Sure, this is probably the gazillionth public service announcement you've seen on inverse/leveraged ETF "tracking error," but it bears repeating at least once more.
You see, a couple of comments on my recent article on the iPath Short Term VIX ETN (VXX), and its 96% drop since 2009, suggested trading an inverse version - the VelocityShares Daily Inverse VIX Short-Term ETN (XIV).
The theory? That if VXX ios doomed to continually fall thanks to the usual contango in volatility, then XIV should continually rise. It delivers the inverse return, right?
Yes, on a daily basis. But consider that over the past three months VXX is down about 19% while XIV is barely up by 2%.
Take a look at this chart. It shows the daily movement in the VXX ETN along the X-axis against the same day's move in the inverse fund, XIV, along the Y-axis.
All looks well. The returns are almost in perfect alignment. When VXX moved down 5%, you could generally count on XIV to move up 5%.
Now take a look at the same chart, but using the price changes across 10 trading days
Notice how that strong linear relationship is starting to sag?
When you extend this to 30 trading days, things get even more, um, flaccid.
What these charts show is that the longer you wait, the less likely it is that an inverse fund like XIV will track its VXX counterpart very well at all.
In fact, during periods when the 30-day change in VXX was near zero, you could pretty much count on XIV to lose you money - or at least not make any. That's not a great long-term investment idea.
If you want all the gory details, Morningstar has a good article on the topic. Meanwhile, please don't expect that XIV will deliver the opposite of what VXX does over more than a day or so.