Quote from DeeDeeTwo:
You don't understand XIV.
XIV is designed to inverse track VXX *** INTRA-DAY ***...
That means from 9:30 to 16:00.
XIV does not track anything over the long term...
Because you have to rebalance DAILY to VXX....
And you are paying huge rebalancing costs...
Which is similar to paying 10%/month in VXX roll yield.
Both VXX and XIV are designed to go to zero...
The former due to roll yield... the latter due to rebalancing cost.
Going long either is sub-optimal... unless it's intra-day.
http://seekingalpha.com/article/306517-why-i-m-shorting-vxx-rather-than-buying-xiv
Simple test of this: do a comparison of the returns on a graph. Rather revealing. XIV is a better-designed fund than VXX, which is a horrible design. VXX has had to do at least one reverse split as I recall, whereas XIV has a normal split in its price history, another very revealing fact.
As I wrote, XIV has a positive long-term graph, and the reason is in its profile: its supposed to replicate the opposite of the SP VIX Futures Index. That would do an excellent job of exploiting contango. Contango is the shape the curve has the majority of the time.
For the VIX, you have to sit and think about this logically. Volatility is really just a measure of uncertainty in options prices, the same way that interest rates are a measure of uncertainty in the bond world. Thus, in bonds, the farther out in time you go, the more you get paid in interest for holding a bond, normally. In options, the farther out in time you go, the higher the IV in the price of the option, normally.
Sometimes this inverts, but this isn't normal, either in options or bonds.
In VIX futures, the effect is magnified, because normally the futures are higher than the volatility of the expiration of SPX options that underlie the futures contract. As of right now, the August futures contract of the VIX, which has the Sept SPX expiration as its underlying, is a few percentage points higher than the volatility of that expiration, with only a few days left on the contract.
So, in normal times when the curve is in contango, as you get closer to expiration, volatility is gradually declining, and on top of that the futures contract is converging to that declining volatility. (although I wouldn't try anything with that August contract, since so close to expiration the probabilities aren't really in your favor anymore. As with any set of multiple probabilities, it's better to have more tosses of the coin to allow things to be in your favor. And you really really don't want to mess with the SOQ that settles the contract.)
If you compare VXV, three month volatility, to VIX, over the history that they share, you'll see that on average VXV is 7% higher than the VIX. The median, even more revealingly, is 8%. This shows you that contango is the shape of the curve the majority of the time. If backwardation were normal the median would be negative, and if it were 50/50 it would be very close to zero.
Therefore over the long term, if you're set up to exploit contango, which is how XIV is designed, you'll do better than if you're set up to exploit backwardation, which is what VXX is set up to do. But in the times when the curve is in backwardation, VXX will do better, obviously.
