4. The VIX cannot be bought: why its products decay

Derivatives 7: Volatility as an asset

There is a line repeated in articles, forums and even in sales brochures: buy the VIX when it is low. It is impossible. Not because it is a bad idea, which it may well also be, but because the object of the sentence does not exist as a buyable thing. This module explains why, and why that technical detail wipes out a lot of people's money.

An index is not a portfolio

The S&P 500 is an index, but behind it sit five hundred shares: anyone wanting to replicate it buys them in the corresponding proportion. The index is a label on a portfolio you really can hold.

The VIX is not that. It is the result of a calculation performed every fifteen seconds on the prices of a wide basket of S&P 500 options, weighted so that the result always corresponds to a thirty-day horizon. And there is the catch: that horizon renews itself continuously. Tomorrow, the thirty days start counting from tomorrow, and the options that served the calculation today no longer serve it. To hold the VIX you would have to rebuild that entire basket every day, indefinitely. There is no buy-and-hold position equivalent to the index, because the index redefines itself as it goes.

There is also a feature that makes the problem worse. The VIX cannot be stored, lent or transported. In a commodity, today's price and the future's are tied together by the cost of financing and storing the asset, and if that relationship breaks someone arbitrages it. With the VIX no such leash exists: there is nothing to buy today for delivery three months out. The consequence is that its futures curve can drift a very long way from the spot value without anyone being able to force it back.

What does exist: VIX futures

What trades are VIX futures, and each one is a different contract with its own price. They do not track the spot index and there is no reason why they should: a future expiring in four months expresses where the market thinks that thirty-day calculation will be four months from now, plus whatever it demands for taking on that commitment.

In calm conditions, the curve rises. With the spot index at 15, the first expiry may be at 16, the second at 17 and a half and the third at 18 and a half. You will recognize the mechanism from module 1 of this same level: mean reversion from a quiet situation, plus a premium for selling protection forward. The only thing imposing discipline is that, when its day comes, each future settles against the index calculation and therefore has to converge with it.

Out of convergence comes the decay

Here everything comes together. An exchange-traded product promising volatility exposure with a constant thirty-day maturity cannot hold the index, so it holds a basket of futures and every day sells a slice of the near expiry and buys a slice of the next one. With the curve rising, that means selling cheap and buying dear every single day, and on top of that permanently holding contracts that trade above spot and that are going to converge toward it.

The numbers make it plain. A point and a half of difference between the first expiry and the second, on an average level of some 16 and a half, amounts to around 9% of erosion a month if the curve does not move. Compounded over twelve months, that takes away something like two thirds of the value. I stress the condition: if the curve does not move, that is, if the investor is completely right. It is an instrument that loses money when nothing happens, and that is its usual state. It is the contango of level 1 applied to an asset that does not even exist physically.

When the curve inverts, the erosion turns into a tailwind

In a crisis the curve turns over: spot jumps above every expiry. The roll then switches to selling dear and buying cheap, and these products rocket with spectacular violence. That is why they show up in the news exactly once they have already risen.

The problem is duration. Inversions in the VIX curve are usually counted in days or a few weeks, while the erosion runs every day of the year. Buying and holding one of these products is holding a lottery ticket with a daily subscription, and the subscription is always collected.

What to check before touching one of these products

• What exactly it tracks. Almost none of them track the index: they track a basket of futures with a roll rule written in the prospectus.
• Which expiries it uses and how often it rolls. The short-dated ones decay much faster than the mid-maturity ones.
• Whether it is a fund or a note. A structured note is the issuer's debt: it carries credit risk and usually reserves the right to redeem early, as was seen in 2018.
• Whether it is leveraged or inverse. The daily reset adds a further deterioration that depends on the path traveled, not just on the point of arrival.
• The annual cost and the spread on a bad day. The bid-offer spread on these products widens precisely when someone would want to sell them.

Real case: the index goes and comes back, the product does not

The best known exchange-traded product of this family was launched in January 2009 on short-dated VIX futures. Since then the VIX has done everything: it has dropped below 10 in the quiet years and it closed above 82 on March 16, 2020, its highest recorded level, and today it is back in ordinary territory. The product, by contrast, has lost more than 99% of its value and has had to consolidate its units repeatedly so that its price would not be measured in cents.

That comparison is the whole module in one line: the index has gone and come back many times, and the product that claimed to track it has never come back. In the spring of 2020 it managed to multiply several times over in five weeks, the best run imaginable, and even so it gave back almost all of the gain before the summer was out. Not even the best volatility episode in a decade compensated anyone who had held it from the start.

All of this is education about how these instruments work, not a recommendation to buy or sell them.