When it comes to CBOE Volatility Index (VIX) analysis, time frame is everything
Is the
CBOE Volatility Index (VIX) low or high? Is it moving too much vs. stocks or too little? Does a strong VIX mean certain doom, or is it signalling too much fear?
What if I told you I could answer all those questions with a "yes," even though they essentially all contradict each other? I'd note that
it's the time frame that really matters for all the above.
Is VIX low? Well, it has a 15 full now. The
all-time median is 18.21. Even the "Low VIX Regime" median is 17.10 -- at least in this current regime, which I start in July 2009. It's even low vs. 2015; this year's median is 16.56.
But VIX isn't low vs. realized volatility (RV). Ten-day RV in the
S&P 500 Index (SPX) is about 10.3, so options are already ahead of the next volatility pop.
A lot of that modest overpricing is thanks to the fact that VIX has outperformed the market lately. By that, I mean it's overreacting to small market dips. The
SPDR S&P 500 ETF Trust (SPY) dropped 0.27% on Wednesday. VIX has typically moved about -6x SPY in recent months, which suggests a pop of something like 1.8% on Wednesday --
yet it actually popped 6.67%.
These relationships are noisy and imperfect. VIX lifts *more* on Mondays, lifts ahead of anticipated news, et. al. But we didn’t have any of that. Wednesday has no "Day of Week" effect, and there was no real anticipated market news (no,
Facebook isn’t going to move market vol). If anything, Wednesday itself saw actual news that the market often cares about: Fed Chair Janet Yellen commented on the possibility of a December rate hike! If anything, this backdrop suggests VIX underperformance, yet we saw the opposite.
If I carefully choose a time frame, I can "prove" VIX is wildly outperforming. It's up 10% since Oct. 23, while the market itself is up about 1.5%.
So finally, does this bode badly for the market on a smart-money sort of thesis, or is it simply too much fear and a bullish contra tell?
Again, results are mixed and inclusive. I ran 20-day rolling averages of the VIX Move/SPY Move ratio, and then correlated it to one-month forward SPY returns to see if there was any predictiveness about the VIX/SPY relationship. And I found virtually no relationship. The correlation was 0.0147 -- the definition of non-correlated.
I'd guess that the modest recent VIX strength is a bit of a counter reaction to the overall VIX drop since
the August spike. Hey, VIX must be "cheap" now since it was so much higher not that long ago! But the truth is, it's only cheap if you narrowly define it, and the very recent strength isn't all that predictive.
Disclaimer: Mr. Warner's opinions expressed above do not necessarily represent the views of Schaeffer's Investment Research