It's usually best to fade VIX pops from October through January
In running some numbers yesterday, I randomly stumbled on an intriguing factoid. The one-day move in the CBOE Volatility Index (VIX) doesn't tell us all that much in terms of future stock action. It had a correlation of close to 0 (specifically, 0.0249) relative to one-month forward SPDR S&P 500 ETF Trust (SPY) returns.
But when I isolated the correlation of the VIX action on the first trading day of the year to one-month SPY returns, it showed about a 0.25 correlation. That is, of course, not high on an absolute basis, but it's high enough to suggest there's some value. And interestingly, it had a positive correlation, meaning that the stronger the VIX on Day 1, the better for the market going forward. That's kind of contradictory to the usual mindset that early January market performance gives you a signal on the whole year. Remember, VIX moves in opposition to the market.
Anyway, I was curious whether there's any sort of seasonality to the usefulness of VIX. That is, do we see higher or lower correlations to future returns based on the calendar month?
I separated into calendar month and looked at the correlation of VIX one-day moves vs. SPY performance one month out. And here's how it looks:
January is actually the third most "correlated" month, after December and November. It's important to note, however, that these are all very low numbers:
Still, you can see a seasonality pattern. One day of VIX doesn't tell us all that much, but it tells us the most from October through January. And it's a contra tell, albeit a very small one. Here's another look, this time it's correlation using VIX vs. its 10-day simple moving average:
Again we see that December and November VIX show the highest correlations, followed by January:
And these numbers are bit higher. Also of note, the one big outlier: September.
Is a September VIX pop "not" a contra tell? Perhaps. But I suspect we're talking a sample size of two here. We had big VIX pops in September 2001 and 2008 that did not presage market lifts a month later. Quite the opposite. Still, I wouldn't fade a September VIX pop so quickly.
What's it all mean? It suggests that on the margins, VIX pops from October to January are the best VIX pops to fade, September is the worst, and the rest of the year we don't learn a whole lot. It's important to emphasize that these observations are on the margins.
Disclaimer: Mr. Warner's opinions expressed above do not necessarily represent the views of Schaeffer's Investment Research.