Authored by Peter Tchir via Brean Capital,
VIX ETFs and ETNs
We take a deeper dive into the strange world of VIX ETFs and ETNs. We take a quick look at the incredible short interest in both the long VIX products and the short VIX products. This massive short interest in both long and short products seems unique to the VIX world (it reflects a re-balancing trading strategy that works in the VIX space because of the high volatility of the VIX products) (VIX ETFs seem to run 5 to 10 times the realized volatility of the S&P 500).
Then we dig into the prospectus for each of the 4 funds I focus on (VXX, UVXY, XIV and SVXY).
What is important is the Acceleration Event in XIV. The language from the prospectus seems clear that if the VIX Short Term Futures Total Return index moves 80% in a day, then XIV has to unwind. While an 80% move in a single day is very unlikely, I believe that the lower VIX goes, the easier it is for it to occur (VIX at 8 only needs to jump to 14.4 in a day for this to occur), because these VIX products make a ‘mistake’ in my view of converting changes in VIX to percentage changes to provide returns (the VIX futures do not do that for example).
The problem with an XIV acceleration event is two-fold – all of the hedges (short futures positions it has) will need be covered, just as the market is struggling. The second order problem is that many investors who have been waiting for a spike in VIX to sell volatility won’t have an outlet. If you planned to buy XIV on a VIX spike and it isn’t there, what do you buy? It is far less clear what sort of trigger mechanism SVXY has (that is something we are looking into).
While I talk about 80%, several funds have mentioned that they see a 60% move as the start of the unwind process in the market as a whole.
On some regression analysis (which I will try and add over the weekend), something like a 3% drop in a single day would be enough to trigger that sort of VIX avalanche – again, not a high probability, but worth noting.
There may also be a bias for events to occur shortly after VIX futures contract expirations as the bulk of the index is in the more volatile, and lower priced, VIX futures contract – but that is just thinking out loud.
One trade, as a result, seems to be an uptick in volumes and activity in higher strike VIX calls. The premise being that if VIX goes > 60% it is going to have a shock much higher. So rather than spending a lot of premium on puts, spend as little as possible to buy some lottery tickets in the unlikely event that it occurs (I think the contingent probability of if VIX goes to 16 it is going to 25 is the right way to look at VIX – the probability of it gapping to 16 in a day is very low – but if it does, it should move a lot higher, quickly).
That ignores the usual stop loss suspects in a crowded, ‘sell vol’ space and assumes no algo will aggressively attack out of the money puts to drive bid/offer wider – which is an integral part of the VIX calculation.
The steady state grind to higher stock prices seems to be the base case, but there are clear paths to some uglier scenarios.
The corollary of my view is that ‘less liquid’ assets will likely outperform in a VIX spike type event because it will happen so quickly and recover almost as quickly, such that extremely illiquid assets won’t have time to trade down (this happened a bit in August 2015 where ETFs like HYG and JNK performed more ‘normally’ than other ETFs because there were no ‘bad prints’ to drive the algos trading). I will highlight some of those abnormalities in this weekend’s report.
Finding ways to buy the cheapest possible lottery tickets to a VIX or volatility spike is the right trade – the event is too unlikely to be the over-riding concern in your portfolio management – but I think it is a valid enough concern and there are cheap enough ways to do it, that it is a worthwhile exercise.
* * *
Full Report below...
Incredible Short Interest In VIX ETFs and ETNs
I cannot think of another product that has such large short interest in BOTH the long and the short versions. I might not be looking hard enough, but on a quick scan, most other asset classes seem to make due with the option of being either long or short.
Given the year to date returns it is easy to see why investors would want to own XIV and SVXY (both up more than 90%) or to be short VXX (down 54%), but why be short both? (for percentage of float short I used short interest and shares outstanding as of June 30th as that was last date I head short interest information).
Shorting both looks even more strange when over the past year it looks like owning both was the right trade (year to date you could have owned XIV and VXX and generated a 40% return?) The strategy of shorting XIV and XVV, with weekly rebalancing has produced positive returns. That is the reason that both are shorted. There is strategy that involves shorting both, rebalancing weekly and harvesting the differences that occur in these two volatile indices. (I had only run daily rebalancing which was flattish on the year, but weekly rebalancing works better).
The Benchmarks Seem The Same
From the UVXY and SVXY prospectus.
exposure to, forward implied equity market volatility as measured by the S&P 500® VIX Short-Term Futures Index (the “Index”)
From the XIV prospectus.
Each series of ETNs tracks the daily performance of either the S&P 500 VIX Short-Term Futures™ Index ER
From the VXX prospectus.
ETNs is linked to the performance of the S&P 500 VIX Short-Term Futures™ Index TR
I will admit that on Bloomberg I couldn’t find a difference between the total return version and the excess return index or the futures index (that could be Bloomberg’s fault, it could be that I missed a subtle part of the naming convention, or it could be that they all reference the same thing – which is probably the starting assumption most people make).
I can’t even find the index calculations (though I only spent half an hour googling various likely sources), but my understanding is that it is weighted return between the first futures contract and the second one. The weighting is Days into Period/Total Days and (Total Days – Days into Period)/Total Days so that at the start of a roll period, the entire weighting is on the front contract and that gradually shifts into the index using the second contract almost in its entirety, so that by the next roll it is all based on the second contract about to become the first contract.
Now For The Good Stuff - Forced Unwinds
If you aren’t confused by this point either I did a better job wrapping my head around the issues than I think I did (which is not likely) or you have just skimmed the pre-amble to get to the good part (seems likely).
But here we are, and we will start with XIV:
Acceleration at Our Option or Upon Acceleration Event
We will have the right to accelerate the ETNs of any series in whole but not in part on any Business Day occurring on or after the Inception Date (an “Optional Acceleration”). In addition, if an Acceleration Event (as defined herein) occurs at any time with respect to any series of the ETNs, we will have the right, and under certain circumstances as described herein the obligation, to accelerate all of the outstanding ETNs of such series (an “Event Acceleration”). In either case, upon acceleration you will receive a cash payment in an amount (the “Accelerated Redemption Amount”) equal to the Closing Indicative Value on the Accelerated Valuation Date. In the case of an Optional Acceleration, the “Accelerated Valuation Date” shall be an Index Business Day specified in our notice of Optional Acceleration, which Index Business Day shall be at least 5 Business Days after the date on which we give you notice of such Optional Acceleration. In the case of an Event Acceleration, the Accelerated Valuation Date shall be the day on which we give notice of such Event Acceleration (or, if such day is not an Index Business Day, the next following Index Business Day). The Accelerated Redemption Amount will be payable on the third Business Day following the Accelerated Valuation Date (such third Business Day the “Acceleration Date”). The Acceleration Date will be postponed if a Market Disruption Event occurs and is continuing on the Accelerated Valuation Date. No interest or additional payment will accrue or be payable as a result of any postponement of the Acceleration Date. See “Specific Terms of the ETNs—Market Disruption Events.” We will give you notice of any acceleration of the ETNs through customary channels used to deliver notices to holders of exchange traded notes.
Acceleration Event
As discussed in more detail under “Specific Terms of the ETNs—Acceleration at Our Option or Upon an Acceleration Event” in this pricing supplement, an Acceleration Event includes any event that adversely affects our ability to hedge or our rights in connection with the ETNs, including, but not limited to, if the Intraday Indicative Value is equal to or less than 20% of the prior day’s Closing Indicative Value
My interpretation of this, is that the Notes (XIV is a note issued by Credit Suisse AG (Nassau Branch)) ‘mature’ or ‘redeem’ (my words) in its entirety if the value declines by 80% in a single day.
I am not a lawyer, so you can have your lawyers read it, but every time I read it, it seems likely that at -80% on day – the fund goes ‘poof’. If the fund winds up down 90% investors would be entitled to the 10% of value left – but that is not the point.
Let’s say we are on a day where all of the calculation is based on the first VIX futures contract and the contract is trading at 8 (why not, we are in a vol seller’s world in the middle of summer).
If that contract goes to 14.4 then I think a rather nasty series of events occurs.
The valuation of the note would be -80% because although VIX going from 8 to 14.4 doesn’t seem that insane, it would be an 80% move in the index.
Now XIV (which as a note doesn’t show its daily holdings, would have to BUY a lot of VIX protection – either in VIX futures forms or in swaps – but end result is the same – a bunch of shorts will be FORCED to cover). Credit Suisse AG (Nassau Branch) as the issue would direct this, because they have no recourse to investors if the fund went down -110%. In that case CS AG (Nassau Branch) would lose 10% - and I suspect they aren’t in the business of giving away free options which is why they have the Acceleration Event defined as such – 20% should give them plenty of time to unwind though in my scenario they get stopped at 14.4 and start losing money at 16 – again seems insane to treat VIX as a percentage change).
All else being equal – UVXY will be a buyer of futures into the close as it would need to buy to maintain its 2:1 leveraged exposure.
For SVXY I could only find this:
Termination Events The Trust, or, as the case may be, a Fund, may be dissolved at any time and for any reason by the Sponsor with written notice to the shareholders.
There is also this little blurb in the prospectus
Inverse positions can also result in the total loss of an investor’s investment. For the Inverse Fund, a single-day or intraday increase in the level of the Fund’s benchmark approaching 100% could result in the total loss or almost total loss of an investor’s investment, even if such Fund’s benchmark subsequently moves lower.
They seem to contemplate some sort of a closeout (‘even if such Fund’s benchmark subsequently moves lower’) but it isn’t very clear.
Here is the current holdings of SVXY (according to their own website which actually doesn’t match Bloomberg’s MHD function)
So SVXY would be short a lot of contracts. It has cash against those. Presumably throughout the course of the day they would be receiving margin calls. While XIV made it very clear what was to occur – SVXY doesn’t make it clear at all. Could they get new money to make more margin calls (are margin calls really only at the end of the day as written in some informational section or can they be during the day? How good are their clearing agreements in terms of avoiding termination?
I am not a lawyer and admittedly gave up searching for Termination, Acceleration, Unwind, Closed, Closed-out, margin, margin call, etc….
This is a different situation than a structured note. In structured note, like XIV, the note issuer would be responsible for any losses beyond money received (not a legal opinion).
In an ETF, like SVXY, it would be counterparties to the company, that are left with the losses. If one counterparty allowed the ETF to lose more money than they held, the counterparty could not go after shareholders for the money (as I understand it). So, in my opinion, the prudent counterparty would have rules in place to prevent that, but it isn’t clear at all in the document as I read it.
Can VIX Futures Move 80% In A Day (And Should We Care)
On a simplistic glance, it seems easier to move from 8 to 14.4, then from 12 to 21.6.
Is there a real potential that that a move like that could trigger buying into a fragile market?
Yes, though maybe it would be offset by redemptions in the long products, or shorts of shorts getting stopped out, because who knows what you do with a short interest in a product that goes away (which would be the case of XIV).
Let’s not forget that VIX itself is a complex calculation – please see the wiki link to read more, if your head isn’t already spinning enough.
Maybe the VIX products are too small and hokey to impact larger markets?
Doesn't Backtesting Protect Us?
Home prices, on a national basis, have never been negative.
The probability of credit loss for a super senior corporate CDS tranche is 0.0001% (or some extremely low number). As far as I know, none of the corporate super senior CDS tranches sold by AIG FP would have experienced a loss due to Credit Events and I think even on the CDX IG index tranches, none have reached into the ‘mezzanine’ tranche, let alone super senior – yet that trade left more blood in the water than an episode of Game of Thrones.
These products, the links between them, the short interest, the unwind rules, have not been thoroughly tested.
On the other hand, SVXY has seen some outflows, so maybe we won’t run into some bizarre scenario.
So Many Unknown Unknowns
With stocks at record highs, and the FANG led NASDAQ once again showing resilience, maybe we don’t have to worry about all these strange products? Maybe all of that is too remote or just simply untradeable – especially when the ‘no one could have seen it coming’ argument would be an obvious one to employ if we ever get a VIX led problem.
Quite frankly, despite writing this, I find it hard to be alarmist right now. I can’t get up the energy to run through town crying ‘the vix are coming, the vix are coming’, but a nominally low level of VIX makes it easier to set something into motion.
While I have been writing this, several people have pointed out the rebound in the S&P 500 Implied correlation index from an unprecedented low yesterday.