Barry Gordon on Building an Index

Back when dinosaurs roamed the earth, all ETFs tracked well-known third-party indexes like the S&P 500. But it wasn’t long before all of the major benchmarks were spoken for and ETF providers began to commission new indexes for their products. In most cases, they formed partnerships with firms that specialize in index creation, such as Standard & Poor’s, MSCI, Russell, Morningstar and PC Bond Analytics, creators of the well-known DEX bond indexes.

I’ve always been curious about how these new ETF indexes were created, so I called Barry Gordon and asked him to share his experience. Gordon is president and CEO of First Asset, one of the youngest ETF providers in Canada, and he’s launched a handful of funds in 2012. These include four equity ETFs tracking new indexes based on Morningstar’s CPMS strategies, which advisors and portfolio managers use to select dividend, value and momentum stocks. First Asset also launched three bond ETFs using barbell strategies pegged to DEX indexes.

Here’s part one of our interview. I’ll post part two later in the week.

When you created your new ETFs, why did you decide to base them on existing methodologies from Morningstar rather than creating your own rules-based strategy?

BG: As a new player in the ETF industry, I think it adds a lot to our offering in terms of people’s confidence in dealing with us. If Morningstar believes in us, and has partnered with us, there is some additional credibility that comes with that.

Any newly created index, whether it is cap-weighted or it uses another rules-based methodology, is subject to manipulation or data mining. That’s why you have to respect the index provider and believe they have integrity, so they are not going to say something like, “You know, if we got rid of everything under a $250-million market cap, this would look a lot better.” That’s why we chose Morningstar. They have integrity in spades: they are very well respected, and we were looking for that in a partner.

How collaborative is the process when you work with an index provider?

BG: Not very. When I approached Morningstar, we said we love the CPMS methodologies and they are well respected with advisors, so can we turn them into indexes that we can replicate and you can put your name on? They went away and we didn’t talk to them again for about six weeks. Then they came back to us and said, “Here is the basis on which we will do this and attach our name to it.” For example, they said they would have to get rid of the small and micro-cap stuff because it isn’t scalable and investable.

Now to be fair, they asked us for our thoughts. I asked questions like, “Why 50 stocks [in the US dividend index]?” or “Why 30 stocks [in the Canadian dividend index] when your original CPMS methodology has 25?” And they would give me the reason. For example, they believed their indexes offered better diversification in an ETF, because there is a lower concentration per name when you use equal weighting.

So it’s collaborative in that I’m able to ask questions about what they’re doing. We had an open dialogue as they created the index, but there wasn’t a single time where we said, “We don’t like that, please change it.” If we had done that they would have said no—to their credit.

That’s interesting, because I remember back in 2009 the S&P Dividend Aristocrats Index kicked out several of the big banks because they’d failed to raise their dividends. The Globe reported that Claymore was considering approaching S&P about changing the index rules “to make sure that anomalies are taken into account.” And I thought, you’d better hope they said no, because if they ever did that their credibility would go up in smoke.

BG: S&P would never do that. Never.

[Note: A reader pointed out that S&P actually did go on to change the index methodology in response to this “anomaly.” A company can now qualify for the Aristocrats index even if it does not raise its dividend for five consecutive years: it is permitted to “maintain the same dividend for a maximum of two consecutive years within that five year period.”]

Yet there’s a potential conflict of interest for index providers. If you decide you don’t like their index, you’ll stop licensing it and they lose revenue. But if they’re thinking longer-term, they don’t want to get a reputation as an index provider that makes arbitrary changes to suit their ETF providers.

BG: That’s exactly right. Morningstar understands that in the States there are billions of dollars in ETFs from iShares and Northern Trust that are linked to their indexes, so are they really going to risk that? Are they going to put their reputation at risk by doing something [inappropriate] with a small Canadian outfit? The answer is no. So they really told us how it would work and we trusted them on that. This goes back to that notion of having trust in the index provider.

How close to the original CPMS methodology is the Morningstar Canada Dividend Target 30 Index tracked by your ETF?

BG: It’s fairly close, but with some subtle differences. All of the changes they made were to ensure the index was liquid and replicable. That effectively gets rid of the very small-cap stuff. And interestingly, when we were talking to advisors who are using CPMS, they were not disappointed with that. I had thought it might be a detractor, but they told us they’ve learned from experience that sometimes with the smaller-caps, if there is a sell recommendation after the stock has performed very well and there are a hundred investment advisors following that stock, it is hell getting out. So they were actually quite happy that we don’t have the small-cap stuff in there.

That raises an important point, because I assume the performance data of the CPMS methodology assumes that transaction costs are zero.

BG: No, that’s not true actually. Again, the more credible your index provider is, the more confidence you can have in them. The published CPMS numbers actually assume a cost of seven cents per share traded. For something with high turnover like the momentum index, that works out to about 2% to 2.5%. For the methodologies that have much lower turnover, it’s less than 1%. But the point is Morningstar is very conscious about having those numbers reflect real life.

As the index provider, does Morningstar provide you with rebalancing instructions?

BG: Yes. The indexes are published: there is a Bloomberg code you can look up any time. But in terms of managing the portfolio, the rebalancing is done quarterly, and they say, “Here are the changes to the index that are coming into effect on Monday at close,” and then we buy and sell stocks appropriately by the close on Monday, and our job is to try to make that as seamless as possible.