Is the VIX any good at predicting future volatility? Sure it is. But there’s another indicator that works just as well, and that’s past volatility. I took some data and created some charts similar to Artur Adib published a few years ago in a post called “VIX silliness.”
Indeed, the simplest measure of volatility–the standard deviation of S&P500 one-day returns over the past month–is an equally good predictor of market volatility for the following month. The plots above speak for themselves; note in particular the nearly identical correlation coefficients r. The data spans the period of 1990-present (data source: Yahoo! Finance).
Here are two charts I made to show what Artur means, updated to December 9, 2011:
The chart on the left shows the correlation between the VIX and the actual realized one-month volatility. The x-axis shows the VIX with the y-axis showing the realized one-month volatility one month later. The chart on the right is similar, but plots “today’s” one-month historical volatility on the x-axis with the y-axis showing realized one-month volatility one-month later.
Here are similar charts, but for the NASDAQ-100 index, using the VXN instead of the VIX. The VXN is similar to the VIX as it tracks implied volatility for the NASDAQ-100 index, but it’s only been published since 2001.
With both indexes, you get similar results. Today’s actual volatility is pretty much as good as implied volatility in predicting what volatility will be like in the near future.
Or is it?
Both sets of charts above use data going back many years, but what if we only use data since April 2009? That’s when the market began rising from the depths of the crash of late 2008/early 2009.
Here's a chart showing the VIX vs historical volatility as a predictive tool:
We see less correlation in this data. One reason could be that historical volatility actually fell to levels well under 10% several times during the period (under 6% at one point), but the VIX didn’t budge much below 15%.
And the result for the VXN was similar during the time frame with less correlation.
One commenter on Artur’s post said that “the volatility forecasting crowd are nothing more than modern day astrologers.” I wouldn’t go that far, but if one month from now you calculate the standard deviation of prices changes in the S&P 500 from today, don’t surprised if it’s pretty close to today’s volatility. Or today’s VIX. Take your pick.