What are the mechanics behind a market crash?

[Groty]

Charles Kindleberger's "Manias, Panics, and Crashes" documents the history and causes of manias, panics and crashes beginning with the Dutch tulip mania. It's a good read. I'll try to summarize his theory:

According to Kindleberger, manias are always a function of a prolonged period of loose credit conditions. Loose credit leads to speculation. Towards the end of a stock market mania, people are generally not buying securities because they believe they represent good value. They are buying because they see stocks only going up and they believe they can sell to someone else at a higher price (kind of a greater fool theory). The underlying reason for the mania is usually a combination of an extended period of loose credit together with a belief a "new era" has emerged (examples: industrial revolution in the '20s, tech revolution in the late '90s; potentially the emerging market industrialization today). Everyone is excited by the prospects of the "new era". And the easy credit makes it easy to buy. So buying begets buying with little regard to risk or value.

It ultimately leads to a final, speculative, blowoff phase in which margin debt is often very high (interestingly, I'd note it's currently at record levels).

Then some catalyst causes an intial wave of selling. Often its a perception that credit conditions will tighten and restrict liquidity. Because many market participants are leveraged, the initial selling wave triggers margin calls. The margin calls result in more selling. So the selling feeds on itself as people attempt to preserve gains and front run the guys who will be forced to sell to make margin calls.

[DesertCat]

[ QUOTE ]

It ultimately leads to a final, speculative, blowoff phase in which margin debt is often very high (interestingly, I'd note it's currently at record levels).

[/ QUOTE ]

This article explains why it probably isn't. Essentially net margin debt is much lower than it appears because cash balances are much higher then they were during the bubble.

One other point is the S&P 500 is worth significantly more now than it was in 2000. Not based on market capitalization, but on earnings power. Earnings have grown but PE levels have been cut in half. I.e. the S&P's price stays roughly the same, but it generates significantly higher cash flow. This makes the risk of margin debt much higher in 2000 than today because the market had much more downside.

Not saying I'm recommending anyone get levered up, just don't get too scared about todays margin levels.

[DcifrThs]

[ QUOTE ]
[ QUOTE ]

IIRC, something that is reflexive, according to soros, feeds back on itself.


[/ QUOTE ]

Another example of this is the crash of 1987. This article is the best I could find to explain it and it's just so-so. Essentially my understanding is that if the market declines, portfolio insurance would involve writing puts against the the S&P 500. The put buyers would sell S&P 500 shares to hedge their exposure, driving the S&P 500 price down further, forcing the automated portfolio insurance systems to write more puts and buyers to sell more shares, etc. driving the market down in an enormous spiral of destruction.

[/ QUOTE ]

this is close but not exact.

during the run up to the black monday (oct 1987), the newfangled option mathematics wizzes looked at the formulas and said "hey, we can replicate put options without paying the premium."

basically, they were saying that by combining a given # of shares and risk free instruments, they could copy the payoff profile of a put option. so, they sold their systems that traded this as "portfolio insurance." meaning, they were sold at a "discount" to the cost of a put option.

the problem with this methodology is that when the market moves sideways (creating tons of transaction costs) or when the market plumments quickly (creating an inability to sell shares at exact prices as the price falls in big jumps), the actual put option replication falls apart. that is to say, it doesn't work anymore and EXACTLY when you need it to, the "portfolio insurance" you purchased doesn't deliver.

so, in the end, the port. insurance sellers were blamed for the crash when in fact they likely on had a small, if any, exacerbating effect on it.

we had this debate back and forth among some of the sharpest people i've met to date (strategists who ran goldman's commodity group, folks who directed risk management for years and oneo f the co CIO's of my old fund) and the overall conclusion was that the causes of the 1987 crash were numerous and that portfolio insurance doesn't work when you need it too (i.e. they couldn't even agree that portfolio insurance was definitely a cause...or even a serious exacerbating factor).

hope that clears it up a bit.

Barron

[icetonez]

Do stop-loss orders always work in a meltdown?

[DcifrThs]

[ QUOTE ]
Do stop-loss orders always work in a meltdown?

[/ QUOTE ]

no. they will get you sold, but not at the price you desired.

for instance, if you stop-loss at $100 and the security is at $102 while the market all of a sudden tanks, the security may suddenly drop by 5 points to $97 which would instantly trigger the stop loss. however, you wouldn't be able to sell at $100 since nobody would be buying at that price (b/c they could buy at $97), so you'd at most get to sell at $97 if not lower

Barron

[kimchi]

[ QUOTE ]
Do stop-loss orders always work in a meltdown?

[/ QUOTE ]

The stop-loss order still needs a buyer in a falling market and a seller in a rising market. Your order won't be filled until one can be found.

The futures markets can also experience limit-up and limit-down days when traders can become trapped in a fast-moving market.

You can use guaranteed stop-losses which trade with an increased spread. These are usually only available with certain derivatives such as CFDs.

Some derivatives don't technically trade on a market where buyers and sellers meet. They are 'fake' markets offered by some brokers. Your stop-loss fills with these can be manipulated, sometimes very unfavourably, by your broker. This is common with spread betting and certain CFD providers.

[eastbay]

[ QUOTE ]
Excellent summary Barron. The Soros book is "The Alchemy of Finance" (1988) He provides several examples of boom/bust processes. FWIW, Soros predicts a real estate crash in the USA in his last book "The Age of Fallibility: Consequences of the War on Terror" (2006)


[/ QUOTE ]

later, Soros wrote that in most situations his theory "is so feeble that it can be safely ignored."

http://www.gladwell.com/2002/2002_04_29_a_blowingup.htm

I tried to read the "Alchemy of Finance" book and couldn't find much if any there there.

eastbay

[APXG]

[ QUOTE ]
[ QUOTE ]
Excellent summary Barron. The Soros book is "The Alchemy of Finance" (1988) He provides several examples of boom/bust processes. FWIW, Soros predicts a real estate crash in the USA in his last book "The Age of Fallibility: Consequences of the War on Terror" (2006)


[/ QUOTE ]

later, Soros wrote that in most situations his theory "is so feeble that it can be safely ignored."

http://www.gladwell.com/2002/2002_04_29_a_blowingup.htm

I tried to read the "Alchemy of Finance" book and couldn't find much if any there there.

eastbay

[/ QUOTE ]

Soros unlike most of his counterparts tends to overstate own flaws in everything he says. As he mentioned once when asked about the past 40 years, the secret to his success is that he is always wrong.

[pig4bill]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]

IIRC, something that is reflexive, according to soros, feeds back on itself.


[/ QUOTE ]

Another example of this is the crash of 1987. This article is the best I could find to explain it and it's just so-so. Essentially my understanding is that if the market declines, portfolio insurance would involve writing puts against the the S&P 500. The put buyers would sell S&P 500 shares to hedge their exposure, driving the S&P 500 price down further, forcing the automated portfolio insurance systems to write more puts and buyers to sell more shares, etc. driving the market down in an enormous spiral of destruction.

[/ QUOTE ]

this is close but not exact.

during the run up to the black monday (oct 1987), the newfangled option mathematics wizzes looked at the formulas and said "hey, we can replicate put options without paying the premium."

basically, they were saying that by combining a given # of shares and risk free instruments, they could copy the payoff profile of a put option. so, they sold their systems that traded this as "portfolio insurance." meaning, they were sold at a "discount" to the cost of a put option.

the problem with this methodology is that when the market moves sideways (creating tons of transaction costs) or when the market plumments quickly (creating an inability to sell shares at exact prices as the price falls in big jumps), the actual put option replication falls apart. that is to say, it doesn't work anymore and EXACTLY when you need it to, the "portfolio insurance" you purchased doesn't deliver.

so, in the end, the port. insurance sellers were blamed for the crash when in fact they likely on had a small, if any, exacerbating effect on it.

we had this debate back and forth among some of the sharpest people i've met to date (strategists who ran goldman's commodity group, folks who directed risk management for years and oneo f the co CIO's of my old fund) and the overall conclusion was that the causes of the 1987 crash were numerous and that portfolio insurance doesn't work when you need it too (i.e. they couldn't even agree that portfolio insurance was definitely a cause...or even a serious exacerbating factor).

hope that clears it up a bit.

Barron

[/ QUOTE ]

I'm not sure you can even call 1987 a crash. The market still finished up for the year. It was not much of a surprise either. The market had been showing weakness and a number of analysts had been calling for a tank.

The interesting thing about it is there would have been bigger moves that there were if people had the communications infrastructure we have now. The "tape" was 45 minutes behind the market on the big down day, so people weren't really sure how far the market had dropped. Then in the following days, I was trying to buy with both hands but I could only get through to my local broker. All 800 lines to national brokers were jammed, and when you did get through you were put on hold. I sat on hold with Fidelity for 3 hours, then gave up.

[deluz35]

Soros is not a great writer, but he does explain the boom/bust process in "The Alchemy of Finance". The epilogue is on the Crash of 1987. The idea of reflexivity is that far-from-equilibrium conditions can be created when participants do not adequately recognize that fundamentals are correlated to their own perceptions (e.g. credit ratings being calculated by debt ratios, which were affected by the banks' lending activity).

In most cases, the reflexivity effect is negligible. But if there is a sudden shift in sentiment or perception, in can result in a rapid reversal of supply/demand and a crash.

It usually seems to result from too many people chasing easy money in a speculative cycle (e.g. Nasdaq bubble 99-00). Too much leverage is applied in one direction until it is exhausted. Whether the reaction becomes a crash or a more gradual reversal depends on the insight and psychology the participants. If the brakes are applied too suddenly on a car going too fast, we get whiplash.