[W]ere VIX futures in contango pre banking disaster?
Just to review, contango is a condition in the structure of a futures market in which spot or short-term prices for a commodity are lower than the prices for longer dated contracts. For non-perishable goods, contango is a common state of affairs, since longer-dated contracts will include carrying costs such as warehousing and forgone interest. Contango may also result from expectations that the underlying will rise in price by some future date: if traders expect the price of corn to be higher 6 months from now, but not necessarily by next month, they will be willing to pay more for futures contracts with a longer time horizon.
In this case, we’re wondering whether there are any notable features about the term structure of the VIX futures prior to the financial collapse in 2008. The following two charts track closing prices for the contracts trading during that period, along with the spot VIX for reference. The first chart examines the VIX futures term structure through August and September:
To answer the question at hand, VIX futures were in contango prior to the financial crisis, up until mid-September. But the range was just a few points wide, and as late as September 12th, the whole complex had tightened and was still trading around 24-25. So it’s not as though the January 09 contract was trading at 50 and giving some sort of unheeded warning.
A few other features stand out on this longer term chart. First, notice that as individual futures contracts get closer to expiration, they begin to track spot VIX more closely. This makes perfect sense: a contract trading on predictions of where spot VIX will be a week from now obviously has less time in which to price in some mean reversion than does a contract with a month left, which means traders will price the shorter contract closer to the spot VIX number.
Conversely, notice how slowly the December and January contracts moved even during the worst of the crisis. In late November, as the VIX closed a hair above 80, the December and January contracts called for expiration in the low 40s. Which side of the divide was “correct”? Well, both sides, really: the shorter contracts accurately tracked the day to day spot movements, and the longer contracts accurately priced weekly and monthly movement, including reversion to the mean. The tendency among option premium sellers is to see a high spot VIX and lower futures and think that short term index options are a good sale. In late October, that wasn’t a bad idea; but the trouble is that the very same approach would’ve had you selling premium starting in mid September all the way through the worst of the crisis. Apparently more than a few premium selling funds and CTAs blew up last year for precisely this reason. That’s a failure of risk management more than anything else, though – even the best strategies will fail some of the time.
The chart above also illustrates the point we made earlier about the complexity of the VIX futures ETNs. Anyone holding VIX futures or these new ETNs as a portfolio hedge would have been severely disappointed in late October had they expected a 1:1 correlation with, or even approximation of, the spot VIX – the November contract didn’t even close above 60 in the month of October. And as for the tendency for contracts to track spot more closely near expiration: remember that both of these ETNs will track a weighted average of futures, not just one contract. We don’t mean to sound reactionary, and we certainly aren’t Luddites when it comes to financial innovation, but with products this complex it makes sense to err on the side of caution.