As usual, the VIX term structure is in contango. I don't mean the obvious positive slope; I mean that each future is overpriced compared to its statistical expected value. I mentioned this in a prior article of mine, True Contango: The Real Reason to Short VXX, and in this article I want to bring some evidence to the table.
Although this "expected value" might seem like nonsense to some, I believe it can be estimated like this.
The VIX can be fitted very well to the following model:
I used historical experience since 12/31/2009 and slopes seen in the current term structure for calibration purposes. This time frame seemed to provide a fair but conservative (from a short volatility standpoint) portrait of current VIX behavior.
The model performs 10,000 Monte Carlo simulations of 252 appropriately scaled incremental changes from the current spot rate to expiry for each month's future and adjusts for simulation bias.
Here is a graph of the error in the survival distribution of the average of 10,000 calibration simulations:
As you can see, the model simulates values according to a distribution within plus or minus 8% of the historical distribution of the VIX. The error that does exist is used to develop bias correction factors which are later applied to simulated values for futures and options for greater accuracy.
The model also matches historical statistical measures fairly well:
All of these measures match the historical experience within 12% except for the min, which still matches petty closely within 23%.
Therefore I believe it mimics the behavior of the VIX pretty well.
With it, I believe it is possible to estimate the "fair" value of each month's future by taking the average of all the simulated expiration values for each month.
In doing 10,000 simulations (per month), the following term structure arises (please note the market prices are as of early Thursday morning):
Here is a 99.5% normal confidence interval for the implied expected returns of each month:
If this model is credible, the market is still clearly in contango. Granted it is probably the case that the current behavior of the VIX is different than the behavior from 4 or 5 years ago (which was relatively elevated). If so, the model is positively biased, and the fact that the estimated "fair values" still undershoot the market suggests the term structure is in serious contango.
So as usual, shorting volatility seems to be the way to go. There will undoubtedly be plenty of noise in the short run, but so long as futures prices remain in contango, a long-term long position in a short-vol ETP like SVXY or XIV or a short position in a long-vol ETP like VXX or TVIX should continue to profit off this statistical arbitrage.
Disclosure: The author has no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
The author wrote this article themselves, and it expresses their own opinions. The author is not receiving compensation for it (other than from Seeking Alpha). The author has no business relationship with any company whose stock is mentioned in this article.
Additional disclosure: I hold a number of positions and options on the VXX, TVIX and SVXY which profit when VIX futures decrease in value.