Does historical VIX data suggest we're moving into a higher volatility regime?
So, we all know exceptional CBOE Volatility Index (VIX) pops tend to mean-revert over time, and often produce decent entry points for fade trades. With exceptions, of course, as sometimes the "fade" works spectacularly poorly. See 2008, for example, and also 2011 in a smaller way.
The theory is that the extreme in emotion on the fear side tends to reverse. The inverse isn't particularly true. Extreme complacency is kind of an oxymoron and not a great market driver, as we saw early last week. But hey, that's my opinion, and I only looked at it relative to the moving averages ... and I only looked at one-month and three-month windows. That's far from the definitive word on the topic. Dana Lyons takes a different look at it:
"We originally searched for all occasions in which the VIX first returned to below 20 after having spiked above 40. Upon looking, however, we noticed that there were a few near-misses that would be interesting if included. Therefore, we adjusted our search for spikes above 37. Yes, this is pure data-mining, and we pride ourselves on avoiding such gimmicks. After all, our ultimate goal is to make our clients more money. Thus, mining for trivial data points is a waste of our time. This study certainly falls into that trap, however, we think the results are interesting enough to publish – and may serve a valuable lesson too."
He came up with 13 distinct instances. And his results were quite mixed, to say the least:
"...The most interesting thing is that this signal was not at all a consistent harbinger of calmer, happier times for stocks. In fact, in the longer-term, i.e., 1-year out, this signal was the opposite of consistent. The results were binary. They were like Babe Ruth's batting output: a home run or a strike out, all or none.
"Of the 13 prior signals, 8 led to 1-year gains between +10% and +22%. The other 5 resulted in 1-year losses of between -12% and -40%."
He found the key to which way it went was whether the market itself was shifting into a higher vol backdrop. I call those "regimes," as in now we're at the tail end of a low-vol regime. These regimes are somewhat loosely defined, and last about four to six years. We're going to see a higher median vol this year than last year, which I'd suggest is a sign we're starting to gradually morph into a higher-vol regime. So, perhaps this was a warning shot of higher vol to come?
More likely, though, it's just a data point along the way that's not going to signal much of anything. The August VIX pop was at record speed in percentage terms, and it had an unprecedented stretch above its 10-day moving average. So it's not shocking that we saw a rather large drop back. "Fair" VIX, of course, changes over time. But that's over a relatively long frame of time. It doesn't tend to double in a couple days and then just stay at that new high level. As such, the drop back is better thought of as a return to the mean, IMHO, than much of a signal, and his work seems to agree with that thesis.
Disclaimer: Mr. Warner's opinions expressed above do not necessarily represent the views of Schaeffer's Investment Research