VXX vs. XIV: A $100,000 Experiment

We all know that iPath S&P 500 VIX Short-Term Futures ETN (VXX) declines over time. It's down 99.6% since inception just over six years ago, so I hope we all know that.

We also know that we can go short VXX via either: going short VXX (brilliant!) or going long VelocityShares Daily Inverse VIX Short Term ETN (XIV). Vance Harwood has a good post up on the merits of shorting versus buying. In theory, the long can do better than the short over time, in that the short can only go to zero and the long can rise over 100%.

Another way to look at the comparison is that a short is a de facto negative-gamma play. If you short $100,000 worth of "Whatever" and it goes down 5%, you're now short $95,000 worth. If instead you went long the inverse, you're now long $105,000 worth. But since it's an inverse, you're actually short $105,000 of the underlying.

Of course, as came up in the discussion, you can simply roll either the short or the long inverse. That is, you can adjust at the end of each day so that you are now the equivalent of short $100,000 of "Whatever."

A static (non-rolled) XIV position can outperform a static short VXX position over time. But it probably won't thanks to the laws of compounding. VXX would have to move pretty much in one direction for that to happen. But a rolled XIV long can keep up with a rolled VXX long, right?

Well, I thought so. In fact, I thought it would come out almost identical, and would outperform in an extreme directional move. Essentially, you would run out of capital later if it went way against you.

So I ran some data, and it didn't come out quite as I expected. I took a hypothetical $100,000 portfolio and compared calendar-year returns for Straight VXX Shorts (held one year), Rolled VXX shorts (rolled daily such that you came in each day short $100,000), Straight XIV Longs, and Rolled XIV Longs. XIV first listed on Nov. 30, 2010, so I started with 2011. The worst choice each year is in red, the best choice is in green.

Remember that we started with $100,000, so a $90K loss -- like we took daily rolling an XIV long in 2011 -- is pretty awful. Anyways, I found the results somewhat surprising. I'm not sure why rolling a VXX short outperforms rolling an XIV long, much less this soundly and consistently. My only guess is that XIV pays out a distribution, but I can't find one.

Shorting VXX in some fashion clearly trumps owning XIV. It's not clear that rolling daily to keep the $100,000 exposure adds any value. It works in some stretches, not in others. You have to roll at some point lest you have no position left over time, but it's unclear exactly the criteria (time? price?). We can back-fit a "system," but I'm not sure that would work going forward.

Disclaimer: Mr. Warner's opinions expressed above do not necessarily represent the views of Schaeffer's Investment Research.