Of VIX, Correlation, And Building A Better Mousetrap

Tyler Durden's picture

We have discussed the use of correlation (cross-asset-class and intra-asset-class) a number of times in the last few years, most recently here, as a better way to track 'fear' or greed than the traditional (and much misunderstood) VIX. As Nic Colas writes this evening, a review of asset price correlations shows that the convergence typical of 'risk-off' periods in the market is solidly underway. While we prefer to monitor the 'finer' average pairwise realized correlations for the S&P 100 - which have been rising significantly recently, Nic points out that the more coarse S&P 500 industry correlations relative to the index as a whole are up to 88% from a low of 75% back in February. In terms of assessing market health, a decline in correlation is a positive for markets since it shows investors are focused on individual sector and stock fundamentals instead of a macro “Do or die” concerns.  By that measure, we’re moving in the wrong direction, and not just because of recent decline in risk assets.  Moreover, other asset classes such as U.S. High Yield corporate bonds, foreign stocks (both emerging market and develop economies), and even some currencies are increasingly moving in lock step.  Lastly, we would highlight that average sector correlations have done a better job in 2012 of warning investors about upcoming turbulence than the closely-watched CBOE VIX Index.  Those investors looking for reliable “Buy at a bottom” indicators should add this metric to their investment toolbox as a better 'mousetrap' than the now ubiquitous VIX.

Nic Colas, ConvergEx: Building A Better (Volatility) Mousetrap

The most patented machine in American history has nothing to do with the Internet, or automobiles, or space exploration; rather, it is the humble mousetrap.  According to noted historian on the topic (yes, there is a mousetrap historian… See link after the note) Jack Hope, since the U.S. Patent Office opened its doors in 1838 it has awarded over 4,400 patents for distinct approaches to catching mice.  The process for obtaining a “Terminate mouse with extreme prejudice” patent includes categorizing your invention into a distinct subclass – “Impaling,” “Choking or squeezing,” or “Electrocution and explosive,” for example – and then proving that you have a unique approach to the process.   And in case that’s not enough mousetrap trivia for you, consider that the still-popular “Snap trap” was first patented in 1903 and is among the handful of inventions in this realm to ever turn a profit.  According to Hope, 95% of all mousetrap patents have gone to first-time applicants and fewer than two dozen have ever made any money for their inventors.

The desire to build a better method of catching mice is positively ingrained in American lore, exemplified by the old Ralph Waldo Emerson quote, “If a man can write a better book, preach a better sermon, or make a better mousetrap, than his neighbor, though he builds his house in the woods, the world will make a beaten path to his door.”   Innovation is a hallmark of American entrepreneurial spirit, whether it is pointed at a field mouse breakfasting in your pantry or fixated on developing the next tablet computer.  The mousetrap is, for better or worse, the symbol of that drive to create both.  Emerson’s observation may have seemed random at the time, but hoards of patent-seekers have proven him right since then.

In the same spirit of invention, I would like to offer up a “Better mousetrap” for measuring market sentiment.  The existing product – the CBOE VIX Index – has been around for +20 years in its current form and has done yeoman’s work at defining the conversation around how to measure “Fear” in equity markets.  Technically, the VIX is simply the price of short term portfolio insurance on a basket of large-cap U.S. stocks.  That insurance comes through the options market, where every contract is priced against the Black-Scholes five-input model and “Implied Volatility” (IV) is the unknown variable.  The higher the IV, the more investors are willing to pay for options that protect their portfolios against near term price declines.

My mousetrap is the actual price correlations between the S&P 500 sectors and the index as whole, measured on a trailing monthly basis.  Right after the text of this note is a summary table that shows the current levels for each of the 10 major industry sectors as well as some other asset classes such as commodities, international equities, currencies and fixed income investments.  We’ll get to the latter categories in a minute, but let’s focus on sector correlations for a moment.  A few points here:

  • Different industries typically do better at different parts of the economic cycle.  As the U.S. moves from recession and into the early stages of recovery, financials tend to outperform.  Towards the end of an economic upturn, commodity producers take the lead since their products have hit scarcity thresholds that offer them pricing power and cash flow.
  • Historically, active money managers have focused intensely on getting their sector over/underweights correct, and the common wisdom was that these decisions were 50% of outperforming the S&P 500 index benchmark.  The other half was picking the right stocks, but every manager knows the pain of picking the right stock in the wrong sector.   You are often better off picking the worst stock in the right sector.
  • Over the last few years, U.S. industry sector price correlations have yo-yoed between their long term historical averages (50-70%, depending on the group) all the way to +90% in any given month.  For example, the average industry price correlation for the 10 major sectors in the S&P 500 is 88%, measured against the index itself.  This is essentially the average of the past two years (87%), but still higher than “Normal” markets.  Average sector correlations have been as high as +95% and as low as 75% (historical charts attached).
  • While the average sector correlations we’ve outlined here tend to track the VIX pretty well (69% correlation, as shown in the accompanying graph), there has been a strange divergence in the past five months.  Simply put, the VIX is down 21% over the course of 2012, but the average sector correlation is +1.5 points over the same period.
  • Does the current investment environment look and feel like it is 21% less risky than the end of 2011?  Only if you are in cash, gold, shotguns and a cabin in rural Maine, I think.  The VIX has been unexpectedly quiescent in recent weeks, touching 17.5 on June 20th.  Even with Monday’s sell-off it barely held 20, its long run average.
  • Conversely, the correlation data is sending up a very visible warning flare about future market direction.  Back in February/March 2012, this indicator hit its low (good for stocks, since they are moving distinctly and separately) and began to trend higher (bad for stocks). Now, sector price correlations have essentially gotten back to their 2 years averages – 87/88%.  This would be like the CBOE VIX Index reading 28, which is its post-Financial Crisis average.  And frankly, a VIX at 28 would feel about right.
  • The real “Buy signal” from sector correlations is when then hit 95%, some ways from here.  As the accompanying charts show, this occurred in mid 2010 and Fall 2011 – both good times to buy U.S. stocks.

Just to round out the discussion, I would point out that there are asset classes that aren’t clustering around stocks like scared sheep in a thunderstorm.  Precious metals, for all their whippy action this past month, are fulfilling their promise to act independently of financial assets.  Gold’s correlation to the S&P 500 index was (18%) last month, and Silver was just +11%.  I have read some notes recently that were critical of gold’s ability to provide diversification in choppy markets.  This month, however, gold and silver are acting exactly as they should – with no eye to other asset class price movements.  U.S. high yield bonds are at the other end of the spectrum, and now trade more like domestic stocks (87% correlation) than sectors such Utilities (59% correlation to the S&P 500) and Consumer Staples (81%).    International equities – developed and emerging economies alike – trade more like US stocks than most U.S. stocks.  Their correlations are 91% (EAFE stocks) and 89% (Emerging markets), versus that 88% sector average.

My conclusion is that sector correlations are a useful adjunct to the widely-followed CBOE VIX Index when it comes to assessing how much risk is really priced into U.S. stocks.  It isn’t actually a “Better mousetrap,” in that the VIX is widely traded and tracked.  But it is a useful add-on tool, and one that is more accurately reflecting the risks imbedded in U.S. stocks at the moment. 


[ZH: as an addenda, we track implied correlation and cross-asset-class realized correlation almost every day in our various market posts, and furthermore, while Nic's approach at analyzing the sector correlations is extremely valuable in our eyes, we find the greater sensitivity of the average correlation of the entire 100 names of the S&P 100 (and the high-yield and investment grade credit indices) is a more accurate and better indicator for turning points in macro 'fear' and 'insensitivity']

The chart below is the intraday cross-asset-class correlation for today - this measures the minute-by-minute changes in commodities, rates, credit, and FX relative to stocks for a sense of whether stocks are moving systemically or idiosyncratically...

Clearly, stocks began to move on their own after mid-morning as correlations fell and then after-hours this evening, correlations have picked back up as stocks have resynced with systemic risk movements.

Below is the chart of 1-month and 2-month rolling average of all pairwise realized correlations in the S&P 100.  Compare this to the chart in the upper left of the charts above (above measures the high beta correlation between the 10 sectors and the S&P; below measures the much broader and more sensitive correlation across all 100 names in the S&P 100)

Three things should stand out:

1) We are following a very similar cyclical pattern of idiosyncratic to systemic fear rotation once again - i.e. as Fed measures lift and we are left to fend for ourselves so risk rises and systemic fears creep back into the market, slowly at first and then rapidly;

2) We are quite a way from any capitulative 'fear' level across the 100 names of the S&P 100 - i.e. there is considerably more pain to come as correlation is expected to rise; and

3) This realized pairwise correlation is much closer to the implied correlation levels we see by tracking the variation between index volatility and the average of all individual volatilies - i.e. this is a better day-to-day tracker for the premium in implied correlation (the real 'FEAR' index) over a realized correlation.


Finally, the chart above shows the average pairwise correlation for a rolling two-month period across all 125 names in the IG18 credit index. Three things should be apparent:

1) IG credit is considerably more sensitive to swings in systemic risk than stocks - i.e. the variation high-to-low is greater and more rapid;

2) We are at a more extreme level of realized correlation in IG credit currently, perhaps indicative of an increase in dispersion aboout to occur - or post-hoc a major market dislocation; and

3) the rapid rise in pairwise correlation post the JPM-CIO-Whale Debacle as investment grade credit spreads in IG become much more index-driven than idiosyncratic risk driven thanks to the need to unwind their position - i.e. a pronounced turn-down would be indicative of a slowing in the unwind. 


Bottom-Line - Don't worry about VIX, it's simply too contemporaneous with risk to be useful; understand and identify useful relationships in correlation to comprehend real 'fear' in the market. Right now, correlation is indicating a systemic AND risk-off mode in markets.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
RobotTrader's picture

Risk Off, Bitchez!!!

mcguire's picture

i was surprised that the correlation of silver to es was so low...  today was a welcome move.  


sablya's picture

Brilliant!  I've watched the correlation between markets come and go and when it is high it is like lemmings marching to the cliff.  And if there was ever a better indicator to buy VIX, I don't know what it would be.  

DeadFred's picture

Why is VIX so low? As it's described it seems it would be difficult to artificially manipulate. It also says that the real price of options to protect against a significant drop is not that much more than in the days when all the talking heads were saying the market would continue to ramp all year. 

wisefool's picture

Also, The NAR chick was on bloomberg today. Business hours, even.

Excursionist's picture

Good question - it's been a head scratcher for me as well.  Having been in the equity derivatives biz for a number of years, I can say that, at a gut level, the current environment 'feels' like something significantly north of 20.

A few guesses include:

  1. To generate income / juice yields when ZIRP is in effect, institutional guys typically not found in the derivatives sandbox are now selling VIX courtesy of both a belief in (eventual) mean reversion in the VIX and the 'Bernanke put';
  2. Increasing popularity of dispersion strategies that short the VIX and go long vol in individual components of the SPY;
  3. Declining volumes in SPY puts and calls are an unlikely culprit - easy to say investors are punching out of this bizarro environment, but CBOE monthly volumes are holding up for example.

Then again, perhaps I ate something bad for dinner, and my gut is completely wrong...  Perhaps the VIX is indeed correctly gauging future S&P 500 volatility, and one or more "strong policy responses" will keep markets at bay in the near- to mid-term.

lewy14's picture

I think your dinner was fine, and you're on to something.

Selling vol is what all the cool kids are doing - you know a) Bernanke et al will CTRL P in any significant downturn, and b) any upturn will be self-negating via oil and commodity spikes. So sell vol all day because nothing's gonna move. Evah.

The correlation thingee is awesome but it's just tracking risk-assets as they slide from slightly under-fucked (euphoria) to slightly over-fucked (panic). About the same these last three years.

Milton Waddams's picture

VIX has nothing to do with 'fear', instead it measures the market's perception on future volatility; which is fancy way of saying the market's expectation of the average daily price change and average daily price range (VIX was rising throughout the blowoff phase of the dotcom bubble).  It just so happens to be that periods of decline, on balance, typically are more 'violent' -- or faster paced -- than advances.  It may take one month to advance x% and five days to wipe out the entire move.

Interesting article nevertheless.  I expect the 'herding' of money to be a subject of ever increasing study (Maria B's mindnumbingly repetitive query of her guest, during every outsized decline, of 'so are you using this weakness to add to positions?').  Behavioral finance / economics is the new black.

macholatte's picture


·  Talking Numbers: Fear the Fear Index?

What's behind the 15% surge in the volatility index, with Brian Stutland, Stutland Volatility Group, and Carter Worth, Oppenheimer.


ZippyBananaPants's picture

Maria B and her two guests are idiots!

She should pull one more guy in from the floor and see if they can get her to go 'air tight'.

Dead Canary's picture

So you're saying asset price correlations.... wait...VIX Index gamma can... no, that's not it. Uh.... Oh hell I'll go back and read it again.
This is why us stupid people like the MSM. We're not stuck with serious thoughts all day.
Damn you Zero Hedge. DAMN YOU TO HELL!

CrawdadMan's picture

Could it could simply be because investors are becoming more diversified. did you consider that, OP?

slewie the pi-rat's picture

that contextModel correlation :: ES  chart for today isn't exactly normal is it?  looks like love or something to me  L0L!!!

 from 99% to 0% in a few hours?   and then back to 70+?   whoa, baby!  grains/commodity indices do that?  

sitenine's picture

VIX will spike soon enough.  Frankly, it doesn't matter though, which I guess is what Tyler was getting at with the article.  I don't know if you've noticed, but MSM headlines have all gone from "hope of ______" to "blame everything on Europe."  Some very large banks are very concerned, and you should be too, regardless of how many vols the VIX gains or loses each day. Do you really think paper of any kind is trading at fair value, anyway?  Does someone out there still believe we have free markets?

q99x2's picture

Risk off sideways or risk off down or risk off algos up or allgoes up or down.

Alejandrito's picture

Another good indicator is the number of news published in zerohedge.

chump666's picture

If you haven't got access to credit spreads/indices charts.  Use Asia as a proxy with USD bid/sells (you'll be up all night).  As for the VIX again good proxy, study Asian trading patterns and VIX correlation which looks similar to various crashes in history (VIX trailing 16, 17 then spiking re: 1987 onward etc).  The market is still correcting but hasn't priced in doomsday.  Which is probably close and it's NOT Europe.  Meantime we have swing trades in ranges, with US markets now starting to correlate with the sell off in Europe/Asia. But oversold now.

It's all Asia, three months of USD bids and market volatility has lead to clockwise cyclical trades on European and American markets.  Commodities being sold i.e oil, usually sends equities lower (deflation trade).

The European saga is bank/liquidity/credit related and since the ECB is now becoming irrelevant (they print, buy bonds doesn't do sh*t any more), then Asia is the cue and of course the mother of insane: The Fed.  Which will monetize directly into Europe (ala The NY Fed swaps has been doing: 500million in June). 


chump666's picture

hang seng and shanghai have gone bid. 

hugovanderbubble's picture

TVIX will hit 15 during Summer

ciupaquo's picture

maybe, VIX (future) is low because market is betting on FED's action shortly ? We'll see one/two months of market rise....

Mercury's picture

Cool.  But I like ZH's more broad-based indicators better, especially the first one.

Measuring the S&P 500 vs. it's 10 component sectors doesn't seem very useful.