"This Has Only Happened Twice Since 2009" - A New Source Of Concern For VIX ETFs

Last week, ahead of the FOMC, we reported that in the latest striking development involving volatility derivative products, the total outstanding Vega across the entire levered and inverse VIX ETP space had reached $375 million, an all time high. 

This, as Bank of America showed, was the result of the highest positioning in levered long VIX ETPs since July 2016 offset by a record high exposure across short ETPs.

And while the surge in long vol ETFs vega is something that had been duly documented over the past year, the offsetting move by inverse VIX ETPs is certainly a novel development. So much so that Barclays writes today that the recent increase in AUM in Inverse VIX ETPs has raised concerns that investors in these products are likely to amplify fundamentally driven moves in VIX and VIX futures.

As Barclays' derivative strategist Maneesh Deshpande writes, since the mini-selloff in August driven by North Korea related geopolitical concerns, AUM in inverse VIX ETPs, or Exchange Traded Products, has increased significantly. In other words, for whatever reason whether due to Capital, Price, and Roll flow, investors - mostly retail - are now betting not so much on a mean reversion in volatility, but a continuation of the current record low vol trend (a trend which, as we showed earlier today, resulted in the September VIX settlement being the lowest on record).

As  shown in the chart above, the AUM of inverse VIX ETPs is now ~$260 MM vega which is comparable to that in long volatility ETPs at ~$280 MM vega. As a result, the net VIX ETP demand and supply is almost completely in balance, a situation which has happened only twice before since 2009. The only two previous occasions when the AUM of inverse ETPs caught up with their long peers, was after the October 2014 Ebola scare and right after the August 2015 ETFlash crash, when vol sellers emerged after the VIX surge (note VIX was briefly halted for dissemination on August 24, 2015 when the market literally broke). 

According to Barclays, the increase in AUM is partially being driven by the fact that equity markets have so far shrugged off the aggressive escalation in missile tests being conducted by North Korea and increasingly heated rhetoric from both the U.S. and North Korean administrations, which in turn has sent the S&P 500 index to new record highs while the VIX has reverted back to ~10.

More importantly, and as we have cautioned ever since we first noted the dramatic increase in overall vega across the VIX ETP complex, this increased AUM has - according to Barclays - raised concerns that a fundamentally driven selloff in VIX could be exacerbated as the investors in inverse VIX ETPs cover their short volatility positions by redeeming their holdings in these products. And since holders of these products are likely to be retail investors, "the argument is that they are more likely to panic."

Barclays also notes that a more important source of concern is that the inverse products are also Leveraged VIX ETPs. "Since Leveraged VIX ETP managers have to buy (sell) VIX futures when they go up (down), the concern is that they might exacerbate the moves if their trading volume is large."

And with attention likely to focus increasingly more on the VIX ETP sector as a potential source of market instability during upcoming VIX spike, here are some additional observations from Barclays:

Interestingly, we see that over the one-day period there is almost no correlation with the changes in the VIX future for both the Long and Inverse Volatility ETPs. In other words, on the day that the VIX spikes, VIX ETP investors as a group do not significantly change their holdings. We emphasize that this is only true as a group since the secondary volumes in these products increase significantly when VIX spikes and so these ETPs do change hands across investors. Figure 4 plots the gross secondary volumes and the gross Capital Flow for all VIX ETPs and we see that former is nearly 20x that of the latter. It is remarkable that the ETP volume actually crossed $1Bn vega in August when VIX only increased to ~15. Thus, even discounting the fact that a lot of this volume could be driven by intra-day speculators, there appears to be a significant diversity of opinion across VIX ETP investors and thus it would be a mistake to assume that they are a homogenous group.

Looking at trading patterns in the various ETFs, Barclays observes that, as one would expet, when VIX spikes the flow is negative and thus investors monetize their long volatility positions to lock in gains. And yet, there is a major caveat:

Remarkably, even the Inverse VIX ETP investors sell volatility during VIX spikes. In other words, despite the fact that their returns are negative, as a group, these investors have on average added to their short volatility positions. The “as a group” caveat is critical: it is likely that some investors do cut their positions as they face losses but there are enough other investors who are willing to step into their shoes. As highlighted in these figures, this pattern has held even for large VIX spikes such as during August 2015.

In other words, instead of just BTFD in equities, retail investors are now doing that through vol ETFs. This, Barclays continues, can be a problem:

Thus the key risk, as we have highlighted before, is that of a “one-two punch” scenario which unfolds in two phases. In the first phase VIX spikes moderately due to adverse fundamental news and then stabilizes for a few days at an elevated level. The VIX ETP investors follow their play-book of selling volatility expecting a mean reversion. However, if at this point, the fundamentals were to deteriorate further, it would create an air-pocket phenomenon where the subsequent spike in VIX could be amplified.

That "air pocket" scenario is just the waterfall analysis that Morgan Stanley put together back in July when it previewed what could happen to the VIX ETP sector, and the market in general, "If The Vix Goes Bananas."

Summarizing the above (there are more nuanced observations in the full note), the following interesting picture emerges:

  • On the day of the VIX spikes, the leveraged VIX ETP managers buy VIX futures.
  • However, on the first day, the end investors (as a group) do not change their holdings although there is fair amount trading within the group.
  • Over the subsequent few days, the end investors sell volatility (long volatility ETP investors redeem and inverse ETP investors add to their positions).
  • As a result of these creations/redemptions VIX ETP managers are now forced to sell VIX futures.

Obviously the swing in flows causes a fair amount of volatility of VIX futures.

And while the two competing flows have partially compensated each other so far, it is not obvious that same balance will hold in the future:

  • As the AUM of inverse VIX ETPs has reached new all-time highs, as shown in Figure 1, the rebalance flow from Leveraged VIX ETP managers has doubled to ~ $60MM vega for a 10% move up in 1M VIX future (Figure 12) and is also now at all-time highs.
  • Importantly, the Capital Flow from investors in unleveraged Long Volatility investors is also negative but the AUM in this category has remained fairly stagnant.
  • Finally, the Short Interest in Inverse VIX ETPs has dropped sharply over the last few weeks. As we had pointed out recently the hedging activity from the counterparties of the short interest in leveraged ETPs would offset the rebalance flow from the ETP managers. The recent drop means that this compensating flow has decreased.

Barclays troubling conclusion:

Thus on balance the total demand from VIX ETP managers is likely to be higher relative to a few weeks ago and is a significant source of risk especially during a large fundamentally driven shock. The only caveat we would keep in mind is that the volume in VIX futures is likely to cross $1Bn vega during a shock scenario.

And an even more concerning extension: if the VIX Complex is indicative of overall markets trends, and if a similar surge in risk exposure, gamma and vega has taken place among the other two prominent volatility players, risk parity funds and annuities, using the Morgan Stanley short-hand of extrapolating exposure, it would imply that the overall volatility vega could be $10 billion or substantially higher. Such exposure would lead to dramatic moves in the overall market if the next VIX spike or 5 or more vol points does not suddenly reverse and "fill the gap" as has been largely the case on most previously similar episodes.