A VIX-Replicating ETF Would Be Problematic and Expensive

As this past summer has seen a long-term regime change in the systemic market volatility, discussion of the VIX and other volatility indices has picked up dramatically. Though volatility has been an issue many times in the past, however, one thing has changed since the last period of high volatility - more and more investors expect more and more complicated products to be tradeable in easier and easier ways.

For example, the ETF market has provided a large number of products specializing in commodities, bonds, and even inverses. None of these are assets that are impossible or sometimes even difficult to trade, but in sum, there are hundreds of ETFs and CEFs new in the past 10 years that provide nothing but access to these holdings.

As we discuss volatility, though, there are a number of problems. First, although the VIX is the most understood volatility proxy, its products are by no means the most liquid. Secondly, the VIX is designed as a theoretical price, not an asset - this is advantageous in that it is quoted in units of the underlying asset, but volatility is much more expensive and difficult to hedge than variance.

This is obvious when one compares the liquidity of the variance swap market to any VIX-related derivatives, though somewhat confusing to the an outside investor who is told that they are merely related by squaring or square-rooting.

Having admitted that the variance swap market is very liquid, and that the two are related by such a simple mathematical operation, why then is a volatility ETF still so far away? The problem is likely due to the fact that the reason outside investors want a volatility proxy is diametrically opposed to the safety of the ETF issuer. When ETF investors discuss the VIX and their desire for such a product, they are not thinking of realizing the arbitrage profit between implied and future realized volatility - they're thinking of realizing a hedge against the Poisson jump part of the Merton market model.

The problem is that these larger shocks are nearly impossible to hedge against without high expenses, and given the volatility of volatility in the market presently, such losses would likely sum to a figure that would have the ETF holders even more upset than various oil commodity ETF holders earlier this year. Add to this the fact that nearly all products are OTC markets and the asymmetry of risk for buyer and seller in any risk market and you probably have those insurance and risk analysts putting up roadblocks left and right.

Don't get me wrong, I am all for a volatility proxy ETF myself, and have even seen hits on this site originating from various ETF issuers on the very topic of replicating the VIX, but I'm not too optimistic in the short term for a product with a safe and sustainable structure. While I'm out on the line making predictions though, let me add this last one as a hedge - if we do see a volatility-tracking ETF in the next 2 years, expect it to come with one of the top 5 expense ratios in the market.

Comments

[ETF Wanderer_:]

Interesting article, but a very important "quant" reason is left out - the additivity of variances rather than volatility.

Variance swaps are more popular than volatility products because of the simple reason that they are linear. For example, lets say there are two products with zero correlation; then:
Var(a+b) = Var(a) + Var(b)

but Vol(a+b) < Vol(a)+Vol(b)

The linear nature of variances can help in a lot of trading strategies - e.g. in spreading the index variances with a "weighted" sum of stock variances (essentially an average correlation play).

Volatility - although conceptually just a square root - is non-linear and cannot represented as a "weighted" sum of stock volatility. Similarly, to arb an index vol with stock vol, once has to do a lot of delta-trades for the non-linear parts which is either a pain or transactionally not efficient any more.

2007 Sep 11 11:13 AM

[Michael Bom...:]

Hello etf admirer, thanks for the additions. This is exactly the reason I said that it was much more difficult to hedge volatility than variance from an issuer's standpoint.

The point you make about the non-linear parts is especially valid for the longer term swaps, and it's cumulative approximation error like this that led me to make the comment recalling community's less-than-pleasant response to the predictable contango issues earlier in the year.

2007 Sep 11 11:54 AM

[ETF Wanderer_:]

Michael,

Quite interesting reply.

Regd your statement: "community's less-than-pleasant response to the predictable contango issues earlier in the year":

Can you give some more details on what was the "predictable contago issues" and what was the "community's less-than-pleasant response"? Also what is the "community" that you mention?

2007 Sep 12 10:36 AM

[Michael Bom...:]

The contango issue involved the various west texas crude ETFs and differences in return on spot price and the ETF returns. When I say community, I'm referring to a number of financial papers and sites, this one included. Here are some links to get you started on it if you're not familiar:

* www.google.com/search?...
* www.marketwatch.com/ne...

2007 Sep 12 01:56 PM