The Demons Were Real: Monday’s VIX Carnage And The 13-Sigma Event

The Demons Were Real: Monday’s VIX Carnage And The 13-Sigma Event

And so, legions of retail investors, hordes of previously cock-sure newly-minted “money managers”, and scores of popular pundits who swore to you they knew what they were talking about, are left to ponder the stark reality of a market structure-driven nightmare scenario.

Turns out, the whole VIX ETP rebalance risk was a real thing (who knew?!) and not only that, the knock-on effect for the systematic/programmatic crowd was real too. Go figure. Who could have predicted that? I mean, besides every goddamn derivatives strategist on Wall Street.

So this morning, as all of the poor fuckers who used to spend their daily commute gleefully tweeting out pictures of the S&P tracking earnings growth spend this morning’s commute staring blankly at the person sitting across from them on the metro, Wall Street is busy explaining what went wrong.

Earlier today we brought you Goldman’s take along with some color from the usual suspects at JPM and RBC and we thought, given the rather high level of interest, we’d expand on that via a couple more excerpts from a separate Goldman note and also some color from Barclays.

As Goldman’s Charles Himmelberg writes, the chaos that unfolded late Monday in the U.S. was clearly the product of a domino-tipping event. That is, while there were plenty of reasons for people to be nervous, none of those reasons suffice to explain the ferocity of the move in the VIX.

Remember last week how we kept pounding the table on the vol. moves being noticeably larger than what the drop in the underlying seemed to imply (one example here)? Yeah well as Himmelberg notes, what we saw in the VIX was “orders of magnitude larger than that of other asset classes [and] in particular, the move was roughly 5x larger than would be expected vs a beta to the 4.2% move in the S&P.” Have a look at the visual on that:


Goldman goes on to note that while the Wells Fargo drama, some earnings “disappointments” and the handover of the keys to the universe from Yellen to Powell might have been sufficient to explain a rough day (especially in light of the bond selloff and Friday’s action), “none of these narratives are adequate to explain the size of yesterday’s moves.” This was a 13-standard deviation event:


Now you might be asking yourself this: “how the fuck is that even possible?” Well, we’ll tell you. It’s possible precisely because a lot of people didn’t understand how dangerous the evolution in market structure truly was despite the best efforts of the people on Wall Street to warn them.

“We have long been nervous that [short vol.] strategies could be more negatively convex than the recent benign market conditions would suggest – owing to new and as yet untested changes in the regulatory framework for market making, liquidity provision could potentially prove unexpectedly scarce in a down market,” Goldman goes on to write, adding that “our vol strategists even forewarned that the biggest threat was a one-day, end-of-day vol spike, which would push issuers to rebalance positions quickly to avoid unhedged overnight risk or excessive tracking error.”

Right. And see they are hardly the only bank to have warned about that. JPMorgan has been pounding the table on it, so has Deutsche Bank, so has Morgan Stanley, etc. etc. Oh, and Barclays, whose analysts wrote a veritable manifesto about the risks inherent in this setup last year. On Tuesday, they’re out with a new piece, and it’s the same story. Here are the details…

Via Barclays

Spike in VIX index and futures driven by technical factors

The VIX index and VIX futures spiked significantly with the VIX closing at 37.32 (Figure 1). Even VIX futures rose significantly with the Feb VIX future closing at 33.2.


The initial increase in VIX products was driven by the continued selloff in equities. However, Figure 2 shows that the bulk of the move in both VIX and UX1 (the frontmonth VIX future index) occurred towards the end of the day. The bid-offer spread for the SPX options used to calculate the VIX index widened substantially in late afternoon and was driven by a lack of offers rather than substantial increase in the bid price. On the other hand VIX futures only increased significantly after the cash-close at 4 PM EST


Indeed prior to late afternoon, the VIX futures appear to have been underperforming the move in the SPX index. Figure 3 quantifies this by plotting the expected level of the intraday value of SPVXSP based on a regression of five minute intra-day returns of SPVXSP versus S&P 500 futures over the past one month.


In our opinion, the major driver for the move post the cash close was likely driven by the demand from managers of leveraged VIX Exchange Traded Product (LETPs). As we have discussed in detail in the past, these products target 2x or inverse of the daily return of SPVXSP. In order to maintain this leverage, the managers of these products need to buy (sell) VIX futures as their price increases (decreases). The precise demand/supply depends on their NAV at the close and hence their tracking error is minimized if the trade is close to this pricing point. Note that the NAV is determined by the futures’ close at 4:15 pm. Based on level of SPVXSP just prior to cash close at 4 pm, we estimate that the total demand from this channel was ~$120M vega or 120,000 VIX futures contracts. It appears that the lack of liquidity in the final 15 minutes caused a significant increase in futures prices. Indeed over this 15 minute period the total volume traded in the two front VIX futures was only ~80,000 contracts. Since the demand increases as the expected close price goes up, this likely led to further buying pressure, causing the VIX futures prices to increase further. In the end, the Feb futures increased by a remarkable 10 volatility points in the last 15 minutes of trading.

Like Goldman and like JPMorgan’s Kolanovic, Barclays is feeling a little bit better about the situation now that the rebalance risk has been flushed out, but as we cautioned this morning, the question now is what happens to the systematic/programmatic crowd if volatility manages to sustain a spike.

Whatever the case, we would gently suggest that maybe next time all of the people who seem to know what they’re talking about are telling you that something is a problem, you at least consider whether they’re right rather than simply doubling down on your XIV bets at the first 1% move lower on SPX.

As for what happens next with CTAs and vol. control funds and pretty much anything else where vol. is an input, we would just say (again) that the demons are real…

Image result for evil dead we're gonna get you gif

8 thoughts on “The Demons Were Real: Monday’s VIX Carnage And The 13-Sigma Event

  1. Can we agree that the use of sigma in the context of a standard-normal Gaussian distribution – basically how it is universally used on Wall Street – is complete and utter garbage and only useful for Joe Schmoe manager of a widget factory and MBA students being drilled for their first stats test? Explaining something as a 13-sigma event (and this is not specific to your article H, as I understand you’re using the parlance of “the street”), something with such a minuscule likelihood that it could reasonably be expected not to have occurred once in the history of this universe, in the context of a completely expected occurrence (to anyone who was paying attention) is patently absurd.

    You can pin wings to the back of a pig and call the pig a bird, it doesn’t mean that pig is a bird and you’ve gained any insight into the workings of the pig because you’ve previously observed the habits of birds.

    1. Ha ha! The guy is telling you that, yes indeed the tail can and will wag the dog with these leveraged derivative products, and you wanna pick nits over *how* it is described rather than the utter mind-blowing significance of what it is describing.

      1. roberto… IMHO…
        It wasn’t mind-blowing; I expected it w/o knowing when.
        It was built in. The significance to me is:

        1. The particular “vector” taken to get there.
        2. The actual teaching event, again – inevitable, that Wall St. will build pos-feedback mechanisms that blow up (by definition), given the means & motivation. It won’t stop.
        3. That the same are so monumentally ill-suited and incompetent by license and ignorance – and greed. That won’t stop, either.

        Journeyman is correct. Part of the hubris & incompetence is applying the blather of terms like “sigma”, pseudo-intellectually, part of the fallacy of option pricing to begin with.

    2. Sigma is just sigma. It has a formula, and implies nothing about the underlying distribution. That a 13 sigma event occurred, implies one of the tails of the underlying distribution is morbidly obese. Thus, pointing out that something is a 13 sigma event is very useful.

  2. Sometime last year, you mentioned how such an event as occurred yesterday could trigger a cascade of algorithmic selling, multiplying a 5% market drop into a 20-30% drop. Things seem a bit quiet today. Do you have any update on that thesis?

    Asides, I just want to say how much I appreciate your blog. I am by no means a finance professional and I consider myself fortunate to be able to learn from you.

Leave a Reply to Jamal James Cancel reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

NEWSROOM crewneck & prints