Description

In summary, I recommend going long XIV at any time that spot VIX reaches 25.  Right now spot VIX is at around 21.5, but the time to research and underwrite this strategy is not when vol is spiking and portfolios are in disarray, so I am posting it now.

But let's review VXX briefly.  As described in the VXX writeup, VXX is an inherently flawed security for long-term investment.  VXX owns a mix of forward months 1 and 2 vol futures, and rolls a portion of the position every day.  The negative "roll yield" is significant (averaging 10% monthly over the past several years) due to the upward "contango" slope of the volatility futures curve.  So, VXX buys high, and sells low, almost every single day.  And since the entire portfolio is turned over every 30 days, this negative roll yield has a devastating effect over the course of a full year.  For example, over the trailing 12 months, VXX is down approximately 80%.

But how to play this?  We could short VXX, which would likely work, at least until it doesn't!  And of course there are the usual drawbacks of shorting.  We could also buy put spreads, but these are actually less appealing than they would at first appear.  The VXX options are priced, in layman's terms, such that you don't "earn" the majority of the spread until the last weeks prior to expiration, and therefore have to take short-term risk of a vol spike.  In other words, the vol of vol seems to increase significantly as expiration approaches, such that even deep in the money put spreads aren't priced to allow for an appealing closeout of the position prior to the last minute (especially taking into account large bid-ask spreads on a percentage basis).  And short-term risk of a vol spike is not a risk I am interested in taking.  So neither of these alternatives are ideal.

Enter XIV (VIX backwards, get it?).  XIV is an exchange-traded note that is the inverse of VXX.  It shorts forward month 1 and 2 vol futures, and on a daily basis rolls them forward.  As a result, XIV typically sells high (month 2), then buys low (month 1), whenever the vol futures market is upward-sloping (as it typically is).  So, if you see VXX as a fundamentally flawed security, then you should see XIV as a fundamentally superior security.

How to play this?  There are no options on XIV (yet).  There likely will be at some point this fall, but I really don't think they are necessary.  One approach would be to simply buy and hold.  This would likely be very profitable over the long term, and is the simplest strategy.  Assuming a typical 10% monthly roll yield, and no overall move in vol levels, you would roughly triple your money over the course of a year.  The other approach, which I recommend, is to opportunistically go long XIV whenever there is a mini vol spike (which I define as spot VIX reaching 25), and then close out whenever spot VIX returns to a more normalized level (maybe 17 or so).  This strategy limits risk by buying in only at more elevated vol levels, and limits the amount of capital utilized, though it likely leaves a lot of profit on the table. 

Relative to shorting VXX, or buying VXX put spreads, I believe that utilizing one or both of these XIV strategies is a far superior strategy.  And since vol will inevitably revert to the mean (unlike almost any other security), I believe that opportunistically going long XIV upon a small vol spike offers a very high likelihood of strong returns.

What are the risks?

1.Vol futures curve could move to backwardation (rather than contango).  In recent years, the futures curve only goes into backwardation briefly, and this only happens when spot VIX spikes (and the future months move up, but not as much and are temporarily lower than spot VIX).  But why is the curve typically in contango shape, and will it remain so?  Most commodities do have contango futures markets, but that may be irrelevant here.  I believe vol is normally in contango due to the desire of long-biased market participants to buy "insurance" in the sense of being long vol, as well as the mechanics of hedgers (either in ETFs such as VXX or in the OTC futures market) rolling forward these hedges to out months (sell current month, buy out month).  This dynamic could change, but if it did we could simply exit the position.  It is important to note that XIV holders can force the issuing bank to redeem the security at a formulaically-calculated value, and therefore "sentiment" regarding the likelihood of contango continuing in the future plays no role.  XIV trades right at the formulaic value, as any differential is immediately arbitraged away.  Interesting dynamic.

2.Credit risk of counter-party.  XIV is actually an exchange-traded note (ETN), rather than an ETF, and is therefore an unsecured obligation of the issuing bank (Credit Suisse AG).  Should it default, the value of XIV would be at risk.

3.Closeout risk on vol spike.  Should vol spike such that the formulaic value of XIV touches zero at any point during a trading day, as measured at 15 second intervals, XIV will terminate at a zero value.  This is designed to protect the issuing bank from a force majeure type event where it would be unable to hedge its risks.  So although XIV resets daily, it faces this test on an intra-day basis.  What type of move in VIX would result in this outcome?  Based on a correlation analysis, XIV has historically had an inverse correlation of around 0.5 relative to spot VIX.  So if spot VIX doubles (up 100%) intraday, I would expect XIV to fall by around 50% all else equal.  In other words, historically, spot VIX is far more volatile than the average of forward months 1 and 2.  So spot VIX SHOULD need to triple intraday (up 200%), to result in the average of forward months 1 and 2 doubling, to result in XIV being worth zero.  This seems unlikely, but is a risk to be aware of, and is one reason I like to buy XIV on vol spikes only (as that greatly lessens the risk of VIX going up another 3x or anywhere close to it).  Note that this security did not exist during the flash crash, but I backtested it and it passed rather easily.

4.Trade gets too crowded.  This is an interesting question.  If this is such a good trade, will it be competed away?  I doubt it but would be interested in others' thoughts on this.  My belief is that the entire vol market is essentially linked, as any differences would be arbitraged away.  So the ETF vol market, and the much larger OTC vol market, and the vol inherent in all S&P 500 stock options, is all one market.  XIV is, in comparison, a flea on an elephant.  Could enough XIV copycats be issued to eliminate contango?  Maybe, but we would certainly see that coming, and again, could simply exit without taking a hit from changed sentiment regarding the likelihood of a contango shaped futures market continuing.    More importantly, funds are free to execute this strategy in the OTC futures market right now, if they so choose.  But they don't, at least not in sufficient size to eliminate contango.  I believe this is due to the large demand to be long vol as a hedge or insurance, and the lack of appeal of being short VIX futures on a regular basis.

Catalyst