VIX index and futures performance during extreme events

Comments (on the site Seeking Alpha) on my latest VIX related article, “How to profit from VIX ETNs and futures contracts” lead me to think it would be interesting to see how the VIX index behaved immediately preceding and during the 1987 market crash.   One thing to keep in mind is that the method of calculating VIX changed in 2004, so this historical information might not be completely comparable to today.  Never the less, here is a chart of the VIX index leading up to the 1987 Crash:

What we see is that going into October 19th, the VIX index had ramped up 70% within 10 days, with a good portion of that climb occurring in the five days before “black Monday”.   In other words, the week leading up to the big crash was itself a fairly extreme market event, that by usual metrics sent the VIX index soaring.  Though VIX futures were not trading until 2004, based on their behavior since that time we know that when the VIX index Jumps anything like 70% within a week, we know that they tend to move into steep backwardation.   I also noted that a simple 10-20 MA system, when applied to the VIX index data, would have triggered an exit on October 12, more than a week before the big crash.   This of course is not completely comparable, but I think it does provide useful information.  Lets look at the 1987 Crash:

In general the VIX index has greater volatility and a larger trading range than VIX futures, particularly during extreme events.   If VIX futures been trading during this time, we would not necessarily expect the October 19, 1987 Spike to have been 100% reflected in futures.  We will observe this tendency when we look at three extreme events that have occurred since VIX futures started trading in 2004.  However, this extreme event does illustrate that risk planning is essential to all strategy and speculative traders/active investors.

Lets examine the relationship between the VIX index and and the October 2008 and November 2008 futures contracts leading up to and during the extreme events of October 2008.

What we see is that the VIX forward curve was in backwardation for the entire period (9/2/2008 – 10/31/2008) that we are reviewing.   This means that a trader who had used forward curve analysis during this time would have been long biased, or at the very least not maintained a short position during this massive spike in volatility which would have created extreme losses for short positions.  One other important feature to notice is that the VIX index spikes up much higher than the near month futures contracts.  As was mentioned above, this is a regular feature of how the VIX curve tends to behave during extreme events.  The VIX Index peaked at 103, the Oct. 08 contract at 63, and the Nov. 08 contract at 45.

So in thinking about the events leading up to the 1987 crash, it seems likely that this same type of Forward curve dynamic would have developed (though I don’t think we can know for sure).

Next, lets evaluate how VIX behaved during the March 6, 2010 “Flash Crash”.

The flash crash was interesting in that there was very little warning that an extreme event was on the way.  However, we still note that The VIX index jumped from 19 to 25 in the three days prior to the flash crash, and that the Index price went above the near month futures contract.

Lastly, lets look at how VIX acted leading up to and during the August 2011 selloff:

Note that the forward curve went into backwardation between the Index and the August contract on 7/26, and into backwardation for both August and September starting July 7/27.   Lets look at how the S&P 500 (represented by the futures contract) was performing during this time:

Note that the severe selloff started August 1st, which was after the VIX Forward curve went into backwardation.

So, while this study is not comprehensive,  we do find that during two recent major Volatility events, the VIX forward curve did provide meaningful information.  This is in fact not surprising, as VIX futures have historically had a positive expected return when the curve is in Backwardation.  I believe with examples like the above, we can begin to better grasp why this is the case.  We also have reason to believe that had VIX futures been trading during 1987, the 70% ramp up in index volatility (which occurred before the Monday crash took place) would have put the forward curve at that time into backwardation as well (though we can’t know this with certainty).

Good trading.

You might also like these articles:

Leave A Reply (2 comments so far)

  1. Nau 23
    2 years ago

    Great Post! I’m honored that my comments inspired this in-depth analysis. Although I’m very, very, skeptical about any sort of market timing strategy, maybe you’re onto something here when it comes to partial early warnings before the vix skyrockets. On the other hand, my inner contrarian would always be tempted to double down, and I would likely have been carried out to sea in 2008, though if I was able to hold on it would have turned out fine in the long run….definitely food for thought.

    Putting a little spin on your results, it seems like there are two types of dislocations that vix shorts should be afraid of: stock market crashes and flash crashes. The first, such as happened in 2008 or 2011, correspond to generally deteriorating investor confidence. The second is more a result of quirks in market structure, like 2010 and (at least partly) 1987. The latter seems really difficult to predict. Out of curiosity, do you ever have some sort of hedge for short positions (such as 75% short 25% long), my risk management tool of choice, or is it mostly stop loss and being selective about periods of exposure? 

    Michael Nau

  2. kainvest
    2 years ago

    Would the circuit breakers prevent the repeat of 87 style crashes? 
    Hence the VIX won’t spike as dreadfully as seen in the 87 crash?