January 17th, 2013

Debunking the VIX Myth …. Again

imageI’ve had enough of this and it’s now time to again debunk the VIX myth.  This isn’t a new campaign of mine and have written about it here before:

  • In Debunking VIX Predictive Value of December 18, 2010, I compared the VIX on each of those 2,387 trading days (6/22/2001 and 12/18/2010) against the change in the S&P 500 Index over the following 30, 60, 90 and 180 days.
    • The VIX has been below 20 nearly as many times as it was above; when VIX was below 20, the market has risen 53% of the time, regardless whether over 30, 60, 90 or 180 days.
    • The VIX has been above 20 52.6% of the time and the market declined 61.7% of the time whether over 30, 60, 90 or 180 days.
  • In Ignore The VIX As A Market Trend Indicator of March 17, 2010, I included a chart that had both the VIX and the S&P 500 Index and wrote:

    “True, the VIX had an inverse relationship with the Index during 2008-09, but throughout the long 1986-2000 bull market, the VIX trend bore little connection with the Index having trended down and then up (indicated by the solid blue 300-day moving average) during most of the period…Falling VIX is interesting but I welcome it rather than finding it something to be fearful of.

  • In Does VIX Inform or Mislead? of October 2, 2008, I reacted to everyone’s call that a VIX in the 40′s (then unusually high) indicated that a bottom was near.  I wrote that “the VIX is a crude instrument and may be good at anticipating a short-term bounce but, until it hits extreme values (50 or more), it gives false indications that a bottom has been reached. Today, the VIX closed at a value of 45.26. I wouldn’t bet that we’ve touched bottom yet; a bounce perhaps, but not a bottom.”  My Market Momentum Meter indicated that the market was still far from having bottomed and, actually, would actually decline another 39%over the next five months to the true bottom on March 9, 2009.

We’re again hearing that the VIX is at all-time lows leading to diverging conclusions.  For example, a few days ago, the American Enterprise Institute in their blog wrote that “the VIX Index closed today at 13.45, the lowest level since May 2007, more than five years ago, and more than six months before the recession started in December 2007.  Investor fear in the stock market has been consistently subsiding, and market volatility is now back to normal, pre-recession levels.”

On the flip side of the coin, there are those who are contrarian investors and believe that a low VIX indicates too much complacency whish in turn leads to a market top.  For example, one blogger recently wrote “The VIX is the best measure of Fear in the market. Fear just hit 5 year lows and is about to hit its long-term lower trendline. Each prior lower trendline visit over the past few years predicted a fairly substantial market collapse in the ensuing weeks…..”  And Fox Business News summed it up by this interview of Bespoke Investment co-founder Paul Hickey on the outlook for the markets.

Let me take another cut at getting to the truth about the VIX.  In short, it is a lousy indicator of both near- and long-term market direction.  This time I did a statistical analysis and discovered that there’s very little correlation between the level of the VIX on any given day and where the market will be 50, 100, 200 or 300 days later.  I went back to April 6, 1998 and compared the value of the VIX on each day with where the market was at the latter date.  The statistics, after over 3000 observations, were unconvincing.  For those of you who are statistically oriented, the regression analyses produced the following results:

Regression ResultsThe most interesting finding in the data is how little the VIX contributes to explaining where the market will be 20, 50, 100, 200 or 300 days hence.  The data say says, for example, that the VIX’s level on 3400+ trading days contributed only 0.14% to where the S&P 500 Index actually wound up 20 days later, a miniscule amount.

Interestingly, the contribution of the VIX increases to 5.12% in accounting for where the market ends up 300 trading days later, still a negligible amount.

Perhaps the accuracy of the VIX in predicting the market’s future direction is best transmitted in a picture (this one is really “worth a thousand words”);this one is shows the market’s percentage change 20-days after the VIX reading (click on image to enlarge):

VIX 20 day

As you can see that the S&P Index changes very little over the 20 days following the VIX reading regardless of whether the reading is below 20 or higher than 50.  Furthermore, the market is just as likely to increase over the next 20 days as it is to decline…..again regardless of the actual level of the VIX, be it 15 or 65.

The major difference between the above chart for 20 days and those for longer spans between the VIX reading and the change in closing value of the S&P 500 Index is that the dispersion around the average change is wider.  The longer the time that passes, the greater the opportunity for more dramatic market changes.  To see those charts, click on 50 days, 100 days, 200 days, 300 days.

So if you hear or read a talking head warning of dire consequences too much complacency, of a VIX reading that is “too low by historical standards”, then remember this statistical analysis showing that the market is as likely to increase as it is decrease at nearly all levels of VIX readings …. that VIX and future market trends are virtually independent.

Technorati Tags:

December 18th, 2010

Debunking VIX Predictive Value

Before we all get involved away from the market, I want to wish all a Merry Christmas and a Happy New Year .

I also want to thank you for being here with me. I can’t believe it but am proud to say this is the 600th post of “Stock Chartist”. All of them can be found in the archives (I accidentally deleted some of these earlier posts so I don’t have a true count of total postings but I’m guessing it closer to 700). I started the blog in 2005 and hope to be celebrating its Sixth Anniversary on July 1.

Friends ask why I continue this labor of love for anyone who’s interested (or the small fee for those who want a more personal form of interaction by subscribing to Instant Alerts). To them I answer that it’s been fun, I earn psychic rewards when I turn out to be more correct than the well-paid “talking heads” and Wall Street pro’s and by helping you readers by offering a unique point of view and, hopefully, perhaps even some financial benefits.

=====================================================

Someone asked whether we should be concerned because the VIX is below 20 and appears to be making a head-and-shoulder pattern indicating that it probably will be heading lower (click on image to enlarge):

I’ve written before about what I believe is the lack of predictive information in the VIX (“Does VIX Inform or Mislead?” on October 1, 2008 and “Ignore The VIX As A Market Trend Indicator” on March 17, 2010) but the question continues to come up. How low could the VIX go and would it be ominous for the market even if it did decline? The “talking heads” always tell us that a VIX below 20 is a sign of complacency and a sure indication that the market is vulnerable to a correction or major fall. Some say that’s been the pattern ever since the Tech Bubble. A year ago, for example, CNBC posted an article entitled “Is the Low VIX a Warning for Stocks?” and repeated the old Wall Street saying: “VIX low. Time to go.”

But lets see what the statistics actually say. The S&P was 1225.35 on June 22, 2001, nearly 10 years ago but almost exactly the same level as last Friday’s close of 1243.91 (that tells you something but that’s the topic of another post). I wanted to see how predictive the VIX actually was over those ten years. I compared the VIX on each of those 2,387 trading days against the change in the S&P 500 Index over the following 30, 60, 90 and 180 days. I plotted the results on an XY-graph and came up with this amazing picture when all the possible combinations of 9165 trading days were overlaid (click on image to enlarge):

If you’re not interested in seeing the chart close-up I’ll tell you what it says:

  • The VIX has been below 20 nearly as many times as it was above; when VIX was below 20, the market has risen 53% of the time, regardless whether over 30, 60, 90 or 180 days.
  • The VIX has been above 20 52.6% of the time and the market declined 61.7% of the time whether over 30, 60, 90 or 180 days.

I conclude, once again, the VIX has no longer-term predictive value. For individual graphs of the different time periods click here:

March 17th, 2010

Ignore The VIX As A Market Trend Indicator

A friend said “A major correction is coming.” When I asked how they knew, the answer was “Because the VIX has declined significantly from its high of near 80 at the end of 2009 and is now under 18″.

If it was so easy to predict future market direction from the VIX’s level then why should I spend so much time with trendlines, moving averages and, when needed, other technical indicators like stochastics and Fibonacci levels. The answer, not surprisingly, is that the VIX isn’t a good forecasting tool other than perhaps extremely short-term fluctuations. Since this question surrounding the VIX seems to be on many investors minds now that the market seems to be moving out of its narrow, 6-month lateral trading range, I decided to do my own form of research.

Coincidentally, I found Laszlo Birinyi coming to essentially the same conclusion in a recent Bloomberg article entitled “VIX Doesn’t Work as Signal for U.S. Stock Returns, Birinyi Says“. Birinyi writes:

“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 VIX is a coincidental indicator,’’ Birinyi wrote. ‘‘It details, perhaps better than other measures, the volatility of the market today but not tomorrow or the day after.’”

If you’ve been brainwashed by some of the talking heads on CNBC then perhaps this long-term chart of the VIX with the S&P 500 Index superimposed on it for comparison might persuade you (click on image to enlarge):

True, the VIX had an inverse relationship with the Index during 2008-09, but throughout the long 1986-2000 bull market, the VIX trend bore little connection with the Index having trended down and then up (indicated by the solid blue 300-day moving average) during most of the period. During the last bull market between March, 2003 and December, 2007, the VIX again fell and then rose as the market climbed from trough to its peak.

Falling VIX is interesting but I welcome it rather than finding it something to be fearful of. If the last two bull markets are reasonable precedent, I could see the VIX continuing to decline another 40-50% from current levels to around 10.0 (the previous lows) while I simultaneously watch the S&P 500 continue climbing into new all-time high territory over the next several years.

October 2nd, 2008

Does VIX Inform or Mislead?

Don’t let anyone tell you otherwise. Today’s market action was frightening. The S&P 500 again violated where I estimate might be a support trendline; with what is expected to be a bad jobs report tomorrow morning and the House’s vote for the bailout, the odds are that the market will continue its cascade to lower levels.

But I need some help. I have been hearing and reading recently that we must be close to the bottom because the VIX is over 40 and getting this high each time “coincides with being close to the bottom”. I’m sorry, I don’t see it that way.

A couple of days ago, Barry Ritholtz, a blogger I admire and respect, wrote “Readings of 50+ are rare and obviously mark extreme panic. Assuming you have cash and know what you’re doing, it’s probably not a bad time to buy good companies”. He included the following chart:

So did some research and overlayed the VIX onto the S&P 500 during the Tech Bubble Market Crash of 2000-03 to see whether the VIX above 40 actually presaged a market reversal. This is what I found:

Four times over the three years, the value of the VIX exceeded 40. Subsequent to each of the first three occurrences, the market bounced for a couple of months but, several months later, it was signficantly lower (the dark line is the VIX and the light line is the S&P 500 Index; the S&P 500 scale is the right furthermost). Only when the VIX hit 50 or more did the market actual begin building a bottom.

Based on what I’ve seen, the VIX is a crude instrument and may be good at anticipating a short-term bounce but, until it hits extreme values (50 or more), it gives false indications that a bottom has been reached. Today, the VIX closed at a value of 45.26. I wouldn’t bet that we’ve touched bottom yet; a bounce perhaps, but not a bottom.