The Bust and the VIX

Adam Hamilton     January 24, 2003     3005 Words

 

Actively speculating in this once-in-three-generation ongoing Great Bear bust is endlessly fascinating.  Every week brings interesting new surprises and outstanding opportunities to learn and grow as a speculator.

 

As the bust grinds on and increasing amounts of general investors’ scarce capital is unceremoniously slaughtered and dispatched to financial hell, speculators monitoring the carnage and taking notes continue to grow wiser.  After enough trading and reconnaissance, market behaviors that seemed anomalous and strange months ago now naturally seem a little less puzzling.

 

And of course as old puzzles are solved fresh new mysteries present themselves regularly to keep the speculators thinking sharply and clearly.  As I suspect that everyone who shares my passion about understanding the financial markets would agree, there is seldom a dull moment for speculators.  What a wondrous and glorious time to be alive and trading!

 

On the market puzzles front, there is one that has been sorely vexing me personally for quite some time now.  It involves volatility, the degree to which a given stock index changes from day to day.  Since volatility is one of the most widely used technical tools today by active speculators, it is prudent to attempt to resolve any mysteries or apparently anomalous behavior with volatility as soon as possible.

 

The cryptic apparent volatility anomalies that many speculators have observed in recent months are two-pronged.  Thankfully after a great deal of pondering we finally have some new ideas to present as potential hypotheses to explain this dual-headed volatility mystery.  I have no idea if these hypotheses will stand the test of time and be proven correct or be crushed by the markets as mere foolishness, but I thought I would share them with you anyway.  Perhaps they will help stimulate further discussions.

 

Here are the two volatility anomalies that have been troubling me.

 

First, there is an apparent contradiction between normal volatility behavior in a Great Bear stock-market bust and the expected behavior of the VIX S&P 100 Implied Volatility Index.

 

In a Great Bear bust, general volatility levels rise throughout the bust process.  As discussed in “Volatility Squared” a couple years ago, these rare stock-market bust events have a distinctive volatility signature.  In 2000 I coined these signatures Double Volatility Tops. 

 

As a stock index that is deflating from bubblicious levels to back down under fair value grinds lower, volatility generally climbs higher.  Eventually two distinct volatility super-tops form, the first spawned by the initial crash event right after the bubble pinnacle and the second arising out of the long selling capitulation climax marking the Ultimate Bottom.  Our graphs farther down below highlight this phenomenon.

 

Generally rising volatility throughout a bust is fine and good and makes sense.  As general fear grows so does volatility.  Yet, last September when the bears were noisily calling for a spectacular crash and forecasting an imminent VIX 150 super-spike, I learned something provocative.  While working on the background research for “Handicapping Market Crashes”, it became readily apparent that crash events in history all happen at a certain stage in bubbles.

 

Stock-market crashes in history only erupt slightly after bubble tops as explained in the essay, not years into the bust.  We couldn’t find a single example of a late-stage-bust stock-market crash in a major stock index in history.  Yes, bear markets grind stocks lower and lower to horrific undervalued levels, but the process takes years, not mere days. 

 

By slowly relishing its grisly work, the Great Bear ensures that it can trap and slaughter the maximum number of bulls in the bust.  It is like the old parable about boiling a frog by putting it in cold water first then gradually warming the water. 

 

If the markets are crashing all the time, which is about a 10% loss in a single trading day, the bulls would be scalded by the boiling bearish waters and flee the market pot far too soon.  But, if the Great Bear slowly brings the water to a boil with an endless series of modest down days punctuated by exciting rallies, the bulls will hold out their dogged optimism to the very end and lose everything.

 

If probabilities vastly suggest that we aren’t going to witness the first late-bust crash in history this time around, then a VIX 150 super-spike, which requires a single-day crash to emerge, is not going to happen.  A VIX 50 level was high in 1998, was high after 9/11, and was high last October.  The markets launched strong rallies off of each of these VIX 50ish events.  Drawing the undying wrath of some of the aggressive bears, I believe that VIX 50 will be a high level in the future too. 

 

But, and here is the glaring contradiction that some of you have very astutely spotted and pointed out to me, how can this be?  If general volatility rises throughout busts isn’t it a good bet that the VIX implied volatility index will reach startling new highs too?  If a VIX 150 super-spike requires a crash, but late-bust crashes are unknown in history, how can general volatility continue to rise if the usual VIX 50ish levels are still its highs in the rest of our current bust?

 

These questions are not merely academic, as our Zeal Intelligence subscribers, my partners, and I currently have vast open short positions for this downleg and we need to know whether or not we should close our shorts on a VIX 50ish reading or hold out for a VIX super-spike. 

 

If we fail to close our positions for legendary profits at VIX 50ish levels because we are holding out for a VIX super-spike that then never comes to pass, odds are we will ultimately suffer a huge loss in the next major bear rally.  Conversely, if we close at VIX 50 but a VIX 150 super-spike shortly follows, our realized profits will be dramatically lower than they could have been.  These questions are really important for speculators!

 

Well, I finally have a hypothesis I would like to share with you this week that may explain how general stock-market volatility can increase as it should yet the VIX will most probably witness no mega-spikes far above 50ish levels as the bust grinds lower.  Perhaps there is no contradiction at all.

 

Before we dig deeper though, today there is a second volatility mystery surrounding the VIX.

 

Way back last September in my essay “Volatility Trading the QQQs”, I discussed using the VIX as both a long and short signal for short-term tactical NASDAQ 100 index speculators.  The VIX signal to go long the QQQs is unambiguous, the VIX 50ish spikes mentioned above.  But, the VIX signal to go short the QQQs is much more difficult to discern in real-time.

 

This short signal involved looking for congestion in the VIX chart “under 30, near the 25 range”.  Because it was difficult to spot on the charts in real-time though, we searched the technical indicator world and eventually found the 21-day moving average of the Put/Call Ratio to be a superior short-entry signal, so we ceased using the VIX 25ish congestion as our primary shorting signal.  Continually refining one’s trading models as new information is digested is crucial for speculation success.

 

While our dear subscribers were actively trading on the new PCR 21dma short signal first-thing in December at the latest bear-rally top before I penned an Internet essay on it, I didn’t fully explain our shift in strategy publicly until this month.  In “Trading the NASDAQ Bust 2” I outlined our new binary QQQ speculation strategy, using the VIX for our long signals and the PCR 21dma for our short signals.

 

Nevertheless, even though we have moved on from reliance on the VIX 25ish congestion theory as our primary shorting signal, I am still interested in learning why it didn’t happen this time around as expected. 

 

In addition, there are countless free-riders out there who diligently read my humble rants every week but don’t believe we at Zeal should be able to earn a living to feed our families from our hard work, so they don’t subscribe to our modestly priced private newsletter to support Zeal Research.  Unfortunately these folks totally miss out on our best actionable and specific real-world investment and speculation recommendations and the rationale behind them, as our monthly Zeal Intelligence newsletter is the real-world application of the principles articulated in these weekly research essays. 

 

Sadly some of these non-subscribers still frantically write me every day with great concern that they missed the optimal time to go short because they waited on the VIX 25ish congestion that hasn’t yet happened this time around.  Unfortunately I think they did miss the boat and lost their opportunity to short near the top of the last major bear-market rally.  Active speculators must constantly learn and adapt or be left behind.

 

For our own curiosity as well as for those unfortunate traders who are still waiting on the VIX 25ish congestion, we also have a new hypothesis on why it hasn’t happened yet this time around and why it probably won’t.  Perhaps this second VIX anomaly is explainable as well.

 

With the benefit of a brief background on these two apparent volatility anomalies, we can now wade into some graphs that may offer the potential resolutions.  First this week we have a few updated graphs we have used before in past essays explaining the double volatility top bust volatility thesis.

 

 

While a few decades from now the 2000 NASDAQ mania will probably eclipse the Dow Jones Industrial Average episode of 1929 as the most notorious popular memory of a historical bubble, the original Great Crash episode is still awesome and instructive.

 

It is really interesting how general volatility continued to rise throughout the bust as the stock markets continued their long, slow-motion collapse.  The red line above is the 100-day moving average of intraday volatility for the Dow 30.  Intraday volatility simply describes how much a stock index moves on any given day in absolute percentage terms.  It is calculated by subtracting a day’s low from its high, and then dividing this difference by the previous day’s close.

 

In addition to stocks down and general volatility up as a bust wears on, the classic signature double volatility top is highlighted above in yellow.  The first volatility top always occurs right after the initial crash event within weeks after a massive bubble peaks on unimaginable general euphoria and “New Era” optimism.  Future investors and speculators can use this indicator if they fear a bubble may have just crashed and they want to know if a bust is rapidly approaching like a fearsome juggernaut.  This key tool helped us go short in 2000 way ahead of the thundering herd.

 

The second volatility top hovers around the Ultimate Bottom of a mega-bust.  It is crucial to realize though that this second volatility top doesn’t occur on a single-day crash, but the final selling climax lasts weeks or even months.  Daily intraday volatility may stay above 5% for weeks on end, but daily 10% crashes are not witnessed.  Here is a graph borrowed directly from “Handicapping Market Crashes” that highlights this essential point.

 

 

It is interesting to note how the top 20 big down days in the 1929-1933 Dow 30 bust episode corresponded with the double volatility tops shown above.  While the first volatility top is composed of one or more 10%+ true crash days spaced closely together in time, the second volatility top is fueled by months in a row of 5%ish volatility days that are brutal for the remaining bulls but definitely fall far short of classic crash material.

 

We can observe this same general phenomenon in today’s horrific NASDAQ bust, rising general volatility over time as the index itself continues to plunge towards frightening new lows.

 

 

The first volatility top in the NASDAQ occurred as the market crashed.  And yes, it was a classical index crash as the NASDAQ Composite witnessed a 10% monster single-day plunge of crash magnitude in April 2000, exactly 5 weeks after its March 2000 bubble apex.  Its second volatility top, like others in history, probably won’t occur until around the time the Ultimate Bottom of this bust is reached.  Unfortunately for the long-suffering bulls that’s a lot lower from here based on the NASDAQ’s still-ludicrous valuations.

 

Provocatively, the NASDAQ’s first volatility top above was around 4%, just like the Dow 30’s in 1929.  If history follows true to form, the NASDAQ’s second 100dma intraday volatility top will also spike above 5% at the Ultimate Bottom just as the Dow 30’s shown above.  If this bust isn’t over until the fat lady sings, then the second volatility top along with classically undervalued levels in valuation terms are the most likely fat-lady candidates. 

 

So, our current bust to date is behaving just as history suggests it should, with general volatility rising as stock-index levels continue to plunge.  Yet, is it possible for this general volatility uptrend to continue while at the same time the VIX does not continue rocketing up into an earth-shattering VIX 150 super-spike?

 

Call me a lunatic, but strangely I think it is possible, with no contradictions, for general volatility to continue rising towards The Bottom while the periodic VIX spikes we see are still highly unlikely to trade decisively above the VIX 50 level.  VIX 50ish may indeed still prove to be the effective “ceiling” level marking widespread fear at every major bounce point spawning a temporary tradable bear-market rally from now to The Bottom.

 

This final graph illuminates why this weird hypothesis may indeed prove to hold water over time.

 

 

As this graph using the ubiquitous QQQs as a proxy for the ongoing NASDAQ bust shows, stocks can continue sliding and the VIX implied general volatility index can continue rising without a VIX 150 super-spike!

 

The key to the mystery may not lie in the height of the VIX spikes, but in their girth!

 

Please note the 3 yellow-shaded triangles highlighting the last 3 VIX spikes marking tradable major bear-rally bounce points.  These VIX spires don’t seem to be growing significantly higher over time, as even way back in the 1998 panic the VIX spiked to 50ish at the bottom of that scary episode.

 

While subsequent VIX spikes since 1998 haven’t towered higher and higher, they have grown enormously larger bases since the NASDAQ bust commenced in early 2000.  If you visualize the areas under the 3 yellow-shaded VIX triangles above as pyramids, their bases are growing more and more massive over time even though they are not soaring higher into the heavens!

 

So perhaps speculators out there looking for a VIX 150 super-spike are searching for a classic soaring spire-like VIX chart pattern like the spiring CN Tower in Toronto when they should be thinking of an ever-more massive base on an immensely-solid pyramid like the Great Pyramid outside of Cairo.  Perhaps the size of VIX spikes marking fear-laden temporary bounce points can be measured not only vertically but horizontally through the uncharted seas of time as well.

 

If this hypothesis proves correct, general stock-market volatility can continue rising but VIX 50ish levels can still mark major turning points to temporarily throw long for active index speculators.  The VIX could manifest this rising volatility not in a super-spike to new extremes not witnessed since the unprecedented events of October 1987, but in a generally higher VIX level over time.

 

Maybe the subsequent VIX pyramids will simply grow larger and larger bases to manifest the growing general fear as more and more bulls throw in their towels and capitulate before the Ultimate Bottom is reached.  The VIX in general will still be in an uptrend, but the interim VIX tops could still be reached around the usual VIX 50ish levels.

 

Interestingly, this idea also brings us full circle on our second VIX anomaly as well, the failure of the VIX congestion to develop at 25ish this time around necessitating switching our current Zeal NASDAQ bust speculation strategy

 

If you take another look at the graph above, there are also three white circles.  The first two correspond to the VIX 25ish congestion levels originally discussed in “Volatility Trading the QQQs”.  They were indeed around a VIX level of about 25, marking the end of each major bear-market rally and the birth of the next brutal bust downleg.

 

Yet, if the general VIX level will rise through time with general stock-market volatility as expected in a Great Bear bust, perhaps the VIX congestion levels will rise through time as well.  Provocatively, if some trend lines are drawn through the first two VIX congestion levels on the graph, they create an uptrend pipe that intersects a level around the low 30s today.

 

Almost spookily, the third white circle in the graph marks an episode near the recent bear rally’s top that could conceivably be the VIX congestion level traders were diligently searching for!  If general fear rises throughout the bust, and general volatility rises throughout the bust, and the VIX rises throughout the bust by lingering at higher levels longer, it is not unreasonable to assume that subsequent VIX congestion levels will also trend higher.

 

Indeed, this is what the graph shows!

 

While we did officially switch out the VIX congestion in favor of the far-more precise PCR 21dma as our primary index-speculation short signal in “Trading the NASDAQ Bust 2”, I do think the VIX congestion is still interesting and useful.  We will continue to watch for it in the future as a secondary shorting indicator, but at slightly higher levels in each subsequent bear-market rally until The Bottom is reached.

 

Thank you for hearing me out on this pair of vexing volatility puzzles.  Perhaps our new hypotheses can adequately explain and resolve these apparent volatility anomalies. 

 

Only time and the markets will prove whether these controversial ideas are right or wrong.  Regardless of how it all plays out though, as an active speculator constantly striving to deepen my own understanding I am very excited to watch the action unfold in the coming months.

 

Adam Hamilton, CPA     January 24, 2003     Subscribe