Vix wagging | FT Alphaville

Vix wagging

What do dividend futures have in common with the Vix, and Vix futures?

Our hypothesis: both have seen excessive demand from structured products desks skew price discovery in their markets. And in the case of the Vix, there’s been a breakdown in its function due to the effect of the so-called Bernanke-put too.

But whilst the dividend futures market has been pretty up front about the scale of the structured product footprint in their market, those operating in Vix futures still seem to be in denial.

Now before ye critics out there shout — “oh, but the two are not comparable markets. Dividend futures are tiny in comparison and have seasonal factors affecting them, making the market much more disjointed”. We know this. It doesn’t in our opinion rule out the argument that there is more of a structured product footprint on Vix futures than most people realise.

Here’s the case laid out point by point anyway.

First, there’s the issue of what the Vix is.

As one self-described ‘keen’ market observer explained to us, it’s not just a question of the market’s volatility expectation (as implied by S&P 500 options) that has to be factored in. The Vix is also very much a function of the funding market:

The Vix being implied volatility (through options) it is in effect describing how confident option writers are. Since they tend to be confident in easy money context, the Vix effectively trades down in such periods and more or less mirrors the cost of money. To this extent it is not a great predictor except when it has reaching a floor (say 10-12): this is the calm before the storm.

Second, there’s the question of who is using the Vix?

The same source says:

But increasingly, its meaning has been declining as equity markets turn in circle i.e. 90% of the volumes are meaningless daily micro arbitrages between sell side operations. There are very few orders stemming from genuine buy and hold players. So that prices are fragile, Vix based signals are fragile and extra liquidity on a fragile base is a recipe for disaster.

S&P options volumes may dwarf that of Vix futures, it is true, but if you consider that the bulk of the trades could now be tied to Vix futures arbitrage… then the tail wagging the dog argument does become reasonable. 

Third, as we’ve explained here, there’s growing evidence that investors are using longer dated options to fund the hedging of their tail-end risk, prices of which are not incorporated into the Vix index calculation itself, which covers only front and second month options. Vix is only an indicator of near-term implied volatility.

Fourth, the Vix is dis-correlating versus the cost of insurance generally, for example that implied by CDS prices.

BNP Paribas makes the point on Monday that whereas the Vix used to correlate with five-year CDS on senior financial names for most of 2010, it’s recently broken out of that pattern:

So why is this happening?

Fifth, what’s going on with volatility arbitrage? As we’ve mentioned here, volatility arbitrage strategies reacted very strangely to the Japan crisis. It’s baffled us, and everyone else we’ve spoken to about it.

When realised volatility shoots over implied volatility, as it usually does during a black-swan type crisis, a volatility arbitrage strategy should theoretically generate a loss.  This time it did the opposite:

Sixth, convergence arbitrage is not a reality in Vix. As Theo Casey from Futures and Options World discusses here, the spot VIX itself cannot be realistically replicated because the execution/bid-offer costs are too high.

That makes the convergence arbitrage opportunity pretty limited, meaning any supply or demand imbalances in pricing are harder to weed out.

It’s also why something like the S&P volatility arbitrage index may be very sound in theory, but when it comes to execution… it’s challenging to replicate.

Seventh, contango dominates Vix futures for reasons other than pure arbitrage, and probably has very little with the market’s assessment of volatility in x amount of months time.

In fact, as Javier Mencia and Enrique Sentana write in a recent paper entitled “Valuation of Vix Derivatives” there doesn’t seem to be any definitive way of valuing a Vix future effectively at all.

As they note:

Importantly, since the VIX index is a risk neutral volatility forecast, not a directly traded asset, there is no cost of carry relationship between the price of the futures and the VIX (see Grunbichler and Longsta , 1996, for more details). There is no convenience yield either, as in the case of futures on commodities. Therefore, absent any other market information, VIX derivatives must be priced according to some model for the risk neutral evolution of the VIX. This situation is similar, but not identical, to term structure models.

What’s really interesting though is when they try to apply the two existing and most logical mean-reversion models out there for evaluating the relationship between real and risk neutral measures, they find they don’t fit historical performance at all:

However, both models yield large price distortions during the crisis. In addition, they do not seem to capture either the level or the slope of the term structure of futures and option prices, or indeed the volatility skew.

If anything there is significant evidence of mispricing of Vix futures in the past, based on what these two models (the most logical for valuing Vix futures out there, in the authors’ opinion) indicate.

A similar conclusion, by the way, was also reached by Zhiguang Wang and Robert T. Daigler with regards to Vix options pricing. As they noted:

The conclusion from the results in this paper is that an adequate pricing model has yet to be developed or that the market is substantially mispricing VIX options.

Of course, all of the model mispricings examined here can be caused either by imperfect models or by inefficient market prices.

In conclusion.

If academics are unable to find reasonable Vix futures valuation models, how exactly can banks (or anyone else for that matter) tell what the fair value for a Vix future actually is?  And, we might add, how can they tell when these futures are responding to pure ETN structured product flows rather than anything else?

No defined valuation model further implies multiple imperfect bespoke house models are currently being utilised to value these contracts.

Now that, if anything, creates a perfect petri dish for reflexivity to breed in.