VIX dropped slightly as the S&P 500 powered to more all-time highs last week. The slight drop in VIX can be attributed to the already extremely low level that it closed last week. The futures all dropped a little as well, which may be more about moving along in time than lower volatility expectations.
VIX closed under 10.00 more times this year than any other year in history. The average for VIX is historically low as well in 2017. So, why is the short VIX trade (using futures mostly) being referred to as a ‘crowded trade’? The markets are all about buying low and selling high, right? Well, not necessarily when we are talking about option implied volatility.
We’ve heard at CBOE Risk Management Conferences (CBOE RMC) in the past that volatility traders think selling VIX at 12 is easier than selling VIX at 20, or other similar sentiments. Those sort of statements along with our current market environment got me to do some high level quantitative work.
The two graphics below us VIX data from 1991 through the end of the first quarter of 2017 and realized S&P 500 Index data for the 21 trading days that follow each VIX close. Remember that VIX is a calendar day forecast and on average there are 21 trading days in each 30 day period, it varies based on holidays, etc, but 21 days is a good number in this case.
As a note - We have VIX data going back to 1990, but I’m on vacation wasn’t able to get the 1990 data while I was flying over the Pacific. I’ll update this to include 1990 when I return to the office and we will all give me a pass on this…
My first mini-study involved looking at the number of VIX observations by ‘handle’ that were followed by lower realized volatility as measured by S&P 500 price action. As of the end of the first quarter there were only nine sub 10 VIX closes so I grouped them into a category that was 10 or under. On the top end I grouped all VIX closes of 45 and greater together. The chart below shows the percent days by these groupings that realized S&P 500 volatility was lower than VIX.
This chart is pretty consistent with the exception of the 10 and below reading as well as when VIX closes with a 35 handle as well as when VIX closes at 45 or above. By consistency I mean over 80% of observations resulted in lower realized volatility than VIX. I get the lower percentage below 10 and above 45, but the 35 reading makes no sense and is probably just a result of a lower number of observations.
My second running of the numbers looked at the average realized volatility that followed each closing handle from 9 to 45. I abandoned the over 45 readings as there are several single instances at difference price handles and several numbers that have never be the final VIX print of the day.
The straight line above represents the closing handles and the red line that is not so straight represents the average realized volatility. They both trend higher, but as VIX moves up the spread between the two widens as realized is typically lower by a larger magnitude when VIX is at higher levels.
Hopefully these two graphics clear up the question as to why volatility sellers are still in the market despite the current low level of VIX. When VIX spikes and the futures follow suit, several short VIX futures positions may get stopped out and some of the profits that have been realized over the past few years give up. However, I am going to revive and paraphrase a quote that preceded the 2008 financial crisis. “As long as the music keeps playing, the short volatility traders are going to keep dancing.”