THE BEAR'S
LAIR Math,
leverage and risk By Martin
Hutchinson
The US$2 billion (or maybe $5
billion) loss by JPMorgan Chase went unnoticed for
several months, because the bank was relying on
the value-at-risk risk management metric, which as
we pointed out in "Alchemists
of Loss" is hopelessly flawed. [1]
However there is a cultural question here:
large conservative banks 30 years ago would never
have relied on a flawed mathematical algorithm to
determine their risk position, and would in any
case have been far less leveraged. The combination
of monetary policy as pursued by Federal Reserve
chairman Ben Bernanke monetary policy, modern
ideas about leverage and flawed mathematical
concepts seems uniquely toxic. Is the financial
system to be bedeviled by it from here on forward?
It did not take the 2008 crisis to show
Kevin Dowd and me that
value-at-risk (VaR),
invented in the early 1990s, was utterly unsound.
Indeed its flaws were well known for several years
before the crash - the great mathematician Benoit
Mandelbrot, in his 2004 The Misbehavior of
Markets, had pointed them out with
mathematical elegance we could not hope to match
(Mandelbrot had pointed out flaws in the emerging
underlying theory as early as 1962).
Both
Kevin and I had also written on the subject, with
my own first essay in the area being published in
"Derivatives Quarterly" in early 1998, before the
Long-Term Capital Management debacle had first
demonstrated the fallibility of Wall Street's
models to all who were willing to examine its
denouement impartially.
Naively, we
assumed if nothing else came out of the 2008
debacle, it would at least lead to banks ceasing
to use such an obviously flawed methodology, and
moving to something more satisfactory. However,
when details came out of JPMorgan Chase's losses,
it became clear this had not happened. Indeed, so
engrained is VaR that the new Basel III regulatory
system still contemplates its use, although to be
fair regulators have now issued a discussion paper
suggesting that alternatives might be sought.
The problem is not one of inadequate
mathematicians. The Mandelbrot book is highly
readable, and generally mathematically accessible.
Furthermore, the problem with VaR, its inadequate
recognition of "fat tails" and "long tails", is
obvious to anyone observing its behavior in a
crisis.
When David Viniar, the chief
financial officer of Goldman Sachs, said in 2007
he was seeing "25-standard-deviation moves,
several days in a row", he would have needed only
a minute or so with Microsoft Excel to see that
the existence of such moves wholly invalidated the
VaR methodology.
The problem was the
incentives on Wall Street. Mathematicians were
brought into the major investment banks only in
the mid-1980s. Seduced by the remuneration
available, far in excess of what they could
conceivably get in academia, they became slaves to
the more strong-willed traders and management of
the banks.
Told that their jobs depended
on managing risk in such a way as to allow the
traders free rein, they devised mathematical
models that appeared legitimate and could be
presented to inquisitive top management and
regulators, but in reality were fatally flawed.
The traders also responded to incentives.
Told that their bonus required them to maximize
profit while remaining within the risk management
parameters produced by the mathematicians, they
found ways of designing products that satisfied
the fallible models, but nevertheless allowed them
to take larger risks than top management
contemplated.
Credit default swaps, the
loss on which could be 100 times the premium paid
for them, had "tails" far longer than the VaR
models contemplated, and hence risks far greater
than the models predicted. Collateralized debt
obligations, especially the multi-layered
"CDO-squared" and CDOs on subprime credits, turned
out to be far more internally correlated than the
VaR models predicted.
In other words,
while the "tail" losses on CDOs were of the same
order as VaR predicted, the "tails" were much
fatter - the probability of those losses occurring
was far higher.
Because of the
pathological nature of their risk profiles, CDOs
and CDS were extremely attractive to traders, who
found they could generate far more profit from
these very risky instruments than was possible
from conventional instruments, the "spreads" on
which were arbitraged away by fierce competition.
The poison of misguided incentives also
extended to the top executive suite, by the malign
effect of leverage. Traditionally, leverage in
banking had been moderate. While long-term
interest rates were generally higher than
short-term rates, occasional credit crises kept
bankers honest, convincing them that it was
foolish to leverage up on long-term paper and
borrow in the short-term market.
The 1980
bankruptcy of First Pennsylvania Bank, which found
its capital wiped out by price declines in
government bonds, showed the folly of relying too
heavily on the positive "spread" between
short-term and long-term rates. Generally, while
leverage enhanced the returns bankers made, it did
not turn an otherwise unprofitable business into a
bonanza because borrowing was expensive and
reserves needed to be kept to face credit crises.
Salomon Brothers and other trading houses
showed in the 1980s that leverage could be used to
produce outsize returns. It is thus significant
that the first major bankruptcy of the modern era,
Long-Term Capital Management, was engineered by
former Salomon traders who had juiced up
apparently ordinary returns (and risks) by the use
of unprecedented levels of leverage.
The
bailout of LTCM provided all the wrong lessons to
the market. Just at the time when the market
needed a salutary lesson of greed given its just
deserts of bankruptcy, ideally accompanied by a
term in debtors' prison, it got the opposite
lesson, an entity of exceptional foolishness and
greed being bailed out because it was deemed "too
big to fail".
Continental Illinois Bank in
1984 had likewise been deemed too big to fail
because of its lunatic oil patch lending, but
Continental Illinois was a major bank, seventh
largest in the United States, with tens of
thousands of employees. LTCM was a bunch of spivs
in a Connecticut office park; if it could be
rescued, then any big institution could expect the
same favor, provided its financial connectivity
was great enough.
Needless to say, the
monetary policies of Fed chairman Alan Greenspan
and Bernanke, his successor, played a major role
in the degradation of finance. From 1995, first
Greenspan and then Bernanke expanded money supply
at record rates, then after 2008 Bernanke kept
interest rates far below the rate of inflation.
Naturally, in this environment, excessive
leverage was encouraged. So also was bad behavior.
Long periods of easy money make the warnings of
the prudent seem old-fashioned, and the risks
taken by the Gordon Gekko-like "BSDs" on the
trading floor seem justified by the outsize
profits they produce for the institution and the
bonuses they receive.
If the 2008 crash
had been treated with the traditional Walter
Bagehot remedy of lending amply, but on good
security and at high rates, behavior patterns
would have been modified, as caution was once
again seen to have its virtues.
As it was,
the combination of a bailout that was almost
cost-free to the banks concerned and an
intensification of the easy-money policies that
had caused the problem produced the inevitable
behavior of yet more leverage, yet larger trading
risks, all suitably hidden by the spurious VaR
system.
One might well ask where the
regulators were while all this was going on. The
abolition of Glass-Steagall restrictions in the US
in 1999, while sensible in principle, in practice
allowed the biggest banks to take massive
advantage of easy money conditions, gambling with
the proceeds of deposits backed by the Federal
Deposit Insurance Corporation deposits, in the
knowledge that on the Continental Illinois and
LTCM precedents, they would be bailed out if
necessary. The abolition in 2004 of leverage
restrictions on brokerage houses worsened the
problem.
However, the worst regulatory
dereliction of duty came from the Basel regulators
tasked with setting leverage and risk principles
for the banking industry. Instead of using a
common-sense definition of leverage based on
balance sheet totals, they allowed banks to
"weight" their assets, with government debt having
a zero risk weighting - thus allowing infinite
leverage in that sector.
Instead of
applying the strictest supervision to the
mushrooming trading desks, by which such risks
were being incurred, they allowed the banks to
devise a spurious risk management metric that hid
the most serious risks, then permitted the banks
using this metric to describe themselves as
"sophisticated" thereby allowing them to escape
much of the regulation imposed on lesser mortals.
Not only were the Basel regulators
"captured" by the industry, they have suffered a
"Stockholm syndrome" by which they have adopted
the attitudes and practices of their captors.
(This syndrome also seems to have affected the US
Securities and Exchange Commission, which is
currently trying to use the tools of the
Dodd-Frank legislation, supposed to rein in the
banking industry, to put out of business the money
market funds, useful bank-hated consumer-friendly
entities that were entirely blameless in the 2008
debacle.)
No serious attempt has been
made, nearly four years after the 2008 collapse,
to replace the discredited VaR system with
something that works properly. What's more, while
the Basel regulators have now included a "common
sense" un-weighted leverage limit as well as the
spurious weighted limits, their required capital
is a pathetic 3%, allowing leverage of 33 to 1,
higher than that which led Bear Stearns and Lehman
Brothers to their doom.
Comfort
yourselves: if something cannot go on for ever, it
will end. The current rickety edifice of
international finance, with reckless monetary
policies, excessive leverage, spurious risk
management and Stockholm Syndrome regulators, will
cause a series of crashes, each one more expensive
than the last.
One had hoped a serious
crisis such as that of 2008 would have brought
reform as well as recession. It now appears that
we will need half a dozen such global financial
crises, each doubtless nastier than the previous
one, before common sense returns. But return it
will!
Martin Hutchinson is the
author of Great Conservatives (Academica
Press, 2005) - details can be found on the website
www.greatconservatives.com - and co-author with
Professor Kevin Dowd of Alchemists of Loss
(Wiley, 2010). Both are now available on
Amazon.com, Great Conservatives only in a
Kindle edition, Alchemists of Loss in both
Kindle and print editions.
(Republished
with permission from PrudentBear.com.
Copyright 2005-12 David W Tice &
Associates.)
Head
Office: Unit B, 16/F, Li Dong Building, No. 9 Li Yuen Street East,
Central, Hong Kong Thailand Bureau:
11/13 Petchkasem Road, Hua Hin, Prachuab Kirikhan, Thailand 77110
