XXV: Dancing the contango—without getting mugged


Updated: Aug 17th, 2014 | Vance Harwood | @6_Figure_Invest

Update

I do not recommend XXV.  At its current price of around $30/share  it does not deliver the sort of performance you would expect an inverse volatility security to deliver.   This security can go no higher than $40 per share and can only get there slowly.  See this post for more information.   If you want to be short volatility use XIV or SVXY.

Original Post

The inverse volatility ETN XXV has gained over 50% in the 70 trading days since Barclays introduced it.   However, far from lighting up the investment community the interest in XXV  has been moderate, with an average daily volume in October of 231 thousand compared to 31 million for VXX—its inverse.

If you’ve been paying attention, you know that VXX ( which not surprisingly has lost over 50% in the same period) suffers from two major maladies:

  1. The VIX index, on which the VXX is based (via volatility futures) tends to drop towards the 10 to 15 range when the market is going sideways or up.  There is no prospect for sustained, long term growth.
    .
  2. The VXX ETN must continually update its portfolio to include more longer dated futures and fewer short term futures.    Longer dated volatility futures are almost always more expensive than short term, so the VXX operators are continually buying dear and selling cheap.   In commodities markets this higher price structure for longer dated contracts is called contango

.
With two strikes against VXX, why aren’t people buying XXV, or just shorting VXX?

If these two ETNs were people, VXX might be a depressed alcoholic artist that has occasional bouts of brilliance,  whereas XXV might be an overperforming sales person that normally blows through sales quotas, but occasionally disappears for a couple of weeks and then calls asking you to post bail.

If you want to invest in XXV you should have a pre-disappearance sell strategy.  XXV already has a doomsday $10 “termination event” stop loss built into it, but with XXV currently above 30 a loss limited to 66% or more is not comforting.

A trailing stop strategy is attractive, with perhaps a 2% trigger, but a key requirement for a stop loss order to work well is an orderly market when things are dropping.   If the XXV price gaps down in bad times then a trailing stop order would not offer the protection it implies—potentially filling at a much lower price than 2% off its high.

Evaluating XXV’s behavior during the May 7th Flash Crash would be a great test—except XXV didn’t start trading until July 19th.

Not easily deterred, I synthesized an approximate XXV chart for all of 2010 using VXX intra-day high data.   I used the intra-day highs, because that would map to the intra-day lows in XXV that would potentially trigger a stop loss order.  I didn’t try to include the impact of XXV’s .89% yearly fee in my calculations.

Comparing the chart of the actual XXV values (purple) against my synthesized chart (green) suggests this approach is valid.

The May Flash Crash and subsequent market correction would have dropped XXV to around 15, but would not have tirggered the $10 doomsday stop loss.    The market uncertainty preceding the Flash Crash would have tripped a XXV position’s 2% trailing stop around April 27th, and the rising $VIX would have kept a prudent investor out of XXV until well after the volatility fireworks died down.   In summary, a trailing stop type approach looks like good way to protect an investment in XXV.

This same analysis should apply for just being short VXX, however my broker rarely has VXX available to short, and with an IRA account you can’t sell equities short, so XXV is the best option for me.

Related Posts



Sunday, August 17th, 2014 | Vance Harwood
  • Jamie Tucker

    I am a student at McGill University, and I am interested in backtesting VXX and XXV over the past 5-10 years for a final project in a class entitled “Market Risk Models”

    I am interested in obtaining the excel sheet you used to calculate your backtested XXV data.

    Let me know if you can help!