Gamed ETFs, from Alan Corwin

October 22, 2010 |

I have been intrigued by recent discussions of ETFs by List members, in particular the UNG and UNL ETFs. Not long before this discussion, there was a discussion indicating that UNG got gamed every time that it needed to roll over the futures that provided the underlying asset for the ETF. Finally, there was a mention by Rocky that UNG had lost 78% of its value since inception while the nominal underlying asset had lost 37%.

Is the relatively greater loss in the value of the ETF a product of this gaming? If not, what alternative theories have been posed?

Other ETFs dealing in commodity futures where there are expenses associated with taking delivery would seem to also be subject to similar manipulation. Have the values of these ETFs experienced similar erosion? Isn't any such commodity ETF bound to erode away given enough time?

Finally, it seemed that today most ETFs were down significantly more than their underlying components. Did any news come out that would appear to be distinctly unfriendly to ETFs in general?

Gary Rogan writes:

This ubiquitous article explains enough about the storage costs and their effect on performance. I came across an additional explanation that the predictable patterns of buying and selling on certain days depress/inflate the prices enough in the wrong direction for the holder to matter as well.

Phil McDonnell comments:

I think it is useful to separate the concepts of 'gamed' and 'carry cost due to contango'. Having contango in the related futures market induces a roll cost every time the fund rolls forward into a new month. That would seem to be an unavoidable structural flaw in many of these funds that will eventually lead to their demise. But the gaming aspect is somewhat different. Specifically I mean that gaming is due to the actions of other market participants who front run the roll periods making it more expensive for the fund to perform its roll operations. That activity simply adds to the roll costs that already exist from contango. 

Michael Cohn asks:

Should I be thinking any differently about the deferred option contracts on these products such as VXX (Barclays Volatility Futures ETF or for that matter the UNG discussed here? How do I think about the changing nature of the basket with respect to these term options that are outside of the existing futures basket for the current composition of the ETF and at their own delivery subject to a new basket? I am convincing myself that I need to learn about basket options influenced by the passage of time.  

Rocky Humbert comments:

The VXX currently has some similar roll phenomenons — however, because it is not a physical commodity, it is not bounded by the same physical supply/demand characteristics of things like natgas, crude, wheat, etc. Rather, volatility is a second-order derivative with no physical delivery — and so the roll can swing wildly and remain in a positive carry condition for very extended periods of time. For example, during the 2008/2009 period, VXX experienced the exact opposite condition — and the rolls were very profitable (because short-term volatility was higher than long-term volatillity expectations).

I want to be clear on an important point: If a speculator is bullish on natgas and believes that prices will rise sharply (in a relatively short time frame), then the UNG is a perfectly reasonable vehicle to express this bet. Natgas periodically doubles and triples in a short period of time. However, if you want a long-term exposure to the nat gas market, then this is a horrible vehicle.

Similarly, having a longterm short of the VXX to pick up the rolls is somewhat analagous to selling far out of the money puts on the S&P. You'll make money most of the time. But you will also occasionally wake up and have a dismal mark-to-market and perhaps give back more than you've ever made. Some may argue that this risk can be managed — but that's as much art as science.





Speak your mind


Resources & Links