A Curious Thing About Volatility

12/23/08 9:10 AM EST

Good Tuesday morning...we begin the day with very little movement in the pre-market futures.  Based on the amount of "out of office" emails I've received lately, trading desks are lightly staffed and that's only going to get worse as the hours tick by.  Overseas markets and the other macro tells are also little changed this morning.

I've received quite a few questions about the seemingly odd behavior of the VIX implied volatility index lately.  Most everyone knows about this indicator by now, as it is a convenient, real-time indicator of fear in the market.

Very generally, it measures traders' expectations of future volatility, and since volatility is a reflection of uncertainty, and uncertainty is a proxy for "fear", the VIX can be an effective sentiment indicator.

I'm not a huge fan of the VIX, never have been, but its popularity is undeniable.  Oddly, as the S&P has declined over the past four days, so has the VIX.  This is highly unusual, as the VIX almost always rises on days when the S&P falls.

I could find only two other times when the S&P and VIX both fell for four consecutive days (03/24/93 and 05/11/01).  The S&P rallied over the next week or so both times before falling back, but I wouldn't read much of anything into that.

More curious is the correlation between the two indices.  The charts below show the 5-day rolling correlation between the daily change in the S&P 100 and the daily change in the VIX index.  We can't use the pure price level of either for comparison because of some geeky reasons that mess up correlation calculations.

NOTE:  The Chicago Board Options Exchange, which calculates the VIX index, changed their methodology a few years ago, switching from the S&P 100 to the S&P 500, for calculating the VIX.  For the charts below, we use the S&P 100 and the old VIX index.

The first chart is from the last bear market, from 2000 through early 2003:

There was a cluster of readings early in the bear market, before we really even knew it was a bear market, when the S&P and VIX traveled together fairly often.  When the S&P rose, so did implied volatility; when stocks fell so did uncertainty.

That apparently wasn't a good sign for stocks, and during the length of the bear market, when the 5-day correlation between the two reached a positive 50% or so, it was a pretty good signal that something was amiss and stocks were about to resume their slide.

Now let's take a look at our current bear market:

Curiously, we haven't seen anywhere near the level of positive correlation we did during the last bear market.  The highest positive correlation we've seen was about +25%, and that was near a bottom, not a top.  The two other times it has approached positive territory, there wasn't much consistent about the price action going forward.

Currently, the correlation between the two is still solidly negative.  Even though the S&P and VIX have both closed down for the past four days, the magnitude of their differences has been great enough that it has not pushed the correlation much higher.

I don't like to spend a lot of time on stuff only to conclude "hey, this doesn't tell us much", but I've received an inordinate amount of questions about this relationship.  My conclusion would be that the correlation right now isn't strongly positive (despite the past four days), the last time they were both down four days in a row didn't lead to anything spectacular either way, and even if the correlation was positive, it has not been a consistent predictor during the current bear market.

Yesterday we looked at another measure of breadth, which is currently overbought.  Historically, overbought readings during a bear market are good selling opportunities, but as we've discussed a few times lately, extremely overbought readings tend to give the exact opposite signal and lead to rising prices, not falling ones.  Combined with other intermediate-term positives (like here and here and here), things still seem to look OK on a one- to three-month time frame.

Short-term, we have a few positive biases here, from the pre-Christmas glow (the S&P has been up 69% of the time 2 days before Christmas when the prior day was down), to the "Tuesday Turnaround" phenomenon (the S&P tends to rebound when it closes lower on a Friday and Monday), to the fact that the S&P futures have rebounded 92% of the time (12 out of 13) after having suffered four consecutive down days during December.

It's always hard to trust stuff like this when we know trading volumes will be very light and the slightest trigger could send us moving 10-20 points in a short amount of time and we still have a bunch of economic reports due, but overall I'm looking for higher prices into the holiday.

A move back under 850 on the S&P would have me back-pedaling on the "longer-term rally" idea, and a break of 820 will see me all but abandon it.  What we need to see is a sustained move over 920 to confirm the higher high / higher low pattern that is still up for grabs.

All the best,