Update: An analyst on another investing site recently argued that the credit risk that ETN investors court is analogous to the credit risk that investors unwittingly take when investing in leveraged or short ETFs, as those ETFs employ derivatives that pose counterparty risk.
I don't think that's so for a few reasons.
For one, while I can't give anyone chapter and verse on how those derivatives work in practice, my impression is that the short and leveraged funds concerned are most definitely NOT posting 100% of the derivative's notional value as collateral. Contrast that with ETNs where, in effect, investors are doing just that--handing over every last cent they wish to invest to the issuer concerned. Thus, the credit exposure is inherently different. Second, unless I'm mistaken, any derivative that a short or leveraged ETF enters would have to be cash-settled on a daily basis (how else to meet the fund's goal of delivering x times the daily market return). Therefore, the fund's maximum exposure at any given point would be equal to the difference between what it has agreed to pay and what it's entitled to receive (in the case of a total return swap, that is). Once settlement takes place, both sides are all-square and the process begins anew the following day. Again, contrast that with an ETN, where an investor CAN settle at any time (by selling) but remains exposed for as long as she holds the investment (i.e., there's no daily settlement). Stated differently, an ETN issuer's credit risk increases with time (assuming the index to which it's tied increases, that is). Funds that employ derivatives don't face that kind of risk. Finally, many ETNs pay no dividends or periodic interest income. Presumably, any such income is factored into the derivative instruments that leveraged and short ETFs utilize, meaning that investors effectively receive any income on a daily basis. There, too, those funds have less exposure to the counterparty.
I could be wrong, but it looks apples and oranges to me.
Summary: In the spirit of brevity, here's the long/short of what I'm saying below--I think ETN pricing is out of whack. The following lays out my rationale.
(Disclosure: Morningstar licenses its indexes to certain ETF and ETN providers, including Barclays Global Investors (BGI), First Trust, Claymore, and ELEMENTS, for use in exchange-traded funds and notes. These ETFs and ETNs are not sponsored, issued, or sold by Morningstar. Morningstar does not make any representation regarding the advisability of investing in ETFs or ETNs that are based on Morningstar indexes.)
I've fielded a number of reporter calls recently along the following lines: We've seen a lot of ETNs launch thus far in 2008 - how come?
The (More) Obvious Reasons
There are a few reasons, some more apparent than others. Let's start with the more obvious ones first:
- Typically, investability is a problem with futures-based commodities indexes like the Dow Jones/AIG or S&P Goldman Sachs commodity indexes, and variants thereof. For instance, rolling futures can be a costly endeavor, making it more difficult to track the benchmark closely. ETNs are better suited to tracking those areas. Indeed, ETNs surmount the challenge associated with transaction costs by removing transactions from the equation altogether-the ETF issuer promises to pay the index return, which is devoid of transaction costs. How the issuer comes up with the money needed to deliver that return is its problem, not the ETN investor's. Problem solved.
- Investors are seeking a more tax efficient way to invest in commodities and currencies, gains on which are typically subject to more punitive tax rates. Enter ETNs, which, for commodities and multi-currency strategies at least, enjoy clear-cut advantages over mutual funds. For instance, a commodities ETN investor won't pay a nickel in taxes until she sells her investment. And any gains are taxed at the 15% long-term capital gains rate. Contrast that with other structures, including ETFs, which would have to distribute any realized gains/income at least annually. Further, any gains would be subject to the so-called 60/40 rule-60% of the gains are eligible for long-term capital gains rates, the other 40% are subject to much higher short-term rates. In other words, ETNs are exploiting a loophole in the tax code for all it's worth.
- The glass-half-full part of me says, reassuringly, ‘as commodities have become a more prominent part of the investing scene, investors have grown more knowledgeable about commodities investing and sophisticated in their needs, explaining the demand for a wider variety of commodity-based investment products.' The glass-half-empty part? It snarls ‘More sophisticated, my shoe. They're performance chasing, plain and simple.' As with so many things, the truth probably lies somewhere in the middle.
The Less Obvious Reasons
But why else? My take is that the economics of these notes, and the relatively exclusive nature of the ETN market, are quite attractive to the fund companies and issuers concerned. How so? Let me count the ways:
- There is no management team behind these, or at least not in any traditional sense. The issuer promises to pay the stated return. But how the issuer delivers that return is up to it. In other words, there's no portfolio being assembled. Investors are buying a promise - a synthetic.
It stands to reason that this is a very cost-effective proposition. True, the fund company has to share a portion of the expense ratio with the issuer. And, yes, the issuer presumably has some kind of personnel and resources tasked with hedging out the firm's exposure to the notes (i.e., investing the capital they've received so as to capture the return they've promised to pay...and then some). But let's get real-in some cases, most notably the iPath family of ETNs, the issuer (Barclays Bank) is an affiliate of the fund company (Barclays Global Investors), making the fee split little more than inter-company accounting legerdemain.
And when the issuer is unaffiliated-as with the Elements and Market Vectors families-the market maker concerned isn't hiring an army of personnel to manage the firm's exposure to these instruments. Instead, they just tap existing resources, the epitome of leverage. Again, the incremental cost of an ETN to the issuer is likely a rounding error. The fees ramp with assets. You do the math.
(Full disclosure: One of the Elements ETNs licenses the use of a Morningstar index.)
- It's free capital. No, scratch that - it's accretive capital, as the investor is paying the issuer a fee (the portion of the expense ratio it receives) in forking over her cash. Ok, so the issuer has to put that capital to work in order to hedge out its exposure to the notes. But look at the roster of firms that are signing up to issue these instruments-global banks and market-makers, all. Risk management and arbitrage is the lifeblood of these organizations (or, as recent events well attest, a pathogen when it fails). So, one can't help but sense that no-strings-attached capital is a very sweet deal to the issuer, especially at a time when the banks seem reluctant to lend to one another amid the credit crisis.
- To borrow a bit of Morningstar parlance, ETN issuance is a ‘moat'y' business. That is, it's founded on defensible competitive advantages.
Generally speaking, you need a few things to issue these instruments-a balance sheet, securities inventory, and some faculty for risk management and trading. Asset managers could, in theory at least, issue these on their own balance sheets, especially if they own a chartered bank (as T. Rowe and Franklin Resources, to name a few, do). But asset managers don't make markets in securities, and probably aren't especially keen to start doing so. Further, an asset manager's business isn't predicated on managing risk per se, let alone in managing a sprawling securities inventory. Also, as a practical matter, asset managers--firms which are already heavily levered to the capital markets--probably aren't chomping at the bit to lard their balance sheets with debt, as it would amplify the volatility of their results. So if you ‘x' out the asset managers, who does that leave? You guessed it--the big banks and broker-dealers.
One could of course argue that the relatively low upfront costs involved in launching an ETN-they're not '40 Act funds and thus aren't freighted with many of the costs that normally apply to mutual funds and ETFs-opens the ETN market to greater competition, as the barriers to entry are ostensibly lower. And, in fact, many of the banks that have issued ETNs have probably taken a shine to the space because they can enter it without expending untold sums of capital in the process.
But it's not as if the local thrift is going to be able to launch an ETN. It requires scale, know-how, and, heaven knows, financial strength. Thus, I would expect that at least part of the allure of ETNs stems from the fact that it plays to the banks' traditional strengths and doesn't demand a hefty upfront investment, yet is exacting enough to keep most potential new entrants at bay.
Fair Shake?
If ETNs are a great deal for the fund-company and issuer concerned, it's only natural to wonder whether the party on the other side of the trade-the ETN investor-is getting a fair shake. That is, to ask whether ETNs are priced fairly?
At first blush, ETNs look like a pretty good deal. They levy reasonable fees in absolute terms-typically anywhere from 0.30% to 0.95%. And they still offer all of the benefits of traditional ETFs-transparency, liquidity, they can be shorted and margined, etc. Plus, as noted, they give investors the ability to access heretofore difficult-to-reach segments of the market, and do so with remarkable tax-efficiency.
But what makes ETNs inherently different than traditional mutual funds and ETFs is the credit risk that they pose to the investor. They're junsenior, subordinated, uncollateralized debt instruments. So, if an ETN issuer goes belly-up, ETN investors line up in bankruptcy with all of the other creditors. That's a risk that mutual fund and ETF investors most definitely do not court.
Some have argued that these banks are so strong that credit risk goes out the window. But it's not as if the capital markets give the most financially stable, fortress-like companies a free ride. Last I checked, for instance, if General Electric floated commercial paper, it had to pay interest on those borrowings. And so forth. Recent events-pocked by the near-collapse of one issuer, Bear Stearns, to whom the ‘too big to fail' catch all had formerly applied-pretty much cut that argument to ribbons.
In that sense, evaluating ETN fees on an absolute basis, or pitting them against some benchmark, like the average cost of a comparable ETF or mutual fund, is misleading. It doesn't adequately account for the incremental credit risk that ETN investors assume.
Whither Counterparty Risk
So, are ETNs being priced in a way that takes counterparty risk properly into account?
To answer that question, we must turn to the derivatives market for comparable examples. There one finds what's arguably the closest analog to ETNs-total returns swaps. For those unfamiliar, a total return swap is a credit derivative in which one party agrees to pay another party a specified return (say, the T-bill) on some notional amount of capital (say, $100 million) in exchange for a different specified return (say, the Dow Jones/AIG Commodities Index) tied to that same notional amount of capital. When the swap commences, the payor and receiver don't literally exchange $100 million in capital and then go about investing those monies in order to generate the return they've promised. Instead, they simply net out the returns at specified intervals. So, if DJ/AIG logs a 10% gain and the T-bill returns 3%, the T-bill-payor/DJ AIG-receiver would pocket that 7% difference. Assuming annual settlement, that would translate to a $7 million payday (7% difference multiplied by the $100 million notional value of the swap).
Now, you might be wondering how ETNs are similar to total return swaps. In several respects, they're not. For instance, whereas there's no exchange of notional principal in a total return swap, here ETN investors are actually handing their capital over to the issuer. In addition, the ETN investor isn't promising the ETN issuer any return, the T-bill yield or otherwise. Instead, she's giving that issuer her capital and, in essence, saying ‘do with this what you may'. Furthermore, unlike a typical total return swap, there's no periodic settlement schedule or term to the agreement. An ETN investor can redeem-that is, sell-at any time with no strings attached, but she won't receive any payment on her investment until she does so. Finally, total return swaps are a form of leveraged investing-I don't have to put up $100 million in order to get that equivalent economic exposure-while ETN investors are putting up every last penny of capital and earning a return on those monies and those monies alone.
But the economic substance is quite similar. For instance, it's not at all unreasonable to think that the issuer, in receiving ETN investors' capital, can put that money to work at T-bill-like rates. And the issuer is already promising to pay a specified index return. And the fee that ETN investors pay...what might we liken that to? What I didn't mention is that the T-bill-payor in the swap example I presented would pay a percentage spread atop the T-bill. Why? To compensate the DJ/AIG-payor for incremental counterparty risk and for its services in delivering the index return.
Going Apples to Apples
Given that, what you'd ordinarily expect is for the overnight swap spread to exceed the expense ratio that ETN investors are paying. Why? ETN investors shouldn't have to compensate ETN issuers for any incremental counterparty risk. Remember-the ETN investor has given her capital to the ETN issuer, meaning that the issuer has no exposure whatsoever to counterparty risk. Thus, the ETN investor should only have to compensate the issuer for the service it is providing in delivering the index return. (The spread on an overnight swap is the most relevant measure given the liquidity of ETNs-as mentioned, investors can buy and sell them at any time, making swaps with the shortest duration the best litmus test.)
What's the spread on a typical overnight T-bill-for-DJ/AIG total return swap? It'll vary depending on the creditworthiness of the counterparty, the settlement terms, and other factors. But according to a source at a large money manager, one could have recently entered into such a swap at a 15 basis point spread.
And what does it cost to own iPath Dow Jones/AIG Commodities Index? 0.75% per annum, or five times the spread on a T-bill-for-DJ/AIG total return swap. That's not what one would expect.
What could explain the disparity? Ordinarily, you'd expect a large, lump sum, private party transaction (i.e,. total return swap between a broker-dealer and an institutional investor) to be less costly to the issuer than a series of smaller-dollar transactions with a widely diffused investor base in the open market (i.e,. ETN). And, yes, retail investors probably ought to have to pay a tad more for access to the commodity and currency markets than an institution would-size has its advantages, after all. But it seems like an awfully big stretch to claim that these differences adequately explain the cost disparity.
Minding the Gap
What ought to happen in a situation like this? The notes should trade like a zero coupon bond, where they're initially issued at an amount less than face value and then gradually accrue to face as time elapses. Here's how I'd foresee this working:
An ETN issuer registers to issue a certain amount of notes, say $500 million par. Let's suppose that the notes have a stated term, say 30 years. On the day the ETNs begin trading, assume that each note has a $10 face value. If we were to assume that investors deserved some minimal amount of compensation for the counterparty risk they're assuming, say 1% per annum, then those notes would trade at a roughly $7.42 net asset value (NAV) out of the gate-that amount approximating the present value of an instrument that's worth $10 thirty years hence. What would those notes trade for the following day? Supposing that the index the ETNs are linked to is flat, the notes' NAV should inch up ever so slightly, the increase reflecting the value that's accrued with the passage of a day. Thus, the note's NAV on a given day would be a function of three factors:
- The return of the index to which the note is linked.
- Market supply/demand
- The note's maturity and the issuer's risk profile
This calculus would hold for any note that's issued, regardless of when the issuer floats it. As such, if the issuer in the example above turned around and issued another $100 million worth of notes on day two, it would do so at the prevailing market price-a price that reflected the accrued interest on the note. (This assumes that the ETN's price approximates, if not equals, NAV.) In that way, the issuer doesn't lose any of the flexibility needed to issue and redeem notes, while the investor receives compensation for the credit risk she's assuming.
And what of the risk that an ETN's price could diverge from its net asset value because, say, the market is mispricing the interest that's accrued on the notes? For instance, suppose an ETN is trading for $550 when it should be trading for $551. In that scenario, an arbitrageur would purchase the shares for $550 and then sell then at NAV, pocketing a $1 risk-less profit. In so doing, the arbitrageur pushes the ETN's market price toward NAV, and so forth.
Conclusion
ETNs ought to trade more like...well...the bonds that they are. ETN issuers might frown at the complexity and higher cost that such a change would entail. But the economics of these notes--which appear to favor issuers at the expense of investors--demand it.