When Using the VIX to Time VIX ETPs Works (and When it Doesn’t)

25Sep12

We know that the VIX index itself is fairly predictable.

But we also (should) know that most of that predictability doesn’t carry over to ETPs like VXX or XIV (or the VIX futures on which they’re based).

A simple example…

[logarithmically-scaled, growth of $1, frictionless]

Here I’ve assumed a trader could go long the VIX index the first day of each week (usually Monday), since 03/2004 (close-to-close). I chose to start on that date because that’s when VIX futures launched.

Note the consistent positive returns over the entire period. 59% of days were up, with an average daily return of 1.5% (read why).

Of course, the VIX index is not directly tradable, so below I’ve applied the same Monday-only strategy to the VIX ETP VXX:

[logarithmically-scaled, growth of $1, frictionless]

Ignore the general bearish bias – that’s simply the result of (usually) contangoed VIX futures. What’s important is the average daily excess return (or lack thereof) of just -0.02%.

Put more simply, the Monday observation does not carry over to VXX (or any other directional VIX ETPs).

. . . . .

That’s a simple example of when using the VIX to time VIX ETPs doesn’t work.

I would expand on that observation to say that ALL predictable moves in the VIX that are related to some structural anomaly (like day-of-week, or days till expiration) are useless.

But where I disagree with a lot of folks is that this somehow means that the VIX is never a useful predictor.

Compare, for example, the super simple strategies I showed here (following the VIX futures term-structure) versus here (following the term-structure + timing the VIX).

Note the similar returns each strategy would have produced, but how timing the VIX drastically reduced volatility/drawdown.

Both strategies were silly simple, and there is clearly a ton of room for improvement (ex. protective stops), but I think it illustrates the point I’m trying to make:

The VIX index is predictive when used to time the ebb and flow of volatility (not structural anomalies like day-of-the-week, but non-structural observations like the tendency of the VIX to mean-revert) assuming that we respect the most important driver of VIX ETP returns: the VIX futures term-structure.

The term-structure should always set the tone of the trade, whether it’s biased towards long or short volatility (read why), but once that tone has been set, there’s room to fine tune the trade by timing the VIX itself.

Happy Trading,
ms

Shameless self-promotion: MarketSci’s Volatility ETP strategy is available via subscription. As of today, we’ve produced a 68% annualized return in almost 11 months of trading. All results are real-time and verifiable. Click to learn more.

. . . . .

To stay up to date with what’s happening at the MarketSci Blog, we recommend subscribing to our RSS Feed or Email Feed.