The Physics of Wall Street: a brief History of Predicting the Unpredictable

Epilogue: Send Physics, Math, and Money! 
•
221
has grown, banks have had a strong incentive to come up with other 
assets that they could use as collateral.
corporate bonds, which are just like government bonds only issued 
by a corporation, aren’t a very good choice because their value tends to 
be connected with the corporation’s stock prices. no one would want 
collateral that could be highly volatile, or worse, collateral whose value 
you could try to “game” by looking at how the stock prices are chang-
ing. So firms participating in this shadow banking sector wanted to 
come up with some new kind of asset that worked like a bond, but 
whose value didn’t depend on something that it was easy to get in-
formation about. the solution they happened on was consumer debt
— mortgages, student loans, credit card debt. now, consumer debt by 
itself isn’t a great choice, because it’s possible to predict from a par-
ticular individual’s history whether that person is likely to default. So 
instead of using loans as collateral directly, banks took the consumer 
loans and “securitized” them. this involved combining a large number 
of loans into a pool, and then slicing the pool up into pieces and selling 
the pieces as bonds. these new assets — which included cdos — were 
designed to work just like government bonds (though they were much 
riskier). they bore interest so that when firms deposited money with 
one another, that money wouldn’t lose value.
the quant crisis was the first signal that all was not well in this 
shadow banking system. the whole system was built on the assump-
tion that U.S. housing markets wouldn’t decline. When they did, be-
ginning in 2006, the system began to crumble, and when the decline 
accelerated in 2007, panic set in. defaults occurred, mainly among 
homeowners who were already perceived as high risk, the beneficia-
ries of so-called subprime mortgages. this sudden high default rate, 
in turn, made the securities based on subprime mortgages rapidly lose 
value, as no one was sure whether the promised interest rates would 
be paid. the quant crisis resulted when a small handful of hedge funds 
were told they needed to put up more collateral for the loans they used 
to finance their investment, which in turn meant they needed to sell 
quickly to raise cash. Most of the quant funds used similar methods, 
which meant they often had very similar portfolios — so that when 

---

222 
•
t h e p h y s i c s o f wa l l s t r e e t
one fund started to liquidate, it pushed all of their holdings, includ-
ing the ones that were supposed to act as insurance, lower. this rapid 
and unexpected loss then forced other funds to sell, too, leading to a 
vicious cycle where everyone involved lost a lot of money. (this is a 
perfect example of how Sornette’s herding effects can lead to crashes.)
the quant crisis, and its reverberations later in 2007, were just the 
beginning. the next casualty was the eighty-five-year-old investment 
bank Bear Stearns, in March 2008. Bear Stearns had been a major 
player in the shadow banking system, producing many of the securi-
tized loans that served as collateral. When the underlying mortgages 
started to see ever-higher default rates, Bear Stearns’s depositors 
started to get edgy. Starting in the middle of the month, some of Bear 
Stearns’s biggest customers asked for their money back at the same 
time. first was renaissance, James Simons’s firm, which wanted its $5 
billion. Another $5 billion was pulled out by another hedge fund, d. e. 
Shaw. Soon it was a classic run on the bank, with all of the custom-
ers clamoring for their cash. to stem the bleeding, Bear Stearns was 
forced to agree to a government-backed takeover by another invest-
ment bank, J. P. Morgan.
the crisis was just beginning to pick up steam. the real climax oc-
curred at the end of that summer, when Lehman Brothers, another 
eminent old investment bank, collapsed. this time, the government 
didn’t step in to negotiate a bailout, which only increased the sense of 
panic. Within just a few days in September, another struggling invest-
ment bank, Merrill Lynch, was eaten up by Bank of America. the in-
surance firm AIG was on the verge of collapse. no banks were willing 
to lend money, least of all to other banks, whose fortunes were far from 
certain. the entire shadow banking system froze up, and the financial 
market collapsed beneath the pressure. By october, 40% of the value 
of the U.S. stock market had vanished into thin air.
Surely, the misuse of mathematical models played a role in this cri-
sis. the securitization procedure by which subprime mortgages were 
turned into new products that behaved like bonds was based on a 
model developed by a statistician named david X. Li. Li’s model had 
a fundamental flaw: it essentially assumed that default on one mort-
gage wouldn’t change the risk of default on other mortgages. As long as 

---