Trading VXX/XIV by Blindly Following the VIX Futures Term-Structure
The consistency of VXX/XIV in recent history has fooled many in to thinking that just staying on the right side of the term-structure (ex. long VXX/XIV when front month futures are above/below second month futures) is a safe path to big returns.
And while I agree that given enough time this approach will probably do well in terms of terminal wealth, it’s guaranteed to suffer ginormous drawdowns along the way.
[logarithmically-scaled, growth of $1, frictionless]
The red line above represents a trader going long VXX (XIV) at today’s close when the front month VIX contract closes above (below) the second month, since 03/2004. The position is held until a change in direction.
For context, I’ve also included buying and holding XIV in grey. Results are frictionless. These ETPs have only actually traded since 2009/10 – see footnote to understand how I took these results all the way back to 2004.
In terms of terminal wealth, this simple strategy would have killed it, to the tune of about 75% annualized, but drawdowns would have been so brutal no sane person would have stayed the course.
To illustrate, a drawdown curve for our simple strategy:
The gory numbers: a peak drawdown of -90%, the longest drawdown lasting over 3 years, and four smaller drawdowns of 30-50%. Ouch. And that’s just since 2004.
Even if we only took the less volatile XIV trades (short vol) and moved to cash when the strategy called for VXX, we would have still suffered a peak drawdown of -80% and done little to reduce those other major drawdowns.
And if you can’t understand why it’s foolish to only consider backtested terminal wealth (especially given the short track record of these products), ignoring volatility and drawdowns, you should probably move along to another blog – one more focused on gambling than sound trading.
So what’s the solution?
I think that we as traders can add to the basic concept presented here to try to reduce volatility and drawdowns while preserving most of the meat in these returns.
Protective stops are a no brainer. Better measures of the VIX futures term-structure, timing the underlying VIX itself, or employing more sophisticated options-based strategies could be helpful.
Of course, this is all made even more difficult than usual by how little data we have to consider, but I think that doing nothing and simply accepting these monster drawdowns as a necessary evil is unacceptable.
Shameless self-promotion: to see the (verifiable) results of MarketSci’s solution to wrangling the volatility beast, visit our strategies.
Geek note: VXX/XIV have only traded since 2009/2010, but we can simulate how they would have performed historically with fairly good accuracy. For data from 12/2005 to when each ETP launched, I use the S&P 500 VIX Short-Term Futures TR Index on which they’re based, and for data prior to 12/2005, I estimate returns based on the underlying VIX futures (read more).
. . . . .
Filed under: Trading Strategies, VIX & Volatility | Leave a Comment