Shorting VXX Every Week: It Works Until It Doesn't

| About: iPath S&P (VXX)

Last week I published an article on some of the various volatility-related ETNs, such as the iPath VIX Short-Term Futures ETN (VXX), suggesting that buying it and holding over a long time frame isn't really the best hedge for a stock portfolio.

The reason is that over time, the contango in the market for VIX futures contracts eats away at returns.

In that article, I published this chart, using data from the CBOE, showing the front month VIX futures contract and the next four months going back to early 2011:

When the market for VIX futures is in contango (longer-term contracts more expensive than shorter-term contracts), holding VXX is a losing proposition.

Here's a chart showing how VXX performed over the past five quarters along with a chart of the VIX itself.

It's not exactly a stellar performance. Shorting VXX over the long term would seem to be the better move, but when market panic sets in, like last summer, VXX can easily soar, although it usually resumes its downward trajectory.

A weekly VXX short?

ETNs like these are best for short-term trades anyway, so what if you shorted $10,000 in VXX every week? What would your equity curve look like?

This chart shows the cumulative profit and loss for simply shorting $10,000 of VXX at the end of each week and then closing that position at the end of the next week - beginning in mid-2009.

In actual practice, you'd probably just readjust your position each week to maintain a $10,000 short exposure, so I deducted $5 per week in commissions.

Over the long run, there's an expectation of positive returns, but fairly massive drawdowns during times when the contango evaporates.

Adding one rule to filter trades

Instead of looking at the contango itself, it might be easier to take a look at the CBOE Investable Volatility Index. This sort of measures the VIX contango and puts it into index form.

The Index measures the forward implied volatility of the S&P 500 Index for a three-month window centered approximately five months in the future. The Index level is published on Bloomberg under the symbol "VOL Index" and on Reuters under the symbol ".VOL."

Here's a look at the index

The gray line above the index level shows the top of a 30-day Bollinger band at a 1.75 standard deviation level (2.0 standard deviations seemed too high to be meaningful).

So what if you added a new rule to your short VXX trade? That you'd skip the trade if the index level was more than 1.75 standard deviations above the mean?

Well, that makes a difference. Here's how the short VXX trade would have worked if you'd skipped the weeks when the Investable Volatility Index exceeded that 30-day 1.75 standard deviations to the upside:

The blue line shows the original strategy. The green line shows the results of adding the new rule.

Although it may not appear to make a considerable difference, the drawdowns in the spring of 2010 and the late summer of 2011 were a lot less, about 30% instead of the 50% for the taking the trade every week.

I have not convinced myself that this is a practical strategy. Even applying the filter of the rule I defined, you would have only broken even for all of 2011. That's discouraging (but more encouraging than the 15% loss without applying the filtering rule).

However, it's a start - and at the very least reinforces the idea that being long VXX all week every week probably isn't the best idea.

Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.