In the ETF world, most investors pay scant attention to legal matters unless there’s a default or lawsuit. That’s a mistake. Understanding an ETF’s legal structure could lead to significant cost savings over time.
Investors focus on investment fees and rightly so. Every penny paid in fund expenses is a penny that can’t grow for future use. While many investors look at fund management fees, ETF and index mutual fund investors have a more accurate way to gauge costs. Cumulative tracking difference shows the gap between the fund’s objective of providing the returns of a particular index and the actual returns.
Tracking difference comes in handy for anyone looking to understand the long-term costs of holding an ETF. If your target index returned 15.40% last year, while the ETF tracking it returned only 15.25%, you have a pretty good indicator that, no matter the stated expense ratio, your actual holding costs should come to around 0.15% per year.
When calculating tracking difference, the key is to match the treatment of cash flows for the fund and the index. If one is accounted for differently than the other, tracking difference will be misstated.
ETF returns are best measured by the fund’s net asset value (NAV). Unlike the ETF’s market price, which is quoted at the time of the last ETF trade, NAV is calculated based on portfolio holdings’ closing prices. NAV is a true reflection of the portfolio’s value, free from trading gaps, premia, or discounts.
An ETF’s total return NAV includes the reinvestment and compounding of dividends over time. Total return NAV is an accurate reflection of the investor experience in that it reflects both market appreciation and portfolio yield. Total return NAV can differ dramatically from regular, or price return, NAV, especially when yield is a major component of the overall return. The chart below shows the returns difference since inception for iShares Core U.S. Aggregate Bond ETF (AGG-US).
The difference is dramatic. In the 15 years since AGG-US launched, it has returned 3.75% per year, annualized through May 8, 2018. Yet the price return calculation would suggest a mere 0.32% annualized return. The total return version shows not just the magic of compound interest, but the most basic of bond fund features: coupon payments. All returns comparisons, including tracking difference, require total return NAVs.
When measuring the performance of an ETF, total return NAV must be matched with total return index data. Calculating tracking difference including yield for the fund but excluding it for the index (using a price or excess return rather than a total return variant) is hardly an apples-to-apples comparison. This is a misleading practice called sandbagging.
Tracking difference comes in handy when analyzing complicated asset classes with multiple legal structures on offer. This is especially true in commodities where there are four possibilities: grantor trusts, exchange-traded notes, commodity pools, and actively managed open-ended funds. Different legal structures could have different total costs because of their portfolio management requirements. Tracking difference, calculated on a uniform basis, allows investors to select the most efficient legal structure.
Commodity grantor trusts offer shares backed by precious metals. Grantor trust expenses include vault leases and security personnel, but not brokerage fees. There’s no portfolio manager, because the trust assets are not variable; there’s never a re-balance. It’s pretty much bars in, bars out, with a few sold every month to pay for upkeep. Commodity grantor trusts do not make distributions, nor do the trust assets generate income. There is no difference between regular and total return NAVs for commodity grantor trusts.
Exchange-traded notes (ETNs) also have no portfolio manager. Similar to structured products, ETNs are nothing more than a contract between an investor and a bank, where the bank promises to deliver a payout based on an index level. Nearly all commodity ETNs, including those from Barclays, Credit Suisse, and UBS, specify that they track a total return index. ETNs represent unsecured debt obligations of the issuer rather than shares in a portfolio. ETN issuers are obliged to deliver a pattern of returns, but it’s the business of their risk desks to hedge that obligation. Because they mostly track total return indexes, ETN NAVs are generally stated on a total return basis, although a handful actually distribute income.
Commodity pools invest directly in commodity futures. These ETFs need portfolio managers, and not just for the futures positions. Commodity pool operators are tasked with investing the cash balance, often called collateral. Dollar-wise, collateral makes up most of the portfolio weight, because futures positions have zero initial cost, and their prices are re-set to zero upon daily settlement. While managers need to keep some collateral in a margin account to meet daily settlement requirements, most of it can be deployed into margin-eligible securities such as T-bills or U.S. government securities in money market funds. In these days of rising interest rates, the collateral yield can be significant.
Because futures have short lives, futures-based portfolios require frequent re-balancing. This means that commodity pool operators must intervene often, incurring brokerage charges, and sometimes accruing additional tracking error as they position portfolios ahead of index rolls.
Commodity pool prospectus documents often state that funds track futures price-only indexes (called excess return), but their NAVs include the collateral and its income. Collateralized commodity pool NAVs are effectively total return NAVs, even if the pools make no cash distributions. For this reason, issuers like Blackrock, Invesco, United States Commodity Funds, and WisdomTree benchmark their commodity pools to total return index variants.
Actively managed open-ended funds are the new kids on the commodity ETF block. Billed as a no-K-1 alternative to commodity pools, these ’40 Act funds achieve commodity exposure via offshore subsidiaries. Because there is a limit of 25% to capital allocated to the subsidiary, these funds qualify as diversified per registered investment company (RIC) specifications. The subsidiary can gear the futures exposure to match the overall asset level without issue, as futures are priced at $0.00 daily, subject to margin settlement requirements.
Clearly, these activities require significant portfolio management. Like commodity pool ETFs, actively managed open-ended funds in the commodity space require frequent re-balancing, which brings with it brokerage charges and invites the possibility of incurring market slippage in trading executions. Market slippage is the cause of tracking error. Yet as actively managed funds, these ETFs do not technically track indexes, as the SEC requires these funds to file for actively managed exemptive relief. However, many state that their investment objective is closely linked to delivering the return of a specific index.
Like many open-ended funds, actively managed commodity ETFs distribute income to shareholders. Total return NAVs must be calculated, and will accumulate value faster than regular price NAVs.
Measuring the operating efficiency of each structure by calculating annual tracking difference is a worthwhile exercise, because there are clear bargains to be had.
Let’s compare the tracking difference for funds that track (or allude to tracking) identical indexes. The table below shows the one-year median tracking difference for ETNs, commodity pools, and active open-ended ETFs that track four different commodity indexes, contrasted with the expense ratio. Note that all NAVs and index variants are total return.
True to their mandate, the ETN NAVs reflect the returns of the indexes tracked, minus expense ratios, within one basis point. This ETN structure allows for excellent index tracking. Note that Barclays iPath ETNs have been excluded from this analysis, because iPath ETNS historically used path-dependent expense ratio calculations.
The commodity pools and open-ended funds did not track their indexes as well as the ETNs. All under-performed their expense ratio-based predictions, though the commodity pools delivered less of a gap than the actively managed open-ended funds. In the cases where both commodity pools and open-ended funds tracked the same index, the results were mixed. The PowerShares DB Optimum Yield trackers delivered results within 0.04% of each other, while the GSCI tracker ETFs showed a clear advantage for the ’40 Act structure.
Looking at the wider commodity ETF landscape, we can extend this analysis, and include grantor trusts. A comparison of 35 commodity-tracking ETFs with available total return index data showed that the patterns of the same-index examples held. Grantor trusts and ETNs tracked their indexes within 0.01% of their expense ratios, while commodity pools and open-ended funds posted bigger gaps. While commodity pool gaps were smaller, their higher overall expense ratios drove their overall tracking difference higher.
| Legal structure | Average tracking difference | average expense ratio | average gap |
| Grantor Trust | -0.34% | 0.35% | 0.01% |
| Exchange-Traded Note | -0.65% | 0.64% | -0.01% |
| Commodities Pool | -1.10% | 0.82% | -0.27% |
| Open-Ended Fund | -0.78% | 0.37% | -0.41% |
Source: FactSet
The difference between the commodities pool and open-ended funds might surprise some, as both structures require significant trading activity in the futures market in addition to managing margin accounts and collateral investment. Interestingly, FactSet’s calculated commodity pool expense ratios include estimates of brokerage and creation costs, while the open-ended funds do not, as trading cost estimates are not provided in traditional ’40 Act prospecti.
A bit of caution is warranted in interpreting these results, as most of the actively managed open-ended funds have short histories. Tracking difference could well approach that of the commodity pools as time passes. At present, though, the open-ended funds look to be the better deal in comparison to commodity pools. Yet both seem inefficient in comparison to ETNs and grantor trusts.
Tracking difference, calculated properly using total return NAVs and index levels, allows investors to make apples-to-apples comparisons between legal structures in the commodity space. This holding cost analysis, along with an assessment of trading costs and tax treatment allow for a total cost of ownership calculation.