"Do I Have To Worry About Another Volatility Spike"- A Q&A With Goldman's Head Quant

After a day of conference calls with investors, Goldman's newly-hired quant Rocky Fishman has assembled the most frequently asked questions that are relevant for trading dynamics in the VIX and equities in the coming days.

Here is the result: Goldman's Q&A on the Trading Dynamics of ETPs

* * *

1. Do I have to worry about VIX ETPs driving another similar volatility spike?

Not for now, and not likely anytime soon. The large short VIX products were the primary driver of Monday afternoon's late-day acceleration in the VIX, and with them diminished ($300mm AUM now vs $3.9bln peak AUM), the quantity of VIX futures ETP issuers would have to buy on a further volatility spike is diminished as well. Levered long ETPs remain sizable ($1.1bln AUM), but per unit of VIX futures exposure, they trade less on a volatility spike than inverse products would (filling the gap between long futures that have risen and a fund NAV that has risen faster requires fewer units of futures than between a NAV and futures position that are moving in opposite directions). The higher level of VIX futures also makes an N-point move a lower percentage move than it would be with lower VIX futures prices.

2. What’s the impact of the XIV (and Japan-listed 2049) redemption? What’s the impact of the SVXY continuing to trade?

The previously sizable short VIX ETPs no longer have a significant impact on the VIX futures market because Monday's sell-off pushed their AUM down so much that their VIX futures holdings would be immaterial to the broad market (the SVXY is short less than 1% of VIX futures open interest), regardless of whether or not they continued trading.


3. Why were VIX futures and S&P futures both down throughout much of Tuesday's trading day?

Typically, VIX futures rise when S&P 500 futures are falling; however, VIX futures had such outsized, technicals-driven buying pressure on Monday afternoon that the absence of this was enough to allow VIX futures to fall on Tuesday, even in the context of an initially weaker equity market.

4. Were short ETPs a material percentage of the VIX futures market? What implications do this week’s events have on VIX future trading volumes?


VIX future volumes closely track VIX ETP trading volumes, so we would expect a material drop in ETP volumes due to the lack of sizable short ETP trading activity to result in a similar drop in VIX future volumes. We estimate short VIX ETPs constituted over 40% of ETP trading volume in January, when measured in VIX future-equivalent terms. While VIX derivatives naturally are active when volatility is moving around, we expect VIX future volumes to fall below - perhaps substantially below - their typical 2017 range (1st & 2nd futures volume was 30% higher than the typical 2015-6 volume) as conditions normalize.


5. Is there a transmission mechanism by which VIX futures activity can lead to S&P 500 trading?

Investors may use S&P 500 futures to hedge some of the risk of short VIX future positions. If some of the investors selling VIX futures to ETP issuers on Monday simultaneously sold SPX futures as a hedge against the market risk inherent in their short VIX future positions, their positioning could potentially have impacted equity prices. High volatility can also pressure equity prices via flows from systematic volatility-linked funds, investor confidence, and VAR-driven risk reduction.


6. Was this technical selloff expected? Should vol-of-vol (i.e., VIX option prices) be structurally higher given what has taken place?
VIX option prices had been abnormally high for the current low level of volatility throughout much of January as investors had priced in some risk of outsized short VIX ETPs causing turmoil like Monday's. In effect, option pricing implied that should volatility rise, it would rise quickly. With the short VIX ETP market now small, we think that implied volatility of VIX options should reset to levels lower than the last few months once conditions normalize, which could create opportunities to sell VIX options.

7. If the AUM of short VIX ETPs was only $3.9bn at their peak, why are they important?

Short VIX ETPs were important because a hypothetical spike in VIX futures would economically drive their issuers to buy an outsized amount of VIX futures, and that buying could push VIX futures up more, creating escalating volatility like we saw on Monday. VIX ETPs are small relative to the US equity market, but large relative to the VIX futures market (often accounting for 40% of open interest). These ETPs were also highly correlated with the equity markets and had multiples of the SPX's volatility, giving them a larger impact on portfolios than their AUM would suggest.

8. Why do short and levered long VIX ETPs both buy more VIX futures when volatility spikes?

When volatility rises, both inverse and levered long VIX ETP issuers are economically driven to buy VIX futures: the inverse product issuers do so to reduce a short position that has become too large relative to their AUM, and the levered issuers do so to supplement a long position that has not risen as quickly as the AUM of the ETP itself.

9. Can VIX ETPs cause vol to fall abnormally quickly just like they caused this rise?

Yes and no, in our view. VIX futures tend to spike up more than they spike down, but should volatility drop dramatically, ETP rebalancing could help VIX futures overreact to the downside as well. We estimate that Tuesday's normalization of volatility left ETP issuers with around 40k VIX futures to sell - a large amount but far below Monday's estimated "vega to buy" that exceeded 200k VIX futures.

10. Why do issuers have to trade near the 4:15 futures market close?

The indices behind the VIX ETPs are based on one-day changes in VIX futures levels, measured by their 4:15 PM NY time prices. Every day, an ETP’s NAV change is a weighted average of the one-day returns of two VIX futures, but those weights change every day. It is only at the close of each trading day that the next day’s weights are fully known, because the total dollar amount of futures involved needs to be exactly the right leverage times the price of the product. This process becomes a feedback loop because each ETP’s closing NAV is an input to the size of its position the next trading day. As we approach the close every day, an ETP issuer shifts its portfolio to the next day’s position so it can correctly replicate the next day’s return.

11. Did the XIV and SVXY cause 2017's surprisingly low VIX range? Now that they are diminished will volatility return?

In our view, 2017's low VIX range was driven primarily by low S&P 500 realized volatility (the lowest S&P 500 realized volatility since 1964), which was a byproduct of low volatility across asset classes, low intra-SPX correlation, and benign economic conditions. We do not see evidence that VIX ETPs furthered this in any material way.

12. What is vega?

Vega is a measurement of volatility risk that represents dollars per volatility point. Because the contract has a 1,000 multiplier, each VIX futures contract carries 1,000 vega - i.e. its value changes by $1,000 for each point that the VIX futures price moves. The amount of vega in any one VIX ETP share changes over time, as the amount of VIX futures exposure that can "fit" inside a share is a function of the share's NAV and the price of VIX futures. Short VIX ETPs had seen their vega per share grow quickly in late 2017, but now the SVXY has very little vega per share.


Yen Cross Wed, 02/07/2018 - 22:39 Permalink

 I've been looking at equity future spreads over the last several days, and they're compelling.

 Now lets discuss the delta on eur/usd [puts and calls ] going out to March first.

 If you look at the options lines the price steepens significantly.

 I'm saying this, because I tried to by the dip in eur/usd earlier today, after making 106 pips on a rather large usd/chf trade, [.9298] and looking at many technical levels.

  I'm just not feeling good about being long eur/usd. So I'm probably going to cut it, and switch to usd/jpy, as a proxy. The vol in usd/jpy is much lower. Easier to trade the machines, on the crosses. Maybe gbp/jpy?

Dewey Cheatum … Wed, 02/07/2018 - 22:41 Permalink

What a load of bullshit tripe. these so called quants ain't got a fuckin clue. NEWSFLASH DUMBASS it's a house of highly leveraged synthetic bullshit, that can't take the pressure, when their hand gets called..in the liquidity trap, known as a rogue VIX

Ha! bring it on.

Nelbev Thu, 02/08/2018 - 00:24 Permalink

This is another dumb statement from Goldman's head quant "VIX option prices had been abnormally high for the current low level of volatility."

This is volatility expected over volatility and it was realized head on this last week.  This is what I hate about these math heads in finance, they look at asset prices from past and discern a distribution, mean, variance, skewness, correlation matrices, or MBS payoff rates when i rates fall and people refinace.  It is all based on historical data.  We have a model like Black Scholes or CAPM they learned in grad school or secret formula and plug in a computer.  (Lots of the data is bad too when you have last price of an option which traded last an hour before close used against underlying asset price close instead of last mean bid/ask.)  But moreover the expectations change.  E.g., before an earnings release, time premiums and options prices are high, after the release and the stock move in reaction, the time premiums shrink, any off the cuff options trader knows this.  Basically you cannot say the distribution or tails are fixed based on prior data since the options prices represent the true distribution going forward.  The January options over the long and reverse volatility ETFs were not abnormally high or mispriced, they were only high relative to prior prices, no anomaly, the distribution changes unlike rolling dice.

And these hords of smart quant people with Ph.D.s got the pricing for MBS wrong last crisis by basing charge-off rates and defaults based on just a few decades of prior data looking at basis point spreads, yea, bank does not take a loss because of equity in event of foreclosure, some of the mortgage had been paid off plus down payment, yea, no or limited risk based on prior data and leverage up to get my bonus, but the prior years of data the quant models were based on had no significant fall in housing prices wiping out equity, and in turn the GSEs and banks and taxpayers fleesed bailing out the idiots, plus the related mess we are in now if looking at fed's balance sheet or public debt.  You could smell the VIX when at record lows, that it would not last, but maybe not from some "advanced" engineered finance math model where you follow a trend line based on prior data.  Options prices are not all based on current or past volatility, but more future expected volatility.  They inflated quick last week, unfortunately I wrote calls against my long UVXY position as the fat premiums were too tempting, but that is not abnormal when traders think the probability distribution moves to upper tail and realization says they were right.


Hey, nobody posted a response to my prior question in another post today:

"Can someone tell me why now today's trading near close UVXY is UP 17.56+2.34 (+15.37%) while VIX is DOWN 23.21-6.77 (-22.58%) (and SVXY also up a tad, 12.64+0.40 (+3.27%), but right direction)?

I am confused, did the Russians hack my quotes? Should I call Perkins Coie and Crowdstrike?"

Anyone care to take a hack at explaining it?  Satan with your -666 signal last week, are you there and can answer?, otherwise does someone here who did not just burn up shorting volatility have a reasonable explanation? 

mily Nelbev Thu, 02/08/2018 - 05:07 Permalink

To answer your question,

Short Term ETPs VXX/UVXY/TVIX/SVXY/XIV do not track VIX spot they track S&P VIX Short Term Index which consists of weighted first and second month VIX futures, as of yesterday close VXX held: 

CBOE VIX Future MAR 18 72.54%
CBOE VIX Future FEB 18 27.46%


CBOE VIX FUTURE 03/21/2018 (UXH8)    77.15%
CBOE VIX FUTURE 02/14/2018 (UXG8)    22.85%

So what you are buying or selling via Short Term ETPs is (short term) future expectation where VIX spot is going to be. That's why even VIX spot is down, VIX futures may be up as the market expects more volatility in the near term, see VIX futures curve / term structure at www.vixcentral.com. We are now in backwardation so long VIX Fut ETPs like VXX/UVXY will keep gaining. Once (or if?) things calm down curve will be back to contango so long ETPs price will start degrading.

And thats is also why XIV blew up as all VIX short term ETNs had to rebalance (readjust weight and quantity or in case short XIV/SVXY cover their shorts) on 1M/2M futures after close based on daily move in VIX FUTURES triggering feedback loop causing 1M future to double between 16:00 and 16:15

By the way, in theory TVIX/UVXY could also disintegrate or go down 90% on a single day, when 1M future cuts in half on a single day (simplified as again it is weight of 1M/2M and double leverage), not sure that is even possible but we have witnessed so many things that should have never happened - I would not be surprised, so options writers be careful!

In reply to by Nelbev