.

So, going back to the classic Fama-French, we talk about the three size factors. You can extend that to lower volatility. So, in our opinion, both are complementary, so I will not say that one is better than the other. It depends upon what is the view you're forming and how you actually mix that in your strategic core allocation.

Zimmerman: Jeremy back to you. I want to sort of dig into the methodologies or at least the way each of you think, and your firms think, about low volatility, and we'll come to dividends in just a minute.

Jeremy, I want to stick with you just with this first question. I know that Wisdom Tree doesn't offer a low volatility strategy per se, but you do analytics around that in terms of the products that you do offer, and you and I were just talking about it, and you were talking about a low volatility in terms of beta. Is that the core risk measure that you're looking at when you think about low volatility?

Schwartz: Yes, and let's just piggyback off of what Raman just said in talking about the market being efficient on the long term in forming a market-cap weighted strategy. The problem with that view is that they are so prone to bubbles that when you have bubbles, a cap-weighted index is going to overweight the most overvalued securities. So, … Japan had a 200 price to dividend multiple in 1989. EAFE ex-Japan has returned about 8% a year for the last 25 years. EAFE, because it was 60% in Japan at the peak of the bubble has returned 4%, so you lost 4% a year in the EAFE Index because it was overexposed to Japan at the peak of the bubble. If you looked at the S&P 20 years after the tech bubble in 2000, it'll probably be something similar, where if you did S&P ex-tech 20 years after the bubble, it'll probably be about 4% a year better than the S&P by itself.

So, a dividend weighting or any rebalancing program to rebalance away from the stocks that have just risen more than their underlying fundamentals helps reduce the volatility to get back to your question, by selling the stocks that have appreciated the most.

We see it most evident in our emerging-markets index, which has been characterized by high volatility over the last five years, but in 2008 oil was at $150, you had energy materials stocks running, you sell the stocks that really outperformed, then you go down a lot less in the downturns. So the market was down 53% in 2008, our high-dividend emerging markets was down 37%. So that rebalancing nature to sell the winners is one way to lower the volatility, and the beta of that strategy is about 0.8 since inception over five years. Our U.S. high dividend strategy over the last three years has a beta of 0.72.

Zimmerman: What about total returns, though?

Schwartz: They've both done very well versus the market. In the last three years, high dividend stocks have outperformed the market by 5% to 6% a year, emerging markets 5% to 6% a year above the market. So the rebalancing, I think, is what helps to lower volatility by selling the winners that have really outperformed, and rebalancing everything else.

Zimmerman: Craig could you speak to that, too, on the methodology? I know that you were quite essential in the development of the S&P Low Volatility Index. What's the main metric that you look at in the construction there? And if you could just sort of describe in brief the methodology that S&P uses for its Low Volatility Index?

Lazzara: The methodology that we use for S&P 500 Low Vol, which was the initial entry that we made into the low volatility space about a year and a half ago, and also it has been followed by mid-cap and small-cap and Canada and emerging markets and developed markets. So the same methodology, mutatis mutandis, is applied across the board. What we tried to do was to develop what we refer to as a model independent, non-optimized framework, which is a complicated way of saying what we wanted to do is take a very simple approach to measuring low volatility and see if it worked. And in fact it turned out to work quite well.

The approach we took was simply to say, in the case of the S&P 500 for example, take the 500 stocks in the S&P 500, measure their trailing standard deviation, each of the 500 stocks, the 100 with the lowest volatility become the constituents of the Low Vol Index, and then importantly those constituents are weighted inverse to their volatility.

Zimmerman: Just a point of clarification: So you are looking at standard deviation over a certain trailing period to pick the constituents for the index. What is that trailing period that you look at?

Lazzara: It's a year. It's a year's worth of trading days, which is kind of, I don't want to say an arbitrary choice. It's not completely arbitrary. As you all know, if you have done any quantitative analysis, there is written on stone somewhere that you must always try five years' worth of monthly returns for 60 months. We tried that, and it doesn't work all that well, because it doesn't adapt very readily. If you were to do six months, you kind of get the similar return pattern, or a year and a half. The year is a convenient look-back, and once we saw that it worked, we didn't want to torture the data to find out exactly what the optimal period would have been for the years we were testing.

So, that done, you get the hundred stocks, you then weigh them inverse of vol, and what that mean simply is if you have a stock with a vol of 10 and another with a vol of 20, the 10 guy gets twice as much weight, rebalanced quarterly. That's the process.

Two important things to say: One is, the fact that we weight by inverse vol is a big contributor to the returns of the strategy. In other words, if you took the same 100 stocks and cap-weighted them or equal-weighted them, you would get exposure to the low volatility factor, you would get a different return pattern than the S&P 500, but it is more different and more exposed when you weight by the vol factor, that's the first thing.

The second thing is, what you need to believe in order to believe that this strategy makes sense is simply that low volatility persists. No risk modeling, no optimization. Low volatility persists, and what I mean by that is simply, if you want to know what's going to have low volatility in 2013, ask what had low volatility in 2012. It's not a perfect indicator, but it's not bad, and in fact the results of the actual portfolio that we got simply by this look-back method are typically 25% to 30% less volatile than the returns of the underlying index.

Zimmerman: So, standard deviation is the main risk measure that you're looking at. I'm sure that you do analytics around beta, though. So, what's the beta of the low volatility portfolio?

Lazzara: The beta of the low vol portfolio is, again, depending on when you measure, 0.65 to 0.8.

Zimmerman: So, comparable to what Jeremy was describing?

Lazzara: Yes, it's comparable if you use that metric. When we were developing low vol, we said, suppose whatever we call it, let's just make it low beta, because we have a high beta index set, and that does what it does. Let's just do the inverse, and it was OK, but if you're looking to minimize volatility, you might as well minimize volatility not minimize beta.

Zimmerman: So Raman, could you speak to that question and describe what is, I think, a quite different methodology that you use at MSCI, and what you see as the main advantages of the approach that your shop takes?

Subramanian: So, in case of MSCI, … the construction goes back to the philosophy that ultimately what is the aim we are starting from. We want to tilt it to a specific factor; in this case it is lower volatility. Now the idea is that, you still want to maintain the main characteristic of your underlying parent benchmark. So, if you are starting with the S&P 500, if you apply that to our methodology, we want to still maintain the similar sector diversification, and if you are going to create one thing which is for broader country composites, you still want to have the country diversification.

Once that is achieved, … we want to tilt it to a specific lower volatility factor, and the way we do that is we look at not only volatility as a factor, but the interaction of the volatility across different stocks. So, we are looking at the correlation of the volatility across differ stocks to get to the final end product, and the reason for doing that is, take the case that Craig mentioned, that you sort it by S&P and pick the lower volatility stock. Now if you imagine that you wanted to do this exercise for a global index like MSCI Emerging Market Index. When you do that, take the case of what happened in 2010 and 2011 when Arab Spring was going on. Egypt didn't trade for two months. So, if I have done this rebalancing, say, a month after Egypt didn't trade for two months, and did the same exercise, your index will be comprised of just Egyptian stock, because you think that … these are lower volatility stocks.

And to avoid that issue, what we do is that we compare not only the volatility, but also the correlation among the stock, and take into account that we want to avoid any of this illiquidity factor, which can be built in because … the stock exchange didn't work or something related to that factor was not working.

Zimmerman: Let me ask a question, and Craig I want to start with you on this. It's also skeptical, but not about the title of our panel, but about the investment product side, the marketing side of this.

So you look at a strategy like price momentum, and there is quite a close correlation between a price momentum strategy and a growth investing strategy. And historically it seems to be the case that price momentum has generated greater alpha, so some researchers have argued that investors could profit by not thinking about having a growth sleeve to their portfolio, but a price momentum sleeve.

Is there anything that is comparable to that in terms of either high dividend investing or low volatility investing? Are investors maybe already getting this exposure or this performance pattern from some other part of their portfolio that is maybe not as marketable, but is just as successful?

Lazzara: I would argue not. I think the classic, or second or first generation strategy with which low vol correlates most highly is clearly value strategies or income-sensitive strategies, and in both those cases, low volatility has a factor exposure of its own, as I said earlier. There is, in our view, a low volatility factor, and if you want exposure to it, low volatility or minimum variance portfolios are a good way to get it.

One way that we find helpful to think about that is if you analyze low vol in a variety of different market environments. So, if you look at low vol, for example--and we typically will split the universe, if it's big enough, into quartiles or what's close to quartiles--so if you look at the worst positive months, the months when the S&P 500 is down the most, and then say how did low vol do, or versus how did value do, how did high beta do (God help it) in those months, and see what the results were, and then you do the same thing for the small negative experiences and the small positives and the big positives. What you typically find, or at least what we find, with low vol is that in the worst negative months, low vol does well. It adds, on average in the U.S., roughly 3% relative to a market that's declining on average in those environments maybe 5% to 6%. In the small negative experiences, low vol, again, adds value maybe 1.5%-2%. In the small positives, more or less a tie, and in the big positives low vol lags.

You get a similar pattern with any of our value or dividend-sensitive indices. The important difference is that they are not nearly as good from a defensive standpoint. So, from the standpoint of going down least in a market which is declining, low vol is clearly a better defensive vehicle than S&P 500 Dividends Aristocrats, for example, or Dow Jones Select Dividend indices. It's also fair to say that the dividend sensitive or value sensitive indices will typically participate a little more in the big positive, not in the smaller positive, but in the big positive. So you get a different pattern of returns.

The thing I always say about low vol from a marketing standpoint is that, you should think of it, depending on the risk preferences of your clients, you can think of low volatility as a core holding position, because it gives you protection on the downside. If the markets are down, it's not going to go down as much typically, and that's a fairly reliable phenomenon over various periods of time and over various geographies. So it is, I think, a fairly robust pattern. You get participation on the upside and protection on the downside.

Dividend-sensitive indices and value indices will give you some of that, but they are not nearly as good as downside protection vehicles, in our experience, although they do a comparable or maybe somewhat better job on the upside.

Zimmerman: We'll come back to that in just a minute, because that's quite interesting. So the downside capture ratio, just to stick with the metrics, is superior on the low volatility side, but R-squared between the low volatility strategy and some of the value indices that S&P offers would be relatively high?

Lazzara: A better way to think of it is just to look at tracking error relative to vanilla--vanilla being the S&P 500, of course, in this context. The tracking error of, let's say, S&P 500 Value versus S&P 500 is … maybe about 3% to 4% range, because it has a lot of stock and it's cap weighted. The tracking error of S&P 500 Low Vol versus S&P 500 is on the order of 10%. It's a big tracking error. As the technologists say, that's not a bug, that's a feature. It is designed to have a big tracking error, because that's, in our view, the way that you get maximum exposure to the low volatility factors. So if there is, as the academics call it, this low volatility anomaly which is there to exploit, and there are a number of reasons if you wanted to discuss why that might be the case, but if there was a low volatility anomaly, in our view, taking advantage of it requires you to take fairly bold positions relative to a benchmark like the 500, and so that's what we've done.

Zimmerman: Yes, we'll come back to that as well, because I think the paper that has looked at the longest time series came out just last year, actually 41 years, and the low volatility premium seems to persist in a much more consistent pattern than other premiums that have now been canonized, small cap and value.

So, Jeremy, back to the question that I posed to Craig. So WisdomTree doesn't have a low volatility strategy, right? You have a dividend focus. Why is that, and is it in part at least because maybe they are too closely aligned to really warrant, in your view, having a standalone product?

Schwartz: Well when we first were looking at the market, we are trying to solve for that problem in cap-weighted, which had this non-rebalancing nature, prone to bubbles, and tried to adjust on the concept of value. What's interesting--you talked about, how does it overlap with the other value type strategies? Does growth correlate with momentum?

Just being here again at the Morningstar Conference, if you looked at just the pure Morningstar nine box style map; you go to Morningstar.com, pull it up. You pull up the U.S. mean-variance, you pull up the S&P Low Vol, you pull up our large-cap dividends, they all do plot in the large-cap value bucket.

But I do think you do have to differentiate, like Craig was saying, about, what is the underlying exposure, because there are very different exposures in the different strategies. The S&P Low Volatility has no financials and 30% in consumer staples, 30% utilities, right? So it's a very concentrated 60% in two sectors, no financials.

Subramanian: We don't have any style bets or sector bets, so … that comes to the original question of tracking error. So, … if you build a benchmark which is just looking for the bottom decile by lower volatility, you get some embedded style bets in addition to the lower volatility. So you will be looking at consumer staples and utilities. The way we have avoided that issue of inherent style bets is that, we still want to maintain the same characteristic of the underlying benchmark by sector and countries.

As I said going back to the Egyptian example, you don't want to include the index with just the Egyptian stocks if the market is not traded. So that means a tracking error that we have is inherently lower compared to a pure decile-based lower volatility. Having said that, it depends upon the investors--what you are comfortable with when you are trying to pick up the stocks. If you think that a decile-based strategy, with a 10% tracking error, that's what you want, you can go with that, but if you think that you still want a tilt on your beta strategy towards low volatility, then a minimum weighting strategy is much more appropriate for you.

Schwartz: Just to finish my thoughts here. So within the dividend space, I think you also have to differentiate, just like they have differences in tracking error. … They are trying to minimize tracking error. Within the dividend space, there are very different ways of getting the dividend stocks. So, you can look at the average dividend paying stock. There are 1,300 dividend stocks in the U.S. today. You say that's the core dividend paying market.

Yes, it's maybe tilted a little bit towards value, but there are different ways of slicing that dividend paying stock. There are some people who focus on the growth element of that dividend paying stock. So, they are going to have a below average dividend yield. Say, if the average dividend stock has about a 3% dividend yield, the growth dividend stocks, because there are a lot of people focusing on the growth element, they have about a 2% dividend yield.

You can say the value dividend stocks, if you sort by dividend yield, you select the highest dividend stocks, you're going to have a 4% dividend yield. So, there is a subdivision within dividend indexes that have different exposures in terms of the value stocks, the 4% yields, the growth stocks, around 2% yields.

And then the weighting has different elements. So, Craig talked about inverse weighting for volatility; that's going to be skewed towards smaller-cap stocks. If you dividend-yield weight, a lot of the indexes, the Aristocrats, dividend-yield weights...

We're dividend-stream weighted, so it's dividends per share times shares outstanding for the vast majority of our indexes. It's going to give bigger weight to bigger companies, be more representative of the market. It's different than when you are yield-weighted. That's going to have a mid-cap skew or a dividend per share weighting, so the Dow Jones Select Index is dividend-per-share weighted, mid-cap value skewed. The dividend yield weightings is mid-cap value skewed. So, there are just different ways and different exposures. You got to look at the underlying baskets to see what you're getting in the different ways of doing it.

Zimmerman: So, I want to pick up on a point Jeremy made, but toss the question to you Craig. So the strategies cluster, as you say, in the large-value square the style box. The backdrop of this question is, researchers find what they're looking for. If you have these questions about what is responsible for a certain measure of outperformance, that's what you are looking for, and lo and behold, that's what you find.

So, if the strategies cluster in the large-value square, if you blew the large-value square out as a style box unto itself and divided it into the nine squares, would the funds that succeed the most, or the strategies that succeed the most, be in the lower left hand corner of that, so that you are also picking up a small-cap premium within that space, and even more of a value premium, because the stocks are cheaper?

Lazzara: If you think of S&P 500 Low Vol, the average cap of the stocks in that index as of last quarter-end was about $40 billion. Now that's much smaller than the S&P 500, but $40 billion is a big company. So, it's hard to characterize it as a small-cap tilt. Is it smaller than the S&P 500? Yes, but almost everything is. You'd be hard put to be higher than $100 billion average cap, although there are places that do it. And the S&P 500 Low Vol has about 110 basis point yield premium to S&P 500.

In terms of other value characteristics, it's very much a mixed bag. I mean the P/E isn't distinctly lower. The price to sales is not distinctly lower; in some cases, it's higher. So there is really not a tendency. We did a very simplistic style analysis, just using a returns-based paradigm fairly recently, and if you look back over time--and that kind of analysis is very sensitive to the time period chosen and the look-back period you use--but if you look back over many years, most of the time, low volatility ends up kind of on the value side of the growth value divide. If you look back only on the past three or five years, low vol is … more highly correlated with growth than with value.

That said, the correlation is not very high in either case. The reason for that, and a good way to illustrate that simply, is to just cite performance statistics for 2011. In 2011 the S&P 500 was flat, more or less, the total return was 2% because of the dividend yield. The S&P 500 Value Index was down, I think, 4%-6%, something like that, so growth would have been up complementarily. S&P 500 Low Vol was up 14%, so the performance is very different from a value index on a moment-by-moment basis.

Schwartz: That's financials, right? So, financials were down, … because you're underweight financials, which were the big heavyweights in the Value [index].

Lazzara: Low Vol had--and I'll be first to admit it--it was almost a perfect storm in favor of Low Vol last year, because in our case, the most overweighted sector was utilities, which happened to be the best performer, and the most underweighted was financials, which happen to be the worst. It's important to say that if you look at the history of Low Vol, even though Jeremy is absolutely right, it's roughly 60% in consumer staples and utilities today. If you looked at Low Vol in 2001, 2002, 2003, 2004, 2005, 2006, the biggest sector weight was financials, and financials went from approximately a 30%-35% weighting at the beginning of 2007 to approximately a 1% weighting by June of 2007, as the financial sector got more volatile. So the fact that if you took a snapshot today, you'd see a staples and utilities bias is absolutely correct and fair to say, but that hasn't always been the case.

Can I just add something. I want to say again. We're not here to talk about cap-weighted indices, but I want to address something that Jeremy has raised, and he is not the only one who raises it. It's a good point and that is, the statement that cap-weighted benchmarks are subject to bubbles. And if you look back, obviously, the Japan bubble in EAFE 20 years ago, the tech bubble in the S&P 500 10 years ago. There are plenty of examples of places where cap-weighted indices have been disadvantaged because they had a big exposure to a sector that turned out to be [over]valued and that's very clear in retrospect.

I'll ask you today, in the first six months of the year, the first nine months of the year--maybe it's the last six months, I forgot the exact timeframe--Apple Computer added more to its valuation than Microsoft is worth. So today, is Apple Computer in a bubble? I have no idea. I mean you can make the argument either way. Is Apple a bubble? Maybe, Maybe not. I don't know, but…

Zimmerman: Well at least they just paid their first dividend, right? It's now going to be $10 billion worth of dividends.

Subramanian: For a pure retail investor, if you look at the…going back to the original comment … we sell the winners and buy the losers. So that's kind of a tax inefficient strategy that you have you to consider. Because if you are always churning out your portfolio, if you look at a typical cap-weighted portfolio, S&P 500, very minimal turnover. If you go to a global composite like All Country World Index, you're talking about 1% to 2% annual turnover, whereas any of the strategy indices, including minimum variance strategies, where we constrain the turnover, you're talking about 10% to 20% turnover.

So as a typical retail investor, you have to consider that tax inefficiency which any of the strategy indices can bring in. So it is not straightforward that one is better than the other. It depends upon a lot of characteristics like the investor profile, the tax profile, and what is the ultimate goal. If you're fine with capturing the market efficient data, cap-weighted is better for you. But if you want to tilt it to a specific investment belief, whether it's the valuations driven strategy or a lower volatility strategy, then anything on the strategy index will be better for you. But we have to consider that what is the tax advantage for any of these strategies?

Lazzara: Well that's why it's great that we are here at the ETF conference, because with ETFs, if you are going to have rebalancing, the best place to have rebalancing is within the ETF, where you can do creation redemptions, and actually do a lot of the rebalancing through creation-redemptions and do in-kind transfers of securities to do the rebalance. So you minimize the tax drag within ETFs. So if you are going to do these strategies, the best place to do it is ETFs.

Zimmerman: Craig posed a question to the audience. The audience will have an opportunity to pose questions to the panelists in about five minutes or so. We are going to take one more pass through the high dividend side of today's discussion.

A lot of the investors, and probably investors who are working with folks who are in the room, are looking for yield. And in some cases, thinking about dividend paying equities as a substitute for the fixed-income portion of their portfolio, because the income that's being generated on that side is so modest relative to historical standards. That can be risky.

So a high yield can be a bad sign if a stock has cratered as a result of weakened fundamentals. While the yield looks great, the fundamentals may imply that price erosion is going to take back what the yield gives, and so nobody earns anything.

How do your methodologies address that risk, and what guidance do you have for investors who may be thinking about using dividend-paying equities as a proxy for fixed income? Raman to you.

Subramanian: So for us what they do is that, in case of our High Dividend Yield Index we have two different base screens which we apply. One is what we call a negative dividend screen, so we don't want to include stocks which are negative paying. So we were looking for the growth rate of dividend. But more importantly, we also look at the persistency of dividend, so making sure that they can really sustain the dividend and look at companies which have higher payout ratios, and try to screen them.

For example, we all know what happened to General Motors; their yields were so high just before they went to bankruptcy. Our methodology ensured that we screened out those kind of stocks. And then … we look for the highest dividend paying stock and then we do a calculation.

Zimmerman: Craig, what about S&P?

Lazzara: There are two things to look at, and we have a variety of indices that will use them in varying respects. One thing to look at is obviously the yield per se. That's easy, although it's subject to the danger that Shannon articulated.

The second thing to look at is the history of a stock's ability to increase its dividend. So we have introduced some time ago a series called S&P 500 Dividend Aristocrats, and the requirement to be in that index is that a company has increased its dividend every year for the past 25 years. That's a very high bar.

There is another version called S&P High-Yield Dividend Aristocrats, the bar is lower there, its 20 years, but still a very high bar. So if you apply that kind of screen and then whether you weight by cap or by dividend after you pass that screen, you tend to diminish the risks that Shannon articulated.

Another of our indices that came to us with the Dow Jones joint venture was Dow Jones Select Dividend; that applies a similar kind of screen but much less stringent, and therefore more companies will get in and the yield will be higher.

Zimmerman: And at WisdomTree.

Schwartz: So one of the things I actually find fascinating here is that you have two of the industry leaders in cap-weighted indexing, right? When they have talk about cap-weighted indexing, they want to own the market. They want to own the most broadly diversified portfolio of the market. When you come to dividend-based portfolios, they try to put on all these screens to identify the best dividend-paying stocks, and they are acting more like active managers trying to pick out the company growing their dividends every year, brings a more narrow basket. You own 20% of the market, let's say, versus holding all of the dividend-paying stocks.

Zimmerman: Jeremy this isn't Crossfire.

Subramanian: We are not a fund manager. We are a pure-play company. We don't pretend to be an indexer and a fund manager. We are just indexer, and we provide indices to the marketplace.

Schwartz: I'm just saying it's sort of an interesting contrast. When it comes to cap weighting, we're happy to own the market, but when it's dividend weighting or dividend stocks, we've got to pick these best dividend paying stocks.

So, one of the things we say is just own the dividend paying market. Buy all the dividend payers, weight them by the dividend stream. Get 1,300 stocks in the U.S. You don't filter out the up and coming growers. So, Apple, I mentioned earlier, $10 billion of dividends-- it's going to be the third-biggest dividend payer in the U.S. If you screen out for 20 years, you're not going to own Apple for 20 years, and they are going to be a big part of leading dividend growth.

Cisco just announced a 75% dividend raise. Last year they were the 40th biggest payer, this year they are going to be 22nd biggest payer. And if you have these screens, you're not going to own Cisco for 20 years. And so it's sort of an interesting dynamic of what you miss out when you put on the screens.

So, we just try to own the dividend-paying market and weight it by that dividend stream, which to answer your question about the yield trap is, you're going to get bigger weights in companies growing their dividends every year. So, I mentioned Apple, that $10 billion, Cisco raising [its dividend] by 75%. You're not weighting based on a yield concept; you're weighting based on their dollar value, the dividends per share times shares outstanding and giving bigger weights with companies with bigger dividend streams.

Zimmerman: What about Raman mentioned, the persistence of dividends? Do you look for that as well?

Schwartz: Well, I think it's hard to identify. How many people really predicted all the financials we're going to cut their dividends? If you look at the history of dividends going back to the Great Depression in 2009, you had a one-year collapse in dividends, 20% on the S&P. It's the greatest collapse in dividends since the Great Depression. Every other year, it's 3%-4% fall in dividends. How many predicted it? I don't know that many people did predict it, even with all these persistency screens.

So, I think it's tough to say. With hindsight, we can know if we lose financials, hey you did great; ex-financials, dividends are up 30% right now. But dividends are just getting past their peak because of the financial crisis, and I don't know that they are going to identify them with all these screens.

Subramanian: Indices are not there for prediction. Indices are there to reflect the marketplace. So, as we are not in the prediction business; that's an active management space. So we reflect what's happening in the marketplace, that's reflected through the data out there.

Zimmerman: So, I do want to turn it over to the audience. Do you folks have questions that you want to ask any of our panelists? A pretty lively conversation, particularly there at the end, Jeremy. Thanks for that. Just step to the mic and ask your question.

Unidentified Speaker: I had a question for Raman. Do we defuse or lessen the low volatility impact by being careful about your sector weightings? Could you talk about that?

Subramanian: So, it goes back to the philosophy of what we are trying to achieve. Our belief is that ultimately as an investor, you can capture the core beta by a market cap-weighted index, but if you are really tilting towards a lower volatility, the inherent style or factor tilt that you want is the lower volatility and avoid all other sector and country bias, so that can be built because of the construction.

So, as I said, if you just rank it by a pure volatility and take the bottom decile, you will get probably consumer staples and utilities in the U.S. If you do that for Canada, it will be mostly financials because energy and materials are much more cyclical, and they are much more volatile. So, to avoid that, what you can do is that, you can still get all the lower volatility stock from the entire spectrum of the index, but … avoid the style bet by looking at the correlation factor also.

Lazzara: We take an unconstrained approach, if you will, to low volatility, and I think based on research we have done, you can compare a formal optimized minimum variance strategy to the more rankings-based strategy that we favor. What you find is that typically in really bad periods, the rankings-based strategy does better, and the reason for that--think about how in the past 20 years, for example, there have been cycles of performance. In 2001-2002, S&P 500 Low Vol had no exposure to technology, because tech was obviously very volatile sector then.

If we had adopted a more formal, constrained approach, there would have had to be a certain technology weight in there, and therefore performance would have suffered because of it. The same is true in the financial crisis in 2008. We were completely out of financials as opposed to being constrained to being within a certain index weight. It's fair to say that in the re-bounce, when tech finally hits bottom and bounces up, we're going to lag in that quarter with the rankings-based approach, and the research has shown that fairly persuasively.

Lazzara: I guess, just to say, if you're going to use these in conjunction with cap-weighted indexes, … I think what people are looking towards doing is to combine this with their traditional other allocations--you probably want the more differentiated exposure. If you are already owning the market and you're just sort of putting these constraints on tilting it, you probably want the more differentiated approach.

Zimmerman: Other question from the audience.

Unidentified Speaker: Jeremy, could you go into a little bit more about the growth screen criteria that you were discussing earlier? 20 years I think is pretty onerous, so I think a lot of us would agree with that. Butlooking at a 10 or a five-year growth screen, do you think that could be a little bit more of a hybrid between high dividend and a low volatility play?

Schwartz: In our broad dividend index, we just wanted to own the dividend paying market. What we were trying to be is the Vanguard of dividends. And so you're owning the dividend paying market, you're tilting towards the dividend paying stocks, and we weren't going to put on the screens.

Even when you put on a 10-year screen, you end up only owning 20% of the market. The five-year screen, there is the Dow Jones Select, they own 100 stocks. It's that mid-cap value skewed basket. So, again it's not representative of the market. So a lot of these dividend-screening baskets do only own a small part of the market.

So we are trying to say, here is a valuation-adjusted way of owning the market, raising the dividend yield on the market. S&P at 2% yield, just like S&P Low Vol, about 100 basis points higher. If you dividend-weighted the market, it's about 100 basis points higher. So large-cap dividends, it's not 60% in two sectors, it's 19% maybe top sector consumer staples, tech is now going to a very large contributor of the dividends, utility is only 4%-5%.

So it's just a different way of owning that same underlying 3% yield on the market, with Apple as the third-biggest weight. Those kind of ideas of owning just the dividend paying market tilting back towards dividend once a year, have a valuation adjustment. If stocks double and the dividends don't double, you sell. And it's that once-a-year rebalance that we think adds value to get you a little bit lower volatility, to get you a little bit higher yields. Other questions?

Unidentified Speaker: I have a quick one. Jeremy, I think you said that dividends declined 20% in 2009. Has anyone looked at to what extent either Federal Reserve policies or strings attached to TARP caused banks or insurers through federal directive to cut their dividends?

Schwartz: Absolutely. The financial dividend stream in 2007, the way we monitor it, the peak was $288 billion, financials were $96 billion. So today financials are … a little bit above $50 billion, so they still have to double to get back to even. JPMorgan announced the biggest one, $3 billion. They came back. I think they are now a top-20 dividend payer again. But the Citis, the Bank of Americas, they are still constrained.

So it was financials. Dividends went from $288 billion, at the bottom it was around $215 billion, financials went from $96 to $21. Now they are coming back, but they are still well below their peak when everybody else is 25% to 30% higher.

Zimmerman: We have just a couple of more minutes, and I had alluded earlier in our conversation to this white paper that came out in 2011 looking at the low volatility premium. It's the longest time series I think that's ever been studied to look at the persistence of that. Pretty dramatic outperformance over a very lengthy period time.

At Morningstar, our research has found, and you can look it up on our website, by comparing total returns that a fund delivers versus dollar weighted or investor returns. Among volatile funds, there is typically a quite a wide gap, Craig, you and I were talking before we began, and I cover a fund called Fidelity Leveraged Company Stock. And over the last decade, that's delivered about 16% annualized, and the typical investor has only earned 2% annualized.

Given the persistence of a low volatility premium and given that investors typically don't use highly volatile funds well, should everyone be thinking about having additional exposure to the strategy, the low volatility strategy, either through a pure play, low volatility, or a dividend approach.

Lazzara: Yes, of course you should. Of course you all should!

The drawback I think for traditional value, dividend-oriented, minimum variance or low vol, to a varying degree in all of the cases is, and I say this to financial advisors all the time: If you have a client who is going to fire you when the market goes up 40% and you are only up 30%, you can't do this. I mean that's the nature of the tracking error, and that will happen. If you look at our research in the late '90s, low vol lagged the S&P 500 in '96, '97, '98, '99. As you would expect--the tech bubble was inflating. So there will be periods when it's awful, and you need to accept that.

If you accept that, if the client can accept that I'm willing to underperform in a raging bull market--I will get something, I just won't get the full market--in exchange for that, I'm not going to get slaughtered when the bubble deflates, and I'll get very good participation in kind of middling times. If that description of a client's utility function, that's too grand a term, but if that description of a client's preferences is congenial to you, I think this family of strategies makes great sense as a core strategy, really for two reasons.

One is the [reason] that Shannon alluded to, which is simply that typically investors … with a more volatile set of funds will tend to sell at the bottom and to buy at the top. And because the tops and bottoms are attenuated, there's less opportunity and less risk of doing that in a low vol space.

The other reason is because, again as Shannon alluded, there is a lot of research and not just coming from those of us who have a vested interest in it, but a lot of academic research that suggests this low volatility premium--it goes by different names--but this low volatility premium or anomaly is a real thing. It's real. It's significant, and it's persistent. So to the degree that you can orient a portfolio strategy to capture a part of that premium that's a good thing.

Shannon: We are about out of time. I want to let Raman and Jeremy weigh in on this question as well.

Subramanian: So our experience is that in terms of any of the strategy indices it can be part of the core along with the market-cap weighted.

We know that each of these strategies is cyclical in nature, so sometimes you will see value strategies outperforming, sometimes lower volatility. We will probably look at investors and say, do you want to market-time these strategies. If you don't want to market-time the strategies, one of the more efficient ways is to form this as part of your core along with the market-cap weighted index strategy. So, that's probably the more prudent way to look into these kinds of strategy indices.

Zimmerman: Thank you, and Jeremy close us out.

Schwartz: Well, Craig said he didn't want to talk anything bad about cap weighting, but I think he made the best case. You want to diversify. You want to use this as part of a complement to your traditional allocations. Don't just be in one strategy: use dividends, low vol, equal weighting, earnings weighing. There are multiple factor-based ways to allocate, and they all have a place in different market environments. And the nice thing about dividend weighting is that it does it for you. It does rebalancing for you. You don't have to market-time when to be in and out. It automatically reduces exposure as markets are riding more than they should, and that's the key element of dividend-weighting that we like is that it helps you sort of take chips off the table as the markets are booming.

Zimmerman: Gentlemen, thank you very much. I think we are all smarter about what could be a complex subject. Thank you very much for coming to the panel. Hope you enjoy the rest of the conference.