Don't Overlook These Hidden ETF Costs

Christine Benz: Hi, I'm Christine Benz for Morningstar.com. Many exchange-traded fund investors assume that finding the lowest-cost ETF boils down to expense ratios. But Morningstar's Ben Johnson, director of global ETF research, believes that investors should look beyond that headline number. He's here with me to discuss some research on that topic.

Ben, thank you so much for being here.

Ben Johnson: Thank you for having me, Christine.

Benz: Ben, you wrote a recent article in ETFInvestor where you grouped ETF costs into two major buckets. Let's start with the first. You call them holding costs. This is expense ratio mainly, but potentially some other costs get packed under that headline. Let's talk about the other costs that fall under the heading of holding costs.

Johnson: Sure. And as you rightly mentioned is evident in a lot of the third-party survey data that we've seen, investors are oftentimes fixated on the headline expense ratio, so that fee that is immediately apparent and very much measurable what they will pay with certainty on a year-in and year-out basis to hold a given ETF. And that fee is the largest, oftentimes the most transparent, the most sustainable component of holding costs more holistically, but it is not the only factor that is embedded in those holding costs. So, when I think about holding costs, they are all of the things that can cause an ETF or an index fund's performance to deviate from that of its underlying benchmark index, which by definition isn't investable. So, the fund is just trying its best to match the performance of that index.

Now, there are certain things that will leave it short; fee, again first and foremost. Another piece of that could be accounted for by sampling error. So, in the case of some indexes rather that are comprised of a large number of securities, say, 1,000, 2,000, 3,000 different stocks and/or bonds, it is cost-prohibitive to try to go out and mimic that index on a stock-by-stock, bond-by-bond level basis. So, what the portfolio manager will do in those cases is that they'll try to build a representative portfolio, a sampled portfolio of securities that will best mimic the performance of a given index. Now, more often than not that works quite well. But what we've seen historically, particularly in periods of market reversals, is that that sampling approach can come back to bite the fund's sponsor and that bite manifests itself in the form of substantial tracking error, a substantial difference between the performance of the index fund or ETF and the underlying index.

Benz: In your article, you took a look at an emerging-markets ETF where that sampling error had created a little bit of a disconnect between the index performance and the fund. Let's talk about that.

Johnson: So, the case in point was the iShares MSCI Emerging Markets ETF. The ticker for that is EEM. And if you look at what happened when that fund cornered out of the bottom of the global financial crisis is that the securities that it left out of that sampled portfolio were those that rebounded the most sharply off the bottom at that point in time. And by virtue of having omitted them from that ETF's portfolio, that ETF lagged the index during that period of time by a very substantial margin. Now, subsequently, BlackRock iShares have gone back and revisited that approach. They now own a more representative sample of the underlying index. But certainly, for investors in that fund through that period of time that was a less than satisfactory outcome to put it lightly.

Benz: Right. So, this has the potential to work the other way too, right, in that sometimes this sampling would lead the portfolio manager to hold the most liquid names within an index. I would assume that in certain market environments that would be a good thing potentially.

Johnson: It could very well cut both ways, which is important to note. But what index investors have signed up for is as precise as possible tracking relative to that benchmark. So, as much as I will be dismayed as an investor in an ETF or an index fund should I be lagging my benchmark index, the benchmark index of that fund, by a substantial margin, certainly something meaningfully more than the fee that I'm paying, I will be probably be just as disappointed if I'm outperforming the index by a substantial margin because what I've signed up for, what I desire is near perfect tracking, the highest-fidelity tracking possible of that underlying benchmark.

Benz: OK. Before we leave this topic, you note in the article and I think it's important to point out, some index funds and ETFs just fully replicate their indexes. They don't do sampling. So, one-for-one, they would hold the very same securities in the same weightings as their indexes.

Johnson: Absolutely. So, if you look at an index like the S&P 500 index, there is no need to sample that particular portfolio given that it's made up of the largest, the most liquid U.S.-listed stocks. So, in the case of some of the most widely used, widely held index funds and ETFs what you'll see is a full replication approach employed whereby the portfolio manager is owning each and every security in that portfolio in proportion to its representation in that underlying index.

Benz: OK. So, we've got expense ratios, we've got some of these sampling costs that can come into play for certain ETFs and index funds. What are some other costs that fall under that umbrella of holding costs?

Johnson: There are other sort of lesser-known, less pronounced, less visible costs associated with just normal portfolio turnover. So, even very low-turnover indexes like a total stock market index will still experience a degree of turnover owing to corporate actions, mergers and acquisitions activity, what have you. Turning that portfolio over has costs, albeit more often than not, they tend to be de minimis. What you'll also see is that not all sources of holding costs are necessarily going to detract from the fund's performance relative to its benchmark. There are other ancillary activities like securities lending and the income that's generated from that activity that can actually be additive. So, it can help the manager, help the fund to bridge the gap between its performance and the performance of the underlying benchmark.

Benz: OK. So, the other major category of costs that an ETF investor can incur fall under the heading of transaction costs or trading costs. So I guess the headline number there is, am I paying a commission to trade in this ETF? Some providers do offer commission-free ETF trades. But there are other costs that investors should be attuned to under the heading of transaction costs. Let's talk about them.

Johnson: Absolutely. So, as you mentioned, commission, not unlike expense ratio, the most explicit, the most transparent, the most measurable, it's going to be paid with certainty, though it depends on the broker that you might be using. Many online brokerage firms now offer pretty expansive lists of ETFs that trade on a commission-free basis. Even those that you have to pay a commission to trade, what we've seen in recent days is that those commissions have actually come down pretty dramatically. We're seeing sort of a price war in terms of commission charges as well.

But less explicit, less apparent are other transactional costs. Things like the bid and ask spread, which is something that investors pay. So, the bid price, which is quoted for exchange-traded funds is the highest price which a buyer is willing to pay for particular ETF's shares. The ask price is the lowest price which the opposite side of the trade, the seller, is willing to accept for those same ETF shares. And ultimately, to cross that spread the buyer is paying that spread. Now, what we see oftentimes in the case of certainly the largest, the most heavily traded ETFs is that those bid/ask spreads tend to be very narrow, tend to be a penny wide. But it's important to understand that that's a very real cost.

The other transaction cost, which really only applies in the case of someone trying to trade a very large amount of an ETF's shares, move a large dollar volume into an ETF's shares on a smaller ETF or an amount that represents sort of a large amount of the average trading volume in that ETF, could be market impact. So, your actual trading activity could affect the price of the ETF in a negative way, which creates an implicit cost for you, the buyer, if you're buying in very large quantities and pushing that price up and away from what you might otherwise be willing to pay. So, it's important to understand that as well.

Benz: One thing you wrote about in the article that I think is important to pay attention to is, the extent to which you're bothered by each of these sets of costs really depends on what type of investor you are. So, if I'm a long-term investor the holding costs are going to be more impactful for me. If I'm a shorter-term investor engaging in more trading, some of these transaction costs will take a bigger bite out of my returns.

Johnson: That's absolutely the case. So, for most investors with a time horizon that extends beyond three years, transaction costs will be not entirely an afterthought, but framed against and amortized across a very long time horizon, will be de minimis, certainly as scaled relative to the cost of holding that fund over a very long period of time.

Now, if I'm very actively trading, which I by no means recommend, but this also explains sort of the nature of the ETF marketplace and why there exists to this day many funds that charge much higher fees relative to even direct competitors but still have substantial assets, still see substantial trading volume, is that there is a whole trading community that has sprung up around exchange-traded funds that are doing a variety of different things with them in order to make money, in order to pick up fractions of a penny in front of very large steamrollers millions of times on a daily basis. Those sorts of hyperactive traders will tend to place more emphasis on trading costs than they will on holding costs given that their anticipated holding period might be measured in minutes.

Benz: OK. Ben, useful information. Thank you so much for being here to discuss the whole gamut of costs that ETF investors face.

Johnson: Thank you for having me.

Benz: Thanks for watching. I'm Christine Benz for Morningstar.com.