Handicapping Market Crashes

Adam Hamilton     September 20, 2002     2992 Words

 

As we plunge ever farther into this amazing supercycle bubble bust, one popular pastime amongst the bears is to speculate on the timing of crashes in the US equity markets.

 

Stock market crashes are really interesting phenomena and they can wreak great havoc, so it is not surprising that they have an almost legendary place of prominence in popular market lore.  A speculator who is fortunate enough to anticipate such an event and go short has the potential of reaping enormous profits in mere days or weeks, a very alluring prospect.

 

I have been pondering stock market crashes lately due to some feedback I received on two recent essays on using the VIX implied volatility index to time short-term speculation entry and exit points.  After both “Volatility Trading the QQQs” and “VIX Bounces S&P 500” were published, I received dozens of similar e-mails from around the world dwelling on one issue.

 

While the level of 50 marks a normal VIX extreme in recent years, a contrarian speculation buy signal, there was one stunning event in history where the VIX actually exploded to almost unthinkable heights.  On October 19th, 1987 the VIX rocketed up to close at the breathtaking level of 150!  Merely one trading day earlier it had closed at 36, a normal level.  Over the next five trading days, the VIX closed above the staggering level of 100 three more times.

 

From October 19th to October 30th the VIX closed above 60 for the entire ten trading-day stretch, an unprecedented development never seen before or since.  On a long-term VIX chart, this huge anomaly sticks out like a central banker at a rap concert.  This mega-VIX spike was an incredible event that will go down forever in the annals of market legend.

 

Folks across the planet, including professional hedge-fund managers, graciously offered me some wonderful feedback after my two recent essays discussing using the VIX to time trading in stock indices.  The crux of their arguments swirled around the massive 1987 VIX 150 spike.  They reasoned that since the VIX had witnessed such stratospheric levels before, it could probably revisit them again and therefore the normal VIX extremes in the 50 range might not be a valid trading signal in these tough bear-market times.

 

Did October 1987’s colossal 150 spike in the VIX invalidate current VIX extremes of 50 as being useful as trading timing indicators?  We decided to take a look at this important question for speculators in this week’s essay.

 

As always, it is very important to discuss the VIX 150 mega-spike in its actual historical context.

 

In late 1987, the US equity markets had been galloping in a virtually uninterrupted bull-market stampede for just over 5 years.  From the Dow 30 low of 776.9 on August 12th, 1982, the US equity markets had roared up to 2722.4 on August 25th, 1987.  This incredible 250% march to the heavens in the flagship US equity index was extremely impressive, especially after the brutal famine years for equity investing in the 1970s and early 1980s.

 

All was not well however, as cancerous doubts about the markets’ future prospects began to metastasize throughout the thundering herd of investors.

 

Bull markets never run forever, and 5 uninterrupted years of cutting a straight swath northwards was starting to make many investors skeptical and nervous.  By August 1987, the general US equity markets were trading at 18.3x earnings, cheap compared to our bubblicious markets today in 2002, but very expensive compared to the bottom around 7.7x earnings in 1982.  Were the markets hopelessly overvalued as autumn 1987 approached?

 

Multiplying the growing trepidation exponentially, long-term charts of the Dow 30 looked just like classic equity bubbles, with a long-flat tail to the left growing into a near vertically-sloping bubble blow-off to the right.  Many troubling comparisons between the unsustainable accelerating uptrend of the Dow 30 in 1987 were made with the Dow bubble of 1929, almost six decades earlier.  In addition, wave and cycle theorists were all over the Dow’s parabolic rise like hungry sharks on a drowning swimmer, noting that one whole Kondratieff cycle had transpired since the last crash in 1929 and another crash was imminent.

 

And so it was.  On October 19th, 1987, known forever after as Black Monday, the day the VIX spiked to close at 150 for the only time in its history, the US stock markets crashed.  While we can look back today with 20/20 hindsight and smugly discount the events of late October 1987, they were absolutely terrifying at the time.  If you meet anyone today who was speculating in 1987 who claims they weren’t scared to death about the crash, they are lying through their teeth.  It felt like Equity Armageddon when it happened, exceedingly frightening.

 

Why was it so scary?  Because nothing like October 19th, 1987 had ever happened before in US history.

 

The Dow 30 plunged by an unbelievable 22.6% in that single trading day, by far its biggest single-day percentage drop in history.  The Dow 30 has only plummeted by more than 10% in 3 days in its entire history, and the other two were a 12.8% drop on October 28th, 1929 and an 11.7% fall the next day, October 29th, 1929.  The 1929 days marked the Great Crash that led to the Great Depression, and the 1987 single-day drop utterly dwarfed the worst down day in 1929 or indeed the entire history of the Dow.  Yikes!

 

If you think that is ugly, try this on for size.  That same fateful day in late October 1987, the venerable S&P 500, which obviously contains 500 elite American companies and not just 30 like the Dow, plunged 20.5%!  It was by far the biggest down day in the entire history of the S&P 500 too, dwarfing all others.  Number 2, an 8.3% plunge, happened a week later on October 26th, 1987.  Number 3 followed exactly a decade later, a 6.9% plunge on October 27th, 1997.

 

The S&P 100 off of which the VIX implied volatility index is computed contains the biggest and best 100 companies of the S&P 500.  The unreal VIX 150 spike witnessed on October 19th, 1987 occurred on the biggest single down day by far in the entire history of the financial markets in the United States of America.

 

Could the VIX soar to 150 again?  Sure, anything is possible.  Is it likely that the VIX will soar to 150 again?  Absolutely not, as the US markets would have to crash and again plummet by 20%+ in a single trading day to create the volatility super-maelstrom necessary to make it so.

 

In the entire history of the Dow 30 there have only been 30 days where the index has fallen by 6% or more in a single trading session.  In the entire history of the S&P 500 there have only been 8 days where the index has plunged by 6% or more in a single trading day.  In light of the historical data, the probability at almost any given time of a general US equity market crash and a VIX super-spike is exceedingly low, almost not even worth considering.

 

Investing and speculation is all one giant probabilities game.  Just like in life, there are no certainties in the markets.  The NASDAQ could rocket back to 5,000 tomorrow.  It probably won’t happen, but it always could happen theoretically.  When you get in your car to drive to work, there is a small chance you will perish trapped in the twisted metal and searing gasoline fires of a grisly car wreck.  It probably won’t happen, but the risk is always there each time you get behind the wheel. 

 

Market crashes are similar, exceedingly rare events that can theoretically happen at anytime per probability theory, but they probably won’t at any given time.  The prudent investor and speculator carefully analyze the probability of market events happening and bet on the highest probability outcome, not some tremendously rare event pushed so far out on to the bell-curve extremes that it is hardly noticeable in a dataset.

 

The stock super-bears today advocating theories of another crash in US equities are way out on an obscure probability limb that probably can’t support their weight.  Yes, the US indices seem destined to migrate far lower than current levels to reach undervalued status before the bust runs its course, but that doesn’t imply the mean reversion has to transpire in a day or two.  If history is a valid guide, it will take many more months and maybe years for the bust to finish its job of mercilessly destroying bubble speculative excesses.

 

Speaking of history, this greatest of market professors can also offer us much insight into the timing of crashes, or extreme single down days.  For discussion purposes, let’s arbitrarily define a crash as a 10%+ loss in the level of a major equity index in one single trading day.  A 10%+ single-day loss has only occurred on 3 days in the entire history of the Dow 30 and 1 day in the entire history of the S&P 500!  Crashes are really, really rare events in history!

 

Neither I nor any of my partners have ever heard of a crash occurring at any time other than right near the top of a major bubble.  Crashes, while exceedingly rare, always occur close to bubble tops to the best of our knowledge, not years after the bubble crest well into the normal bust process.

 

If you know of even a single example in world history of a major equity crash occurring years into the bust rather than right at the very bubble top leading into the bust, please drop me a note and let me know.  If we can track down the raw data we will create some graphs and I will write an essay on it.  In all my studies I have never seen an example of a late-bust crash occurring in any major equity index.

 

While we were on this whole VIX-150-inspired research adventure, we decided to create some charts of a couple bubble bursts and busts in history, one classic and one destined to be classic, and look at the distribution of extreme down days and outright crashes.  We were hoping that visually seeing the distribution of the big down days would offer some valuable insights into what to expect as busts continue to evolve toward an ultimate undervalued bottoming level.

 

Our first graph shows the distribution of big down days in the ultimate historical bubble bust, 1929 and the following years in the States.  While the graph shows data both before the bubble top and after the bust bottom, we chose the precise dataset between the ultimate top and bottom to actually analyze. 

 

Between the Dow 30 top of 381.2 on September 3rd, 1929 and the Dow 30 bottom of 41.2 on July 8th, 1932, we took the 20 biggest down days in this actual bust period and marked them on the graph.  The distribution of the big down days, only 2 of which qualified as honest-to-goodness crashes per our 10%+ single-day plunge definition, was quite provocative.

 

 

Interestingly, the big down days were clustered in two regions in the classic 1929 bust episode.  The initial crash event, readily apparent in the chart above, had 7 of the Top 20 down days, including the ruling triumvirate of the Top 3.  This first cluster of big down days, including the outlying #10 in 1930, had an average single-day loss of 7.8%, the yellow number above.

 

Somewhat surprisingly, the rest of the Top 20 down days during the 1929-1932 bust clustered tightly in the final 10 months or so of the supercycle bust process.  These big down days averaged a 5.6% loss, and they seemed to increase in frequency close to the ultimate bottom, probably reflecting growing general fear as the markets excruciatingly kept falling and falling, with no end in sight. 

 

These two swarms of big down days spread apart look like a giant barbell on the chart.  There was a long parched desert of relentless downward action in the middle of the bust, marked with the red arrow above, but this ongoing shredding of investors occurred without any big down days.  This unimpressive bear-market action, unmarked by catastrophic daily falls, is typical.  The lack of big down days helps the Great Bear keep raw fear minimized and seduces the perma-bulls to stay in as long as possible so they lose the maximum amount of capital.

 

While 11 of the Top 20 1929 bust down days did occur late in the brutal bear market, there were none approaching the 10% crash magnitude.  The majority of these late-bust single-day drops were between 5% and 6%, which are nowhere near as rare as the monster 10% crash days.  There have been 66 5%+ down days in the entire history of the Dow 30 and 11 in the S&P 500’s history.

 

The closest big down day to the ultimate 1932 bottom, #16 above, only managed a 5.2% loss, certainly not crash material at the end of a bust!

 

So what are the lessons echoing down through the sands of time since 1929?  In bubble busts, the crash days happen up front, soon after the top, not years into the bust process.

 

On an interesting historical sidenote, the crash days do not plunge immediately from a supercycle top, but generally occur 5-8 weeks after the pinnacle of each bubble.  In 1929 the first crash day occurred about 8 weeks after the top, in the Dow 30 in 1987 about 8 weeks, and in the NASDAQ of 2000 about 5 weeks.  It takes a little time for bubble euphoria to collapse into the immense short-term fear necessary to spawn a crash day.  Nevertheless, crashes are early-stage bust events soon after bubble tops, not late-stage bust events years into the bottoming evolution.

 

Our next graph is constructed based on the same methodology and applies to the ongoing NASDAQ bust.  We took the dataset for our Top 20 analysis from the ultimate NASDAQ top of 5048.62 on March 10th, 2000 until September 18th, 2002, the data cutoff date for this essay.  Once again the results are provocative.

 

 

The ongoing NASDAQ bust has also witnessed two clusters of big down days.  The first occurred during the initial crash event and contains the Top 2 days, including one 10%+ monster of crash magnitude.  This first swarm of big down days averaged 7.5%, not too far away from the 7.8% seen in the Dow 30 in its initial 1929 crash episode. 

 

It never ceases to amaze just how well market history seems to repeat itself!  With the very same greed and fear fortified deep within human hearts driving markets today as in 1929, perhaps it shouldn’t be surprising how well history rhymes.

 

The second swarm of big down days occurred soon after the initial post-crash rally failed and the NASDAQ plunged like a rock.  The majority of these big down days were only between 5% and 6%, however.

 

Note that almost all the Top 20 down days so far in the NASDAQ bust occurred between the initial crash event and the first major bear market rally of early 2001.  As the red arrow above indicates, the right side of this graph is extremely lonely without a single big down day in over a year.  Just as in the 1929 bust, it would not surprise me one bit if we see a new cluster of big down days bloom closer to the ultimate NASDAQ bottom in the future, forming another barbell-shaped big-down-day distribution across time.

 

The only outlier occurred on September 17th, 2001, which marked extraordinary events.

 

Monday September 17th, 2001 was the first day the US equity markets were open after the radical Islamist attacks on the interventionist US federal government.  While 9/11 probably wasn’t the worst terrorist atrocity in world history, it was certainly the most spectacular.  Thanks to the miracles of the Information Age, the whole world watched live on television as the twin towers horrifically imploded into smoldering rubble.  In light of this enormous historic discontinuity, it is not surprising that big down day #6 occurred this far out into the bust process.

 

Now if even 9/11 couldn’t provoke a 10% crash day, and 9/11 couldn’t push the VIX higher than 49 on a closing basis, what on earth could?

 

Last year I had a bunch of open put-options positions that had huge profits in late September after the events of 9/11.  At the time I made the wrong decision and chose not to sell them because the events of 9/11 were so earth-shattering that I figured the probability was high of a vicious slide in the equity markets.  I was wrong and the markets soon launched a huge bear-market-rally assault off of an extreme VIX reading near 50.

 

The super-bears arguing today for another crash in US equities and a VIX 150 approaching should remember 9/11 and also realize that historic probabilities are vastly opposed to such arguments.  Yes, a crash is always possible, but only probable within a month or two of stellar supercycle bubble tops.  Crashes don’t happen years into supercycle busts!

 

The smart money today is indeed gambling on the ongoing bust dragging US equity indices much, much lower.  At current earnings the S&P 500 will bottom under 500, the Dow 30 under 5000, and the NASDAQ under 400.  Yet, just as in history, this ongoing bust process will be slow and painful over many more months or years, exquisitely flaying the last pound of bloody flesh from all the hopeless bulls who still desperately believe the Tech Bubble of 2000 will return in their lifetimes.

 

Don’t get caught up in the popular crash hype that flies in the face of market history and probabilities.  The markets are heading lower before this is all over, but the exceedingly sadistic and vicious Great Bear is in no hurry to end his hellish orgy of carnage.

 

Adam Hamilton, CPA     September 20, 2002     Subscribe