Even before Adam Warner and I were humbled in our attempt to forecast the VIX one month out, I have been slow to warm up to the predictive value of the VIX looking out more than a month or two – all of which makes Eric Boughton’s “The Predictive Value of the VIX: Room for Divergent Opinions” particularly interesting reading.
First, the bad news: Boughton concludes that “knowing what the VIX is today is not likely to give you any aggregate useful information about whether returns will be high or low over the next year.” The good news is that the VIX does a reasonably good job of predicting future volatility over the course of the next 12 months.
While Boughton’s target time frame of one year may not be the best way to test the predictive value of the VIX, the tidbit I found most interesting was tucked away at the bottom of Boughton’s article, in which he notes that “buying the S&P on every day the VIX exceeded 30, and holding it for a year, resulted in an average return of 17.15% (as compared to the average return of 10.28% available for all twelve-month periods during the period in question).” His conclusion? Even if there is no correlation between the VIX and one year returns, “the market seems to pay an outsized return in exchange for the risk of buying when fear is high.”
My thinking continues to be that middling values in the VIX are generally not worth talking about. In the end, it is VIX spikes and mean reversion that matter most – and those factors are easiest to predict over the course of about 5-10 trading days.