Should We Follow The Vix?

Came across an interesting article on Bloomberg the other day, titled VIX Doesn't Work As Signal For Stock, Birinyi Says.

Investors looking for clues about the US stock market should probably ignore the Chicago Board Options Exchange Volatility Index, according to a study of the VIX by Birinyi Associates Inc.

Speculation that equity returns will be positive after the volatility gauge decreases and negative when it climbs has little basis in fact, Birinyi said. The VIX provides a summary of historical price swings and tends to move in lockstep with equities instead of forecasting their direction, the firm found.

“The VIX is alleged to be an indicative indicator and has become a staple of analysts and journalists alike,” Laszlo Birinyi and analyst Kevin Pleines wrote in a report to clients yesterday. “We respectfully disagree and ultimately conclude it is a measure of current volatility with little or no predictive or indicative value regarding the course of the market.”

This sparked a blogstorm.

We have Chris McKhan of Optionmonster. We have MarketBeat from the Wall Street Journal. We also have David Rosenberg via Josh Lipton yesterday in What the VIX Really Tells Us:

The VIX index slides from 42 on October 9, 2002 to sub-18 exactly five years later and the S&P 500 doubled; the VIX then surged to 50 by March of 2009 and the market was down 60%; the VIX then went on to plunge from 50 on March 9, 2009 to sub-18 on January 19 of this year and the stock market soared 70%. Since January 19, the VIX has been flat and so has the equity market.

And why not.., let me chime in further too.

The header and the first few paragraphs are a bit more controversial than the actual work. Here's Birinyi's summary:

For those whose time and patience is limited, we will at the outset present our conclusion: The VIX is a coincidental indicator with limited predictive value. It details, perhaps better than other measures, the volatility of the market today but not tomorrow or the day after.

In my opinion, he's right. There's some notion that the VIX today does some magical job predicting volatility of tomorrow. After all, it's a measure of 30-day forward market prices for volatility. But at the end of the day, the VIX does a better job telling you what has just happened in realized volatility (volatility in SPX itself) than it does in actually predicting what will happen over the next 30 days. Correlation is greater to the recent past than it is to the 30-day future it seeks to price. Now that makes perfect sense to me, I mean what else would you base options pricing on?
I'm not sure I understand Rosenberg's counter-argument, it doesn't refute Birinyi's basic point, which is that the VIX is simply a coincident indicator.
Now none of this is to say the VIX is useless. The VIX can give useful info when compared to ... itself. It generally mean reverts, which we can objectively define as gravitating toward its moving averages. As it stretches, it's not the worst idea to look for a reversion. Well, usually not the worst idea.
The VIX also gives clues when viewed under a subjective lens. Take right now for example, the VIX is at something like 21 month lows in absolute numbers. But at the same token we have a 10-day realized volatility in SPX near seven. Given that the market's rightfully assuming (for now) that realized volatility will pick up, it's understandable to have a VIX premium here. So I'm not reading much into it right this second. If the non-volatility persists though, you have to start seeing the VIX drift again, otherwise it's just too much Fear.
If you're looking for some sort of predictive magic, it's not there. But for some info on the margins, there's plenty to glean.