We also go under the hood of some asset-specific costs.
Last month, my colleague Abraham Bailin wrote an article titled "The Intangible Costs of ETF Ownership," which explained the concept behind Morningstar's proprietary cost data points. (The data points can be found under the Performance tab.) In this article, we will discuss how to use these data points when comparing funds within a category.
First, a quick review of the data points and their definitions. Estimated holding cost reflects how much a fund's NAV has lagged its index over the last year. This can be considered the fund's "cost" to replicate its underlying index. The largest component of this figure is usually the fund's expense ratio, but other contributors may include high turnover (which drives higher trading expenses), higher expenses due to exposure to less-liquid securities (such as small caps or high-yield bonds), sampling (which can result in positive or negative alpha), and share lending (which is a source of revenue and therefore offsets other costs). EHCs are calculated as the geometric difference between the index return and the fund return over the past year.
Tracking volatility measures the volatility of a fund's net asset value performance versus its benchmark index. This is measured as the standard error of the regression of the NAV's performance versus that of the index. The lower, the better. A potential driver for higher tracking volatility is large flows in and out of the fund as a percent of total assets under management. For example, during creations and redemptions, baskets with adjustment for dividends can be off, which can drive tracking volatility. For exchange-traded funds with negative EHCs, investors can check the tracking volatility figure--if the tracking volatility is high, then the negative EHC may be attributable to sampling, and it is unlikely that the fund will be able to consistently outperform its index.
ETFs in the large-cap categories have estimated holding costs nearly in line with their expense ratios given the liquidity of the underlying holdings. ETFs with significantly higher EHCs relative to their expense ratios tend to be funds that track broader indexes that include small caps (which can be less liquid) or funds that track fundamental indexes, which tend to have higher turnover. Not surprisingly, small-cap ETFs have EHCs that are higher than their expense ratios, relative to large caps. Overall, the ETFs with the largest assets under management (which covers different size and style categories) have the lowest EHCs and tracking volatility--in other words, these ETFs are incredibly efficient.
For international-equity ETFs, it is not surprising to see many more funds with high tracking volatility (combined with negative EHCs)--this is consistent with the fact that many foreign equity ETFs use representative sampling. There is also another reason for this--Vanguard, as well as some of the smaller ETF providers, employs fair value pricing. Vanguard MSCI Emerging Markets ETF VWO and iShares MSCI Emerging Markets Index EEM track the same index, but the former had a tracking volatility of 5.81 whereas the latter had 1.21. The reason for this discrepancy is because Vanguard employs fair value pricing to its NAV whereas iShares does not. Mutual funds that invest in foreign securities are required by the SEC to employ fair value pricing, and, because Vanguard ETFs are a share class of their corresponding mutual fund, Vanguard ETF NAVs have to be fair valued. (Mutual funds price their NAVs after the U.S. market close and will apply fair value pricing to their foreign securities holdings whose prices at that time are "stale.") Because ETFs trade during U.S. market hours, it can be argued that the ETF's price is the fair value, and iShares cites this as the reason why it doesn't fair value its ETFs' NAVs.
If the only difference between EEM and VWO was that the former did not employ fair value pricing and the latter did (and assuming that both employed full replication and charged the same expense ratio), we would expect EEM's market price to show greater premiums and discounts relative to its NAV and that its NAV would closely track its underlying index, whereas we would expect VWO's market price to show lower premiums and discounts relative to its NAV and that its NAV would not track its index as well as EEM. But more important, we would expect the price performance of EEM and VWO to be essentially the same, which is important for investors because they buy and sell at market prices and not at NAVs. While EEM has been able to incur a lower EHC relative to its expense ratio, or ER, over the last year, it is not guaranteed that the fund managers will be able to continue to provide this small amount of alpha. Given VWO's significantly lower ER (which is a much more stable figure than EHC), we prefer it over EEM.
For fixed-income ETFs, another cost that is reflected in the EHC is the positive or negative alpha from the cash component of the creation basket. Because many fixed-income securities are not very liquid, some fund providers allow authorized participants to submit creation baskets with a cash component in lieu of certain securities. The fund company then uses the cash to purchases securities to complete the basket.
SPDR Barclays Capital High Yield Bond JNK and iShares iBoxx $ High Yield Corporate Bond HYG provide fairly similar exposure, however, JNK's EHC has been significantly higher than that of HYG for the year to date. SPDRs allows for more cash in its creation baskets, relative to iShares, and it is possible that SPDRs has not been able to efficiently put that cash to work over the past few months, as there has been strong demand for this relatively illiquid asset class, which results in wider bid-ask spreads. Fixed-income ETF NAVs are based on the bid prices of their underlying securities, but in practice, the actual purchase price can be much higher, especially during periods of illiquidity. If a fund company takes a creation basket that is part cash, higher bid-ask spreads for the underlying holdings will drag on the fund's NAV (this can be observed as the gap in performance between JNK's index and NAV, which is greater than that of HYG). In comparison, the iShares product takes creation baskets with less cash, so the additional cost of wider bid-ask spreads on high-yield bonds is reflected in higher market-price premiums to NAV. It is possible for fund managers who take creation baskets with a cash portion to provide positive alpha, but we note that for JNK, that was likely not the case--over the past three years, JNK's NAV has trailed that of its index by 184 basis points, annualized, while HYG trailed its index by 25 basis points. Based on this information, we would prefer HYG over JNK.
Another example of the intricacies of fixed-income ETFs with illiquid underlying components can be found in Timothy Strauts' article, "We Love This ETF: Sell It Now."
Vanguard also employs fair value pricing for its fixed-income ETFs (as fixed-income markets close one hour prior to equity markets), which can contribute to higher tracking volatility for its funds but should not negatively impact long-term performance. Below, we highlight again that despite Vanguard Total Bond Market ETF's BND high tracking volatility, its price performance relative to its underlying index has been almost in line with that of iShares Barclays Aggregate Bond AGG (these two ETFs track very similar indexes).
Generally speaking, international-equity ETFs and fixed-income ETFs incur higher costs because the underlying assets are more expensive to gain access to, relative to U.S.-equity ETFs, and not because they are faulty products. Other vehicles, such as open-end mutual funds, also face similar costs. Exchange-traded notes, however, are a different story.
ETNs are uncollaterized notes issued by banks that promise to provide the returns of the stated underlying index (sometimes advertised as offering perfect tracking), less fees. These banks use derivatives and certain strategies to hedge its position and deliver returns. As such, we would expect an ETN's EHC to be in line with its expense ratio. This is not the case, as the weighted average of the group's EHC minus the expense ratio of 0.17 is the highest among the four categories. This reflects the fact that some ETNs have additional and sometimes opaque fees, such as brokerage expenses, financing fees, and path-dependent fees. In fact, an investor can use an ETN's EHC to determine if an ETN has additional fees, which are usually not explicit in the note's fact sheet and other marketing materials. More details about the risks of ETNs can be found in my colleague Samuel Lee's article, "Exchange-Traded Notes Are Worse Than You Think."