Counting On Complacency: An Incompletely Comprehensible History of VIX   Leave a comment

This is about the point when the moderate din over already subdued market volatility sliding still lower becomes a raucous argument between quixotic bulls avowing a new virtuous market cycle on one side, and those foam-flecked permabears augering Armageddon on the other.

If you’re reading this by accident, maybe you don’t know what I’m talking about; but let’s assume you do.

Like so many, I have been following the near-term structure of VIX, first out of historical and then professional interest (though VIX has very limited tactical utility in my backyard).  As the last few weeks have worn on traders continue to find equities adrift in a strikingly thin market, albeit with a slight, steady bid underneath. Like a feather, short-term implied volatility continues to dissipate, occasionally buoyed aloft by a gentle contrary wind but quickly returning to its silent swaying float lower.

If we step away from our increasingly shorter-term adaptive trading strategies and steady state gap paranoia for a moment and adopt a more generational time frame for the sake of observation, the medium-term (~5 years) does demonstrate VIX is incredibly low.  In fact, since the equity market bottom in March 2009, VIX has closed only six sessions at or below 14.75.  Three of these sessions occurred in April 2011, two in March 2012; and the final session occurred yesterday, 08/10/2012.

You didn’t really think I just happened to staggeringly wander in here to talk about volatility today, did you?

There are probably dozens of articles that have been or are being written about this right now.  I won’t read them, but I’m glad they’re being written.  For an event that obtains on so few (substantially less than 1%) of all market days in the past 5 years, VIX closing =<14.75 is noteworthy.

It was not always so.  VIX saw two sustained low vol regimes during which the index pinged back and forth between 10-15: 1992-1996 and 2004-2007.  14.75 in those days was pretty steep, something like breaching 30 has been the last several years.  These periods (the computation in the former case still a function of OEX) each had several major pullbacks where VIX handily snapped 15 to the upside; but those long vol events were short-lived.  Perhaps not surprisingly, OEX and later SPX registered magnificent bull markets for their relevant periods, bordering on +50%.

So: what about that interstitial period of Feb 1996 – Jan 2003?  During this window, VIX treated ~15 as support rather than resistance.  This was a period punctuated by the Asian Contagion, LTCM, and most notably NASDAQ 5k and the ensuing massacre.  So OEX and SPX underperformed.  …Right?

Well, no.  Despite these and other, lesser paroxysms of risk, OEX registered a comparatively blistering advance (Feb 1996-Mar 2000) from 320-850 (aside: a number it didn’t begin to threaten again by late 2007 whilst SPX was staking out a nominal new high) while SPX matched stride with 650-1550.  Of course, VIX well established from early 1997 in its 19-30 range (this would hold, the customary few outliers notwithstanding, until Spring 2003), somewhat mystifyingly printing around 21 at the SPX cyclical Mar 2000 top of 1552.

I throw that quick-and-dirty “VIX: the Early Years” survey out because a) cascades of wonky statistics are mind-numbing but oddly fun, and b) to set up a timely point: for all the carrying on and claims made to the contrary with great certitude: twisting meaning out of the flaky negative correlation VIX shares with equity returns (the S&P 500 these days) is very tricky business.  Anyone trading for very long whose head is still securely atop their neck knows that; certainly even the novice option trader at all acquainted with trading vol knows that.

For technical reasons much too exhausting, lengthy and well, technical to go in to here – and after all, the above two paragraphs make it plain enough – those undertaking an evaluation of the VIX:SPX relationship looking for linear, inviolable rules will come away disappointed.  Long volatility occurs not only in periods of fear, but relative contentment as well.  “Fear Index” is a popular misnomer just accurate enough to enamor the unwitting.  In general, (esp. short-term, implied) volatility regimes are more relativistic than the recency bias a lot of casual VIX commentators suggest (if you doubt this, go back over the part where I mentioned a cyclical market top occurring at VIX 21).  In other words, that VIX 19 at SPX 1552 didn’t care VIX was twice as “complacent” when SPX printed 796 several years earlier.

Back to the present: VIX =<14.75 – six times – including yesterday amidst dried up desiccated equity volumes – in 5 years.

Do I want to suggest you going heavily long equities here because stock investors are so “complacent”?  No.  I just told you “complacency” could shift into a VIX 50% lower than today’s close if we’re seeing a legitimate regime change a la January 2004.

Do I want to caution you against shorting vol here?  “Caution” maybe, sure: unless April/today are toe-dips into one of the above-alluded to changes, there are significant downside risks if caught wrong-footed and not incredibly nimble here. But then: no.  If you’re going to survive whatever comes next, you’ve probably thought of all this and more already.

As a short-term index/currency futures and spot forex trader, I rarely give more than passing notice to VIX.  But if I do, it is because I hear a shift in the frequency of the technical white noise over which I trade day-to-day.  VIX under that long-term regime dividing line at 15 is like VIX over 30: it’s infrequent; but when it occurs, something fairly significant is taking or is about to take place.

As with most extreme events, outcomes are binary: VIX at this 14.75 volatility Maginot line is no different.

Here’s a look at the previous 5 occasions of =<VIX 14.75 in the last 5 years; and a selection I culled from the beginning of the most recent longer-run low vol VIX 10-15 regime (Jan 2004-Jul 2007) range of data:

Admittedly, this is a pitifully low sample space.

But, look at it this way: there are two scenarios about to play out: either one is going to offer substantial opportunity (if on different time frames and favoring different strategies). They are:

1) Not-Complacency: Vol Somnambulance.  It isn’t likely VIX is entering a third multi-year 10-15 regime here.  It isn’t highly unlikely, either.  If it does, we’ll know soon enough and this is – a few 8-10% pullbacks and maybe a dislocative 15-20% correction e.g. 1998 aside – the kickoff of the second leg of the bull market begun at the 2009 lows.  In that case, the above table isn’t relevant and we become more interested in studying upside opportunities in SPX when VIX is <15, analyzing tail risk (for those brief >15 spikes) and downside risk in SPX where VIX approaches 10.  I’ll post those details as occasion warrants.

2) Complacency: Long Vol Lurking.  Assume instead VIX dips its toe here <14.75, finds the placid low vol water too brisk for its taste decides it would rather not jump in feet first.  The above table is…back on the table.  Every available indication suggest the R for any program long col and/or short SPX where VIX has closed =<14.75 or below is highly asymmetric (notice here columns J and K on the above sheet).
As a trader who only occasionally holds overnight, scenario #2 is by far the most preferable here: adequately quantified risk is much more easily extrapolated assuming implied volatility is bottoming versus passing through VIX 15 on its way to 10.

Departing from the pessimists, we can’t state that scenario #1 is a claim “this time really is different” – that smacks of recency bias with its oblivion to anything pre-dating 2007.  The truth is VIX has broken beneath 14.75 before and averaged well beneath it for years at a time.

Departing from the optimists, scenario #1 has occurred twice in the last 20 years under post-recessionary/mostly irrepressible expansionary conditions in the wider business cycle.  Today, though most agree the US is post-recession and currently out of recession, these conditions are mostly absent.  VIX 14.75 is all-but-unheard of over the last 5 years; and there’s nothing qualitatively different that suggests we’re about to witness this cyclical flip into 10-15.  This gives rise to a series of questions: Under what conditions could VIX revert to a longer-run low volatility environment?  Are there any notable precursors, quantitative or otherwise, that inaugurated these periods in 1992-1996 and 2004-2007?  Once established, what conditions perpetuated these regimes; and what events or cyclical changes ended them?

On balance: a VIX over 14.75 (and thus an imminent long vol bounce) is the baseline scenario until the weight of evidence shifts against it.

08/13/12 note: a few supplemental items to consider here as VIX closes yet lower today:

About these ads

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

%d bloggers like this: