In the investment process, once a high-level investment thesis or allocation decision has been made, the question shifts to how one can best gain exposure. While there are certainly a number of considerations, chief among them is cost.
This shouldn't come as a surprise. For years, cost of access has been a paramount concern. We can trace the trend all the way back to 1975, when, in founding The Vanguard Group, Jack Bogle built an investing empire on the back of low-cost investing. Even our research here at Morningstar has reinforced the idea that costs make a difference. We've shown that over the long term cost is an excellent predictor of relative performance; that funds with lower fees tend to outperform peer offerings.
Within ETF land, the traditional view of cost can be summarized with two words: "expense ratio." Investors often look to this metric as the all-in measure of an ETF's cost. Generally speaking, the expense ratio is a good yardstick. There are, however, a number of intangibles that, while frequently overlooked, do present the investor with very real costs.
I'd like to frame the question a bit differently here. I posit that when considering ETF costs, one should not merely ask what they charge. Instead, ask yourself what you initially expected to receive and whether or not it was what the product has provided historically.
Estimated Holding Costs
Most ETFs are index vehicles, and ideally, these funds should track their index, less the expense ratio. On our US website, under the performance tab, there is a table called "Total Cost & Risk." Our proprietary Morningstar data point that measures the gap in return between the ETF and the index is called "Estimated Holding Costs." In addition to the expense ratio, the estimated holding cost captures the realized cost of replicating an index. Indexes with high turnover or relatively illiquid constituents can be more costly to replicate. For funds that track these types of indexes, we would expect the estimated holding cost to be higher than the expense ratio.
Take for instance two funds tracking very similar indexes, the Dow Jones U.S. Total Stock Market and the Dow Jones U.S. Index. They are tracked by SPDR Dow Jones Total Market (TMW) and iShares Dow Jones U.S. Index (IYY), respectively. Given the similarity between the benchmarks, either could be used to gain the same exposure. They maintain a correlation of 1.00 over the past three years. Both funds charge exactly 20 basis points per annum. Year to date, IYY has returned 7 basis points less than the index. Given that we are roughly halfway through the year, the tracking error falls close to the size of the product's annual fee. The SPDR product, however, has trailed its index by a larger 18 basis points, year to date.
We can attribute this additional lag to the fact that the SPDR fund is tracking a total market index which includes all U.S. securities available to investors, 3,718 securities in total. The iShares fund tracks an index that includes only the top 95% of the available market cap, amounting to a total of only 1,340 securities. In other words, we think the reason for the SPDR fund's greater lag to the index is due to the fact that small-cap stocks are harder to track, given their lower liquidity.
Generally speaking, the gap in return between the ETF and the index can be thought of as an additional cost. Note, however, that these figures will fluctuate from year to year and sometimes can even be positive. Significant and persistent deviations from expected levels of excess return should raise eyebrows. A good way to gauge the potential for deviation from the index is the data point "Tracking Volatility" (which can be found next to the estimated holding cost figure), or rather the volatility of the excess return figure over time. Again, given TMW's greater exposure to small caps, we would expect this fund to have a higher Tracking Volatility figure relative to the iShares fund. As such, investors looking for the best index-tracking ETF would be better off in the iShares fund. Those who are more bullish on small caps and want that exposure can pick the SPDR fund, the trade-off being that this ETF won't track as well as the iShares product.
Market Impact
A second intangible cost comes by virtue of a product's liquidity. Liquidity can be thought of as the sensitivity of an asset's price to the act of buying or selling. Assets that can be bought or sold in large quantities without generating large impacts on the asset's price are said to be highly liquid. Conversely, assets that suffer large movements in price in response to small levels of buying or selling activity are said to be illiquid.
If the size of an order to buy or sell exceeds the number of shares being offered at the most advantageous price (for buyers the lowest price available, and for sellers the highest price available), to fill your order you must accept the next best price. Consider the following example:
You would like to purchase 10 shares of exchange traded fund XYZ. The best offer is $5.00 per share. There are five shares available at this price. You purchase all five. To complete the trade, however, you must purchase an additional five shares. The next best price available is $6.00 per share. There are 17 shares available at the $6.00 price. To complete the order, you purchase an additional five shares at the $6.00 price. The total purchase price of the order was $55.00.
By purchasing shares, you moved the price higher. In effect, your trade had an impact on the market price.
Illiquid markets provide for very serious implications. In the example above, you initially wished to purchase shares of XYZ at $5.00, but to fill this order immediately, you had to purchase half of the order at $6.00. Assuming no transaction fees, you spent a total $55 to acquire 10 shares. Had you expected to fill his entire order at the initial $5.00 level, the trade cost you 10% more than you expected it to. This exemplifies the idea that market illiquidity can pose a real cost.
While there are sophisticated statistical metrics that can be used to gauge and project market impact cost, they aren't necessary for the disciplined, long-term retail investor. Generally speaking, volume can serve as a decent proxy for liquidity. That said, trading in ETFs that maintain higher average daily trading volume numbers will leave you far less exposed to suffering these market impact costs.
Wrap Up: Best Practices
We don't aim to discount the relevance of the baseline expense ratio in your investment selection process. We do, however, posit that there are additional costs that are often less transparent but nonetheless important to consider. For those making decisions purely based on annual fees, note that the incidence of negative excess of return can represent a sizable cost relative to the expense ratio. While the excess return figure can fluctuate over time, any excessive or persistent substantial negative excess return should be taken into account. Also consider the impact your own trading activity may have on the ETF. If your trades are sizable, sticking with the most-liquid options can minimize the cost that your own market impact poses.
We recently published an in-depth study on total cost of ETF - “ETF Total Cost Analysis in Action”. We hope it will help investors understand the actual “cost” being “charged”.
Abraham Bailin is an ETF Analyst with Morningstar
