I published an article titled Beware of MLPs in an ETF Wrapper: AMLP on August 27, 2010. Representatives from both Alerian and ALPS contacted me regarding the story. Here is the written response received from ALPS on August 30, 2010:
We are excited to see all the media focus on our new product and appreciate you taking the time to read through our filings and learn about our new ETF. We also appreciate your candor in noting that you’re not a tax advisor. However as you well know, your followers trust you implicitly, and your sentiments carry significant weight with them perhaps more than any tax advisor.
We would like to take this opportunity to address several of the points highlighted in your article and clarify some of the facts surrounding the tax treatment of the ETF structure – not only for your benefit but also for the benefit of your followers.
Perhaps the three most important points to address from your article are –
1) The notion that the corporate tax election chosen by the ETF is novel or in any way differs from industry convention
2) The notion that the ETN structure was chosen first because it is optimal in any way from a tax perspective
3) The notion that there is anything misleading about our expense ratio
We will address each of the points in succession and by highlighting industry convention regarding the tax treatment of MLPs, as well as the pros and cons of various MLP structures we hope to provide you with a better understanding of the rationale behind creating an ETF and its advantages over an ETN.
1) THE CORPORATE TAX ELECTION
First of all, it is important to note that the corporate tax election chosen by AMLP is not novel in any way, shape or form. In fact, EVERY MLP-focused registered investment company in existence today has made the corporate tax election, including the following MLP funds that you and your followers may be aware of:
1) The Closed-End Funds (KYN, TYG, FMO, FEN, NTG, CEM, etc.)
2) The Mutual Funds (The SteelPath Funds MLPZX, MLPTX, MLPOX).
We would also note that we have not yet seen an article, in the six years since KYN and TYG launched telling investors that they should avoid these products because of the tax status, nor has it impacted the billion dollar Closed End Fund launches of CEM and NTG in June and July this summer. Therefore, from an apples-to-apples comparison perspective it is important to clarify that AMLP is no different from a tax perspective than any of the MLP-focused Closed-End Funds or Mutual Funds that have preceded it. It’s only differentiating factor is that it is an ETF and therefore carries the liquidity, transparency and cost benefits that investors have come to associate with ETF structure.
So the question must be raised, why did all of the CEF and MF products that are in existence today choose the Fund structure over the note structure? The answer is that the Fund structure, even with the corporate tax election is inherently a more tax-efficient structure than an ETN. In fact, the only reason why the ETN structure preceded the ETF structure was because the ETF option was not available from a regulatory perspective at the time that the first ETN was launched. Perhaps the most objective opinion on this subject would be from the creators of that very first MLP ETN, Alerian founders Gabriel Hammond and Kenny Feng. Both Gabriel and Kenny would be happy to speak with you on the record about this decision, and in fact I would recommend it as I think it would be very informative for your followers. However, in the interim I will share with you our understanding of their decision process as they communicated it to us during the development of the ETF.
MSSRS Hammond and Feng created the MLP ETN structure, and packaged the very first MLP ETN (the NYSE: BSR actually, the predecessor to the AMJ) as an ETN, because they could not figure out how to structure an ETF. For reasons which we will detail below, the ETN structure was a band-aid, a 3rd best approach. At the time of BSR’s launch, however, it was the only structure that was feasible under the regulations. Every ETN that has come since (the JP Morgan AMJ, the UBS MLPI, etc. has copied this structure, again because it was the only product choice available, not because it was optimal from a tax perspective. Ishares, Van Eck, First Trust, ProShares, Morgan Stanley, UBS – every one of these institutions has been trying for years to unlock the secret to an ETF, long before the ETN ever came out, for that matter, to rid themselves of the detrimental features of the ETN investment. Fortunately, MSSRS Hammond and Feng finally figured out the key to the ETF structure, and they would tell you: the ETF is the optimal structure, and the ETN did not come first out for tax optimization reasons, it came as a second rate solution but it was the best the two could do at the time.
1) Credit risk – When discussing MLP focused investments it is important to highlight that the only reason why the MLP ETNs do not make the same corporate tax election that is chosen by every other MLP focused CEF, MF and ETF is because they are not technically an investment in MLPs, and as such are not required to make the election. Remember, ETNs do not invest in anything per se, but rather are debt obligations of the issuer. This distinction has important ramifications not only from a credit perspective but also from a tax perspective. It is important to remember that all of the MLP ETNs are debt obligations of the issuer and carry said issuer’s credit risk. This is not a risk that we would understate. – We easily forget the days of 2008 and Lehman Brothers, Bear Stearns, UBS and Morgan Stanley themselves nearly going out of business – this is not an insignificant risk and should be considered when comparing the different structures. At minimum one should consider the price of credit default swaps on the issuer as an embedded cost to the ETN.
2) Distribution taxes – The second disadvantage of the ETN structure relates to the tax efficiency, or rather inefficiency of its distributions. Again, MLP ETNs are technically not investments in MLPs, but rather debt obligations of the issuer. Therefore, unlike every other MLP focused structure whose distributions retain the tax-advantages of investing in MLPs, ETN distributions are taxed as interest payments. Again, this is not an insignificant distinction. MLP distributions are large – approximately 7% per year. Investors in MLP ETNs will have to pay corporate income tax on this 7% yield – this equates to a reduction of nearly 250 basis points in the after-tax yield of an MLP ETN. This is not an estimated cost, or a deferred cost, but an actual cost borne by MLP ETN investors that can be calculated with certainty. By contrast, an ETF that invests in MLPs will have the majority of its distributions considered to be return of capital and tax free. On an after tax yield basis, the ETF has a substantial tax advantage. Considering that the majority of the total return generated by MLPs comes from its distributions and that the majority of MLP investors purchase these securities for the yield, the ETF advantage on distributions is critical. Thinking about the ETF vs. ETN returns on an after-tax basis is a critical part of this analysis that cannot be ignored.
3) Volatility – In your article you explicitly mention the negative features of the corporate tax election chosen by the ETF structure. While it is true that the ETF will have to accrue for a deferred tax liability on any appreciation in underlying MLP securities, it is important to remember that the same process works in reverse. In other words, if there is any depreciation in the underlying MLP securities, the ETF accrues a deferred tax asset. In performance terms this means that in a falling market an MLP ETF should fall LESS than an MLP ETN. This is an extremely important point as it relates to the mention in your article of a ‘risk-free arbitrage’ between the ETN and ETF. If for no other reason than your own liability with regards to your readers, we suggest you remove the comment regarding the arbitrage – a reader of yours that takes this trading position, and watches the AMZ I fall 10% in September, will find that while his MLPI position has fallen 10%, while his AMLP position will have only fallen 6.25% because of the deferred tax asset.
While in a rising market an MLP ETN should appreciate faster than an MLP ETF, in order for its total return to exceed that of the MLP ETF the appreciation will have to be so significant that it offsets the tax disadvantage of the ETNs interest payments. For taxable investors the MLP ETN offers an unfavorable asymmetric risk profile – there is certainty of more risk on the downside and there is only more return on the upside if the appreciation in MLP securities is significant enough to offset the tax disadvantage on interest payments.
To summarize the disadvantages of the ETN structure and why Alerian chose to do an ETF, the ETF offers similar access to the MLP market as the ETN with 1) No credit risk, 2) More tax efficient distributions and 3) Lower downside volatility.
It is true that there is no free lunch in finance. The ETF does concede some upside to the ETN in a very strong bull market for MLPs. However, we believe that most investment committees, risk managers and fiduciaries will be willing to concede this upside in exchange for the opportunity to minimize downside volatility and credit risk.
2) EXPENSE RATIO DISCLOSURE
In your article you imply that our expense ratio is misleading. This is simply incorrect. First of all, we heavily disclose the potential tax impact. Second, we cannot give specifics on what the tax impact will be because it is not a constant -it depends on the performance of the underlying index. Unlike the tax treatment on ETN interest payments, which is always a negative, the corporation election can either work in favor of or against the ETF which makes it impossible to forecast or calculate in the prospectus. It is important also to highlight that neither the closed end funds (KYN, TYG, FMO, FEN, NTG, CEM,etc.), nor the mutual funds (The SteelPath Funds MLPZX, MLPTX, MLPOX) account for the taxes in their expense ratios. We invite you to take a look at the disclosure in these companies’ N-2s and N-1As, respectively. Many in fact on the Closed-End Fund side do not even disclose the tax drag in relation to their expenses, whereas we do (as does SteelPath). For us to try to put in a rote expense plug would be in complete violation of an industry practice that’s going on six years.
Again we do appreciate the interest that you and your followers have shown in our ETF. We sincerely hope that this information provides you with more background behind not only the rationale of the ETF structure but also for the industry convention in the MLP space.
We would ask that you consider retracting or redacting your article until you have the opportunity to follow up with us, and or Alerian, or other third-party independent sources regarding the corporate taxation of Master Limited Partnership investments.
Author’s name and title redacted
1290 Broadway, Suite 1100
Denver, CO. 80203