Full text of Financial Crisis Inquiry Commission : Documents Relating to the Financial Crisis of 2007-2009 : Transcript of Interview With Satyajit Das, May 7, 2010

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

FCIC Interview of Satyajit Das, May 7, 2010

United States of America
Financial Crisis Inquiry Commission

INTERVIEW OF
SATYAJIT DAS

Friday, May 7, 2010

*** Confidential ***

1

FCIC Interview of Satyajit Das, May 7, 2010
Financial Crisis Inquiry Commission
Friday, May 7, 2010
--o0o--

MR. SEEFER:

I just walked in the room, so I

assume they haven’t given you the usual introductions.
So let me do that.
We are all with the FCIC, as I’m sure you
know.

And as you may know, we have been tasked by a

statute passed last year to figure out the causes of the
financial crisis.

As part of that, we are looking into

the role of -- that derivatives may have played in
either causing the financial crisis or acting as a
propagating mechanism.

And, therefore, we are talking

to a lot of experts; and you were, of course, on our
list for that.
Another thing I should tell you is, we are
look -- and this is confidential, so please keep it
confidential, sir -- we are looking at having a hearing
in early June on the subject of derivatives and their
role in the financial crisis.

And you have to be

there -- no, I’m just kidding.
But, actually, if you were interested in
coming out, it would be great.

But, of course, we’re

not going to ask you to travel 17 hours to the United
States for that.
2

FCIC Interview of Satyajit Das, May 7, 2010
MR. DAS:

Okay.

MR. SEEFER:

But my overall question is really

just that in the beginning, and then I’ll let these
researchers that have Ph.D.s, that know this stuff much
better than I do -- I’m just a lawyer here.
If you could just really, in the first
instance, tell us what you think the role of derivatives
were in the financial crisis, whether as a cause, a
contributing cause, or as a propagating mechanism.
MR. DAS:

The first thing I would say is, I

love the word “expert,” because Niels Bohr once defined
it as somebody who has made every error possible in his
field.

So I think on that basis, I qualify.
Now, I think the first thing that I would say

is, you have to look at derivatives as almost two sets
of instruments.

There is an element of derivatives

which essentially used managed risk, which assumes an
underlying position, whether it be credit, interest
rates, or any other asset class.
The second is, you can, of course, use
derivatives to create leverage.
Now, what is also very misunderstood about
derivatives where it’s used to create leverage is, there
are two types of leverage.

There is what I

call “explicit leverage,” which is basically where you
3

FCIC Interview of Satyajit Das, May 7, 2010
effectively actually use the –- actually outlay cash to
buy something, like in a futures contract -- or there’s
a variety of transactions you can do.
However, there is a much, much nastier form of
derivative leverage which nobody really understands,
which is called “embedded leverage.”

And I don’t -- I

can’t remember whether they’re in the papers I sent
you -- I sent you stuff on embedded leverage.
MS. SHAFER:
MR. DAS:

I don’t believe so.

No, I probably forgot.

MS. SHAFER:

I would love to read that.

Please -MR. DAS:

I certainly will send that to you.

MS. SHAFER:
MR. DAS:

Thank you.

And essentially what that means is,

effectively, in most derivatives transactions, what you
do is, you disguise the borrowing to buy a larger
proportion of assets.
The other type of embedded leverage is
actually where you use an event and you scale the
losses.
Let me give you a very quick example.

For

instance, if you have a normal option on, say, the
S & P 500 and your strike price is, say, five -- say,
1,000.

If the S & P 500 goes to 1,010, you make
4

FCIC Interview of Satyajit Das, May 7, 2010
1 percent -- the difference between 1,010 and 1,000.
I’m assuming it’s a call option.
However, I can also structure that so that if
the S & P 500 goes above 1,000 -- say, it just goes to
1000.01, I get paid an amount of 10 percent.

And this

is known as a “digital option.”
The point I’m making is, derivatives create
leverage but not in obvious ways.

And just as an aside,

I would say this is why all the leverage covenants and
the leverage rules that central banks are putting in
place are going to be useless.
blunt with you:

Because I’ll be very

I can develop as many forms of leverage

as humanly possible which will get around the rules.
It’s not going to make any difference.
MR. STANTON:
MR. DAS:

Can I ask a question --

So I think -- in coming back -- yes,

of course, you can.
MR. STANTON:
detail?

Can you explain that in more

I don’t see the leverage part of my buying an

option on the S & P 500.

And I apologize for that, but

you’ve got to do some remedial education here.
MR. DAS:

No, no, that’s fine.

That’s no

problem.
Okay, let’s say -- let’s say I have a million
dollars and I want to buy the S & P 500, right?

So I
5

FCIC Interview of Satyajit Das, May 7, 2010
have to outlay the entire million dollars and my
exposure to the S & P 500 is a million dollars, right?
MR. STANTON:
MR. DAS:

Yeah.

And it’s my money.

So if the S & P

500 goings to zero, I lose all of it, okay.
money was there in the first place.
wiped out.

But the

It’s just been

Which has incidents; but at the moment,

that’s what it is.
So instead of actually buying the $1 million
worth of share, I might be able to buy a call option to
buy the S & P 500 at a strike price, say, of 1,000.

But

the cost of that might only be 15 percent of
the million dollars that I have to outlay.
MR. STANTON:
MR. DAS:

Uh-huh.

So basically, I have the same

exposure to the S & P 500 that I would have, had I
bought all the shares for a million dollars.

But I

effectively now have only outlaid 15 percent of
the million dollars, $150,000.
But my returns are based on movements on
a million dollars.

So it’s like $150,000 gives me

exposure to $1 million of shares.

So it’s like a

6-to-1, 7-to-1 leverage.
Does that make sense?
MR. STANTON:

Yeah, it sure does.

Thanks.
6

FCIC Interview of Satyajit Das, May 7, 2010
MR. DAS:

Now, the next thing, of course, is

the leverage -- and this is the point of what I call the
embedded or amplified leverage, is normally when the
index goes from a 1,000 to 1,010, my 1 million would
have grown from basically 1 million to 1,010,000, right?
MR. STANTON:
MR. DAS:

Uh-huh.

But if I have this strange thing

called a “digital option” and it just goes up slightly,
my principal actually grows effectively from 1 million
to, say, 1,100,000, which is much more than it would
have.
And, see, people think of leverage as
borrowing.

It’s not.

It’s amplification of return.

So

you’ve got to [unintelligible] the end effect to work
back to how you create leverage.

At least that’s what

I’ve done in my work.
So coming back, so the risk management is
fine; but once you get into leverage, we can create any
form of leverage you like.

That’s the first problem.

And essentially in the crisis, what happened was we
created all sorts of this.

And if you’ve got probably

a year, I could sit you down in a room and take you
through each one because they’re certainly different.
Now, the next thing is -- that’s one.

But the

next one that creates problems is what people don’t seem
7

FCIC Interview of Satyajit Das, May 7, 2010
to have actually grasped with derivatives is, it layers
a whole lot of other risks on top of it.
And the simplest one is effectively what we
call counterparty risk.
Now, to go back to our option example,
everybody assumes this is a bet on the S & P 500.
it is and it isn’t.
that contract.

Well,

It’s also a bet on whoever sold you

Because you’re assuming that person is

going to be there to pay you that amount at the end of
the period.

And this becomes particularly important

where you are relying on that contract as a hedge.

And

this is where the, what I would call the problems of
derivatives get effectively very muddy.

Because on the

one side, you think you might be using a contract to
hedge, but you actually end up with a different set of
risks.
So the first risk you end up with is whoever
is the party you’re hedging with.

And the classic

example would be people like Citigroup and Merrill Lynch
who are hedging with these insurance companies, who they
thought would be there to make payments to them when
they needed to on their mortgage portfolios.

But,

unfortunately, they didn’t have enough capital.

And

under those circumstances, that hedge was not
worthwhile.
8

FCIC Interview of Satyajit Das, May 7, 2010
And that counterparty risk is inherent in
markets.

And as you will know, there are a whole bunch

of proposals now to deal with that after the horse and
trolley left the planet.
But effectively, under those circumstances,
that’s the second risk.

And it’s a very, very difficult

risk to manage.
And in one of the papers I sent you, I made
that comment -- and it’s very difficult to measure this.
Because let me explain why it’s difficult to measure.
Now, in most instruments, you know on day-one
what your exposure is.

If I lend money to you -- if I

lend you $100, I know that I could lose a $100.

But if

I enter into a derivative contract with you, it depends
on what happens to the underlying in terms of movement
as to how much you may owe me.

So I already know what

that is in advance.
Now, let me give you a very concrete example
of that.

When people hedged with AIG, they were hedging

what were called, effectively, super-senior pieces.
Now, the idea is in the super-senior pieces, that you
can never lose, or it’s as close to risk-free as we can
theoretically make something.
However, as it turned out, when mortgage
losses mounted, then our models were wrong.

So there
9

FCIC Interview of Satyajit Das, May 7, 2010
was rather a large exposure to AIG.

And under those

circumstances, AIG, if the contract had to be closed
out, would owe us a lot of money.
And that wasn’t anticipated in advance.

And

certainly, if you say that AIG knew that they had this
problem when they entered into this contract, that would
be wholly false.

And if you look at all the evidence

that’s emerged from AIG, they had no clue how bad that
exposure could potentially be.

Their models were

fundamentally wrong.
So the point is the counterparty risk exists,
but also the counterparty is very vulnerable to models.
That’s the second thing.
The third thing is, no hedge is perfect.

So

even if you’re hedging in terms of using a derivative to
hedge, then what happens is, if you look at the match
between the hedging instrument and the underlying, it’s
never exact.

It’s almost impossible to make it exact.

And so what happens is, particularly when
markets go kind of awry as they have, then what happens
is, these hedges start to have some very strange
behaviors.

So it can actually make you gain huge

amounts, but it can also make you lose huge amounts.
And underlying this, of course, is a theme
that you can see is, we become very model-dependent.
10

FCIC Interview of Satyajit Das, May 7, 2010
And the models are invented by people, generally
speaking, who have very limited understanding of how
markets work, and so they make assumptions which are
grossly unrealistic, and so these models break down very
quickly.

And that’s one of the problems that we have.
Now, the other thing I would, just for a

second, loop back to, because it’s related somewhat to
counterparty risk, is also because of the way the
financial structure of -- I’m sorry, this financial
system is structured, we create these massive
daisy-chains of risk.
So what happens is, you know, Warren Buffett
hedges with Goldman Sachs, Goldman Sachs then splits up
the risk into different little bits, hedges with
JPMorgan, Deutsche, and Barclays.

Barclays may hedge a

little bit of that with Deutsche Bank.

Deutsche Bank

may hedge that with, say, a Japanese bank.

And it goes

on and on and on.
And the best comparison I can give you is it’s
within a -- it’s like an electricity grid.

What happens

is, you have this massive thing between generators and
consumers.

But the point is, it flows through all sorts

of people, all sorts of junction boxes, all sorts of
elements.
The problem is, if any one of those fails, you
11

FCIC Interview of Satyajit Das, May 7, 2010
have a problem in the system.

And that is actually what

happened in September, October 2008, and that is
actually what is starting to happen again now.

And that

is very, very important to understand.
And there is some other sort of ancillary
issues here, and I’ll mention a couple of them.
One is liquidity risk.

And to understand the

liquidity risk, you need to sort of go back in history.
I’ve been around this business since 1977.
Now, when we started doing derivatives, generally, the
only counterparties we had were very highly rated
companies.

And by that, I mean, AA and AAA companies.

So this idea of counterparty risk became irrelevant.
Now, what happened in the inevitable nature of
finance is over time, we want to broaden the market.
The way we want to broaden the market, the
counterparties we were dealing with were less
creditworthy.
So what happened sometime in the eighties was,
we were looking for ways to allow these people to trade
with us but with minimal risk.
So under those circumstances, what we did was,
we came up with different ways to enhance the credit.
And the most common way to do that is actually to use
what we call “collateral,” which is bilateral
12

FCIC Interview of Satyajit Das, May 7, 2010
collateral.

And that works very simply, which is I deal

with you, I think you’re okay but not so okay.

So I

say, look, if the contract’s value changes and you would
owe me money, under those circumstances, you have to put
out some collateral, which historically has been cash or
government securities.

But over times, we’ve become a

little bit more adventurous and we used other types of
securities as well.
So under that circumstance, what happened was,
you have to give us this money.

Now, that, in a

theoretical sense, is fantastic because it sort of
safeguards your performance; but it creates liquidity
pressures in the system.

And this is what happened to

AIG.
AIG ended up owing -- depending on who you
believe -- between $14 billion and $18 billion.
Now, the interesting thing about this
collateral is, it’s not an actual loss because the
contract values obviously fluctuate over time.
But the point is -- I’m sorry, at a given
point in time, the issue is that if the contract was to
be closed out, at that point in time you would owe this
money.

So that’s the amount you have to put out.

And

that creates hugely volatile liquidity demands on
people.
13

FCIC Interview of Satyajit Das, May 7, 2010
And no matter how big a company you are, like
AIG, if somebody comes to you and asks you to cut a
check for $18 million -- I’m sorry, $18 billion -- it’s
not easy to reach into your back pocket and just hand
that over.
And as liquidity pressure has built up in the
system -- and this is where leverage gets attenuated
through the system.

So you have somebody dealing with

somebody who has these collateral provisions, but
they’re leveraged.

So basically what happens is, they

have to draw down certain lines which transmits the risk
to other people who are also leveraged.

It makes the

whole system quite fragile.
There are two or three other things here which
I would very briefly mention.

One is complexity and

built levels of knowledge.
In my belief, the system is extremely complex.
And also what is very, very clear, and it’s become
clearer to me over the last, probably, ten, 15 years, is
there’s probably less than 2,000 professionals in the
world who would pass a fairly rigorous test on how
derivatives work and how this thing works with the
financial system.
And given the number of players now in the
world, those 2,000 people are very, very thinly spread.
14

FCIC Interview of Satyajit Das, May 7, 2010
And the people with Ph.D.s in the room will forgive me
for saying this, because I deeply believe that, Joseph
Stiglitz in his recent book, “Freefall,” said something
which I literally had to wire my jaw back up when I read
this statement, he basically said that there’s no point
in asking professionals for advice on derivatives
because, in his view, the multi-lateral development
agencies like the world banks had basically specialists
who would be able to deal with that.

Generally, those

specialists are identified because they walk around with
white canes, dark glasses, and guide dogs.
And this is part of the problem that you’re
going to have, is a problem of expertise.

There is no

way you’re going to get a regulator who gets paid
$100,000 to be able to match wits with people who get
paid $10 million.

That’s just the nature of talent.

And, you know, as George Bernard Shaw once
said, “The most unfair thing in life is to assume that
people are equal.”

And that’s what our markets have

assumed.
And I’ll give you a signed bet on that, the
idea of a sophisticated investor, which is much in
demand -- or much in debate, rather -- with Goldman
Sachs is indictment.

Anybody who thinks IKB, the German

bank, is a sophisticated investor is delusional.

I have
15

FCIC Interview of Satyajit Das, May 7, 2010
dealt with IKB over the years.
other German Landes banks.

I’ve dealt with many

And, generally, when there

is a disaster in the world, there are four people you
always find at the zone of the disaster looking crippled
and deeply injured, and they start with the German
Landes banks, the Japanese banks; and usually, you find
Citigroup and Merrill Lynch somewhere nearby
MR. SEEFER:

We are going to bring you out

here.
MR. FELDBERG:
MR. DAS:

First-class.

Well, if you pay for me, I will

come.
But it is actually quite, quite fascinating to
see that.

And I know because I used to work at

Citigroup and I used to work at Merrill’s.

I know.

I

do understand what I’m saying because I used to sit
there and I used to shake my head as to what that
organization did.
But, anyway, so effectively, there’s a huge
issue of knowledge and complexity and asymmetry.
And the last element is effectively, when we
look at this, just on a knowledge point, is what has
happened over the last 15 or 20 years?
business has become industrialized.

The whole

By

“industrialized,” I mean, it’s become like a production
16

FCIC Interview of Satyajit Das, May 7, 2010
line.
Because when I started in derivatives, there
were kind of, sort of dilettante bespoke products, where
we used to sit there and rub our chin and look at each
other and make profound, sort of, statements about
certain things.
These days, it’s an industrial process.

And

what that’s done, is fragment the knowledge even
further.

The fragmentation of knowledge is actually

extremely relevant to the problems.

The right hand

doesn’t know what the left hand is doing, okay.
And I’ll give you a concrete example of that,
of how this works, which actually was very relevant in
this crisis.

As you know, many banks were writing these

poor-quality subprime mortgages.

If you actually look

at the poor-quality subprime mortgages, the people
writing them -- this is the people who were actually
approving the loans, et cetera -- were not stupid.

They

knew, effectively, that these mortgages were essentially
like hand grenades with the pins pulled out.

It wasn’t

a question of whether they were going to blow up, it was
just when.
So mostly they did it because they could then
repackage them and sell them to investors on the other
side.
17

FCIC Interview of Satyajit Das, May 7, 2010
Now, those banks at the one side, what I would
call the loan-origination and repackaging side were
acutely aware of the risks, and were very, very keen to
basically get rid of the mortgages off their balance
sheets.

And they, to varying degrees, succeeded.
Now, to give you some idea of how poor the

internal coordination in the banks is, those same
securities that were being created -- and these were
being put through rating agencies, models, and made into
AAA securities and so forth -- other parts of the bank
were doing two things to it, which are quite hilarious:
The first thing they were doing was, certain
parts of the bank, because it was AAA, were allowed to
invest in these.

So noted.

You know, one side of the bank is saying,
“This is just toxic nonsense.

Another part of the bank

is saying, “Well, this is now rated AAA.”

They don’t

actually look through to the underlying at all.

And

their risk managers are approving them, and allowing
them to hold tiny bits of capital.
MR. SEEFER:
MR. DAS:

Look at Citi.

Another part of the bank, which is

the prime brokerage, which is dealing with hedge funds,
is allowing hedge funds which own these AAA securities,
like the now deceased Bear Stearns funds and a whole
18

FCIC Interview of Satyajit Das, May 7, 2010
bunch of other funds, to borrow up to 98 percent of the
face value of these securities because they’re AAA,
without even concerning themselves -- in fact, they
didn’t know -- I know this for a fact -- they didn’t
know what the underlying securities were.
And this comes because the cultures in the
bank are so different and there’s a lack of
coordination.

And anybody who thinks banks are properly

managed, you know, can’t spell the word “management.”
And anybody who thinks that the directors are there
other than effectively to provide social airs and graces
are kind of missing the point somewhat, because there’s
just no way they can have the depth of knowledge that
you actually can have.
If I go and look at the AIG board, for
instance, there are several diplomats, there are -there’s, I think, an admiral from the Navy.

And unless

AIG was planning to wage some sort of war with nuclear
aircraft carriers, I’m not sure if they had the right
board for the business that they were in.
And there’s just no way.

I’ve sat on boards,

and I do understand what the problem is.
huge gap in knowledge.

There is a

So there’s a huge complexity

information asymmetry, management, directorial problem,
in this.
19

FCIC Interview of Satyajit Das, May 7, 2010
And the last thing I would say is what you
already know, is the incentive structures in the
business are such that it is palpably open to some
problems.

And -- but rather than bore you with

executive-compensation issues, which I’m sure you’ll get
tons of, I just want to point out some of the problems
in that whole area which people don’t often talk about,
one of which they don’t talk about is the difficulty of
marking these instruments to market.

And this is where

the use of mark-to-market accounting is basically
problematic.
Now, mark-to-market accounting was developed
by a bunch of accountants who basically didn’t have the
personality to become actuaries.
Now, essentially, they have almost no
understanding of how this mark-to-market is.

They have

this mythical idea that somewhere on a screen -- and
when they go into dealing rooms, they see the bank of
screens, and they immediately assume it’s a bit like the
tablet coming down from Mount Sinai carved in stone,
that’s what’s there is actually true.
And so basically, they assume these are
actually numbers.
Now, the fact of the matter is, particularly
in the OTC derivatives market, very few things can be
20

FCIC Interview of Satyajit Das, May 7, 2010
properly valued, because very few things trade with the
regularity that allows you to actually mark these things
to market.
Now, that means that most of it’s
mark-to-model.

And depending on the complexity, some of

the mark-to-models are okay; but there’s a whole bunch
of things which are mark-to-market, which are not really
a mark-to-market.

At best, they’re mark-to-model.

But

as I am fond of saying, and as I have said in print
several times, this is not mark-to-market, this is
mark-to-make-believe.

Because a very small change in

any of those assumptions can have material impact on the
valuation of those positions.
Now, the next thing about that is, there is
no way to verify this stuff, and it’s a very circular
process.

The traders create the models -- or, to be

very honest, they know what they want the model to say.
So they get the quantitative people to create the
models, which give them the answer they’re looking for
in the first place.
And then that model is used to calculate the
risk.

The regulators use the model to, say, calculate

the capital.

Then the credit people use that to

calculate the credit risk.

And then what they do is use

this to mark-to-market their books and show how much
21

FCIC Interview of Satyajit Das, May 7, 2010
money they’ve made.

It’s a completely circular process.

And this is not a process which can be changed
fundamentally because of the nature of these
instruments.

They just don’t trade.

And that’s the

fundamental problem that you have.
The second thing is, the mark-to-markets are
bad, which creates a problem.

But the other problem is,

when you’re PV’ing [phonetic] cash flows back to today,
to actually value these, there are certain hidden costs
down the track which are very, very imprecise and very,
very difficult to actually value.

And that became

hugely problematic.
So the executive compensation, people focus on
lots of things.
But Warren Buffett, at his Woodstock for
Capitalism this year, made the comment which is that he
disliked performance schemes, which were somewhat like
archery, where you fired the arrow; and wherever it
landed, you went and painted the target around, to make
sure the arrow had hit the bull’s-eye.
what most of this industry does.

Because that’s

That’s a huge problem.

Now, the other problem is, in terms of
compensation and so forth, is a somewhat different
issue, which is people seem to misunderstand why
derivatives and risk-taking became so important.
22

FCIC Interview of Satyajit Das, May 7, 2010
Now -- and I would just make the following
observation:

People don’t exist in vacuums, okay.

Bankers don’t create these products because they
essentially think them out.

There’s a great myth about

productivity and creativity in bankers.

Bankers are

among the most uncreative, dull people I have ever met
in my life.

Now, they couldn’t invent anything.

What it is driven by is forces from corporates
and from investors who are looking for ways to make
money.

And I think you have to understand that,

fundamentally, it became very, very difficult to make
money in real corporate activity.

Because one of the

side effects of globalization, which nobody really
focuses on, is that it absolutely crushed profit margins
for most people.

There’s a few people who it helped,

but they were few and far between.
And so what these people now do is they really
rely heavily on financial activity to make money.

And

this is why there were lots of problems with Asian
exporters with certain derivative structures which they
entered into, which weren’t really hedges at all but
were highly speculative transactions.

The reasons they

actually entered into that is, effectively, that there
is very -- there’s almost no way they can actually make
money without speculating.

And, in fact, ironically,
23

FCIC Interview of Satyajit Das, May 7, 2010
it’s even worse than that.

They can’t actually hedge

because the cost of hedging would be far greater than
their profit margin.
And investors on the other side, for a whole
bunch of reasons, effectively are chasing returns.

And

the returns they’re chasing are for two reasons -- and
we can sort of stratify the investor pool quite easily.
For instance, if you have a defined benefit
pension plan.

And we all know -- I mean, the U.S.

Social Security is a very good example, you’re
hopelessness underfunded.

So how do you make that up?

I mean, the logical person would say you put more money
into the fund.

But if you don’t have the money, well,

that’s going to affect your P&L.
Now, under those circumstances, what you do
is, you chase returns.
Individuals, as you know, in the United States
effectively are in particular 401(k)s.

If you actually

look at what they have to earn in terms of returns to
have a reasonable sum left for them at retirement, they
have to chase returns.
So all of those types of things create an
environment when risk-taking became paramount.

And in

my judgment, what happened with derivatives was just a
very elegant way to give them those risks, and that’s
24

FCIC Interview of Satyajit Das, May 7, 2010
what happened.

As we’ve just been talking, there’s a

whole bunch of risks in derivatives which are inherent
in them.

And effectively -- you know, it’s like the old

famous Mae West line:
strayed.”

“I used to be Snow White, but I

So basically, that’s actually, that’s what

happened with derivatives over a period -- not of five
years or six years, but like 20 years.
This started -- I saw this happen -- in the
middle to late 1980’s; and it gradually got worse and
worse and worse.

That’s what happened.

Anyway, that’s a half-a-hour answer to the
first question.
Christmas.

At this rate, we’re going to be here at

But -- so basically, that is what the

problems with derivatives are.
MR. STANTON:

I have a question, please.

I haven’t read much of your work.

After your

marvelous discourse just now, I plan to read a lot more.
But I did read your chapter in your book on risk
management which was, if I may say so, somewhat
disparaging.
And I’m curious -MR. DAS:
vicious.

Disparaging?

I was

basically being

You don’t have to say “disparaging.”
MR. STANTON:

All right.

My question is,

given that we have large, complex financial
25

FCIC Interview of Satyajit Das, May 7, 2010
institutions -MR. DAS:

Yes.

MR. STANTON:

-- we have people in serious

positions who may not understand what they need to
understand.

I’ve only seen two models so far of

companies that managed risk well.

One of them was a

company, JPMorgan Chase, that had a principal that said,
“If I don’t understand it, I don’t do it.”

Just a

fundamental logic.
The other one, of course, was Goldman, where
they presumably did understand it better than anybody
else, and it is a relative game.
I’m curious, for all the rest of the firms in
the world -- financial firms in particular -- what do
you recommend as the most appropriate risk-management
processes, structures, et cetera?
And then later, we can go to the questions you
raised about the board.
MR. DAS:

Okay, the first thing is, let me

just sort of make a small observation about JPMorgan and
Goldman Sachs, or whoever did well out of the crisis.

I

think there is a great degree of mythology about all of
this.

To be very blunt, they were lucky.
MR. STANTON:
MR. DAS:

Okay.

Okay.

Now, that doesn’t mean they
26

FCIC Interview of Satyajit Das, May 7, 2010
did not have the same processes in place that would have
worked very well, but they were lucky.
And I will make one slight observation about
that, is -- and let me explain why I think they were
lucky.
Goldman Sachs got lucky because John Paulson
came to them.

What is very clear in the indictment is,

Goldman’s famous risk management was more a result of
John Paulson pointing out certain things to them, and
Goldman’s light bulb going on above their head, going,
“Oh, my God, he’s right.”
The second one was JPMorgan was lucky because
they would have actually had the biggest problem of
anybody because they had the biggest derivatives books.
MR. STANTON:
MR. DAS:

Uh-huh.

It’s because the part that would

have blown them up was in disarray.

Because when

JPMorgan and Chase merged, their credit derivatives
operation had a lot of problems because the cultures
were different, the people had a lot of staff turnover.
And during that period, when, effectively, everybody
else was loading up, they effectively were in complete
disarray for about three or four years.

And

essentially, they were able to avoid that.
So I think you need to be -- and I’m not
27

FCIC Interview of Satyajit Das, May 7, 2010
saying that they didn’t do their jobs properly.

But the

point I’m making is, they got lucky.
The fact of the matter is, almost every bank
essentially has exactly the same procedures, the same
processes, the same, exact models they use, with slight
variations.
And let me explain why that’s the case.

The

first thing is, the regulatory rules actually prescribe
the models.

So the market-risk amendment of 1994 to the

BASEL I Accord is what everybody still follows.

Now,

what most banks do is supplement that with essentially
some variations to the statutory or the regulatory
models.
Now, there may be some variations, but I have
worked with quite a few of these institutions.

The

differences are there, but they are neither profound
enough or significant enough to actually make the type
or quantum of difference that you’re talking about.
So then the question becomes, if that’s not
the case, what makes the difference?
I think there are certain things.

One of the

better firms that I saw in my working life in risk
management was a company which became part of,
ultimately, now UBS, it’s called, “O’Connor Partners,”
which is a Chicago -- small Chicago specialist
28

FCIC Interview of Satyajit Das, May 7, 2010
derivative firm on the exchange.

And what I learned,

dealing with them, was they were good at it because it
was their own money.

It was actually as simple as that:

It was their own money, and they didn’t have a lot of
it.

So they hung on to their capital very, very well.
MR. FELDBERG:
MR. DAS:

Partners.

That’s Econo Partners?

The second -- that’s O’Connor

O’Connor Partners.
MR. SEEFER:

O’Connor?

MR. FELDBERG:
MR. DAS:

O’Connor?

O’Connor, “O,” apostrophe,

C-O-N-N-O-R, Partners.

It’s a very famous firm.

And then they were bought by SBC, and then
they were bought by UBS.
And effectively, if you actually look at that,
and you look at UBS and you look at what they did, they
adopted a lot of the systems -- not all of them -- but
it shows that the system itself is not the issue.
I think there were two fundamental things that
you also need.

You need to have risk managers who are

very good; but the point is, nobody wants to be really a
risk manager in a bank.
say “yes.”

Why?

Because we all like to

The risk manager’s job is to say “no.”
And the other thing is, if you really don’t

want to be the prom queen or the class captain, that’s
29

FCIC Interview of Satyajit Das, May 7, 2010
a job you really go for.

And I have met very few

masochists in my life who like that.

Because it’s a

horrible job.
And the other thing is, you also have enormous
business pressures within the firm on risk managers.
And no matter what you do to try to protect them, it’s
very difficult.
The other thing is, that’s not only at the
senior levels, those are the junior levels.

Nobody

effectively in the risk-management area, at the junior
levels, sees that as their long-term career.

They

basically want to move into the front office, to trade
or sell, and make $5 million bonuses.

Because in risk

management, you’re never going to make that.
So under those circumstances, what happens is,
you have to deal with traders, salespeople, structuring
people who, probably for 50 percent of the time, you’re
trying to do their job and the other 50 percent of the
time, you’re trying to basically, essentially do a
recruitment interview, so they like you and think you’re
good so you’re going to get a job with them.

And that’s

fundamentally what happens in all of these banks.
And then you have very complex oversight
layers where the blind are leading the deaf and dumb.
So basically, it’s just a complete farce.
30

FCIC Interview of Satyajit Das, May 7, 2010
I mean, Lehman’s had Henry Kaufman, the
original Dr. Doom, because Nouriel Roubini isn’t the
original Dr. Doom -- on their risk management committee.
And Henry Kaufman is not a person without ability, yet
the risk management committee at Lehman only met twice a
year.

In fact, I think they only met twice in 2006 and

2007.
So the processes are there, they’re very
similar; but they’re effectively not used in any
meaningful way.

And getting the quality of staff is

important, getting them to do their job.
And, look, the reality is that there are very
few fearless whistle-blowers in the world.
And the last comment I would make is that you
cannot at all dismiss a culture.

Almost all of these

people go to the same universities or come from the same
universities which the same doctrines are taught.
cultural milieu is very, very similar.

The

They think the

same, they talk the same, their ambitions are the same,
and they have a very, very cohesive world view, which
does not effectively withstand any tight scrutiny.
One book I would recommend you read, which
is –- it’s got nothing to do with finance -- it may be a
little bit of a challenge -- is a book by a woman called
Karen Ho.

And the book is called, “Liquidated.”

It’s
31

FCIC Interview of Satyajit Das, May 7, 2010
an ethnography of Wall Street.
you can skim a lot of it.
cultural constraints.

It’s hard reading, but

But you get an idea of the

And she’s basically an

anthropologist, who’s working in the field of
ethnography.

And it’s very interesting to see exactly

the cultures.
And that’s why risk management can never work.
The risk management can never, ever work in that sort of
environment.
And, you know, I don’t want to be negative
about it to the point of saying it can never be made to
work; it does work sometimes a little better, but
there’s a lot of problems.

And the fundamental thing

is, human beings have an almost infinite capacity for
self-delusion.

And one of the pieces of self-delusion

we have is that risk is something we understand and we
can actually manage.
And I think it’s always useful to go back to
Frank Knight’s concept of risk versus true uncertainty.
What is really problematic is real uncertainty.

And,

frankly, that’s very, very difficult for people to
grasp.

And I think a lot of the modern intellectual

disciplines mask that to a degree, which is extremely,
extremely dangerous.
And so all of those factors in risk management
32

FCIC Interview of Satyajit Das, May 7, 2010
become very, very important.
And, look, I think you can do whatever you
like with the board of directors, you can do whatever
you like with the internal structures -- and I’m sure
you will.

At the end of the day, I do not think it will

guarantee success.
And Mr. Greenspan is not one of my favorite
people on the planet; but I think he was right when, in
one respect, he said, you know, when he was asked a
question, I think at his testimony last year before one
of the committees, where he said, “Look, Senator, I
could regulate and I could assure you nothing will
happen; and I can not regulate, and I can’t give you any
assurance that there won’t be a failure, and that’s
exactly what you’re going to sign.”
Now, that doesn’t mean you shouldn’t try and
improve it, but I think you need to have realistic
goals.
But if you really want to do this, this is not
something you’re going to do, I know; but I could tell
you how I would do it.

One is, you’d want to make sure

that all their net worth is in the sum.

And you would

basically say that if you felt a discharge of fiduciary
duties, you would be basically shot by a firing squad if
you were if you were [unintelligible] found guilty.
33

FCIC Interview of Satyajit Das, May 7, 2010
That would basically focus the mind wonderfully well.
MR. SEEFER:

I think so.

Let me ask you one thing, since I believe when
you were talking about the various areas on these
derivatives, we were primarily, if not exclusively,
talking about credit derivatives.
Do you see -MR. DAS:

No, I was talking about derivatives

generally.
MR. SEEFER:

Okay.

Well, if you can then --

and, obviously, you can see I’m not one of the Ph.D.s in
the room -- can you say something about the role of –I
mean, we have the various kinds of derivatives.

And I

will tell you, we have been certainly looking probably
more at credit derivatives than other types of
derivatives.
But were the other types of derivatives part
of the causes of the crisis or the propagating mechanism
of the crisis, whether it’s interest-rate swaps,
foreign-exchange swaps, equity, et cetera?
MR. DAS:

To different degrees, yes.

But you

have to remember, this crisis, is the crisis of debt,
the amount of debt.

The securitization techniques and

the credit derivatives which basically deal with debt
will have a much greater disproportionate role.
34

FCIC Interview of Satyajit Das, May 7, 2010
If you actually look at the volume of CDS
contracts, credit default swap contracts, at the peak,
it was around about $62 trillion.
MR. SEEFER:
MR. DAS:

Right.

Which is less than 10 percent of the

total outstandings of OTC derivatives.
MR. SEEFER:
MR. DAS:

Right.

But because, effectively, they were

at the heart of the credit complex, they made things
worse.
So this is one of those problems with going
from one crisis to the other.
You know, the ‘87 crisis was a crisis of
equity markets.

The ‘97-98 crisis was one of emerging

debts.
Each crisis, to some extent, is slightly
different; so that the derivatives that affect it will
be the ones that are most germane to what is, if you
like, the central epicenter of a crisis.

This just

happened to be credit this time around.
But also, I think the fact that it was credit
derivatives and securitization which led to the heart of
this crisis, also has a subtle bi-line, and the bi-line
is this:

If you actually look at banks, no matter what

anybody will tell you, the core risk a bank takes is
35

FCIC Interview of Satyajit Das, May 7, 2010
credit risk.

And so the fact that these instruments

related to that is actually extremely important.
Let me explain what I mean by that.

What I

actually mean by that is the essence, the fact that
people thought they could use these instruments to
transfer risk, created the moral hazard issue that they
started to take on more and more risk because they
thought they found the magic, sort of formula for making
money with no risk.

And that really was why they became

important.
And I know you will be asked questions about
things like sovereign CDSs and so forth.

They are kind

of, to me -- the best way to describe a sovereign CDS is
it’s like -- it’s like basically Alice in Wonderland.
It’s like the Cheshire cat.
Can I ask you a really simple question?
The thought of JPMorgan selling protection on
the United States has to be an oxymoron; doesn’t it?
MS. NOONAN:

Is the point that if the United

States isn’t making good on its debt, then JPMorgan
isn’t going to be able to?
MR. DAS:

Exactly.

MR FELDBERG:
MR. DAS:

And we bailed them out.

That’s right.

Because JPMorgan was

selling protection to the United States while you were
36

FCIC Interview of Satyajit Das, May 7, 2010
bailing them out.

It was quite funny, I thought this

was just completely like the Cheshire cat.

Eventually

it’s going to grin at you and disappear.
MR. SEEFER:

It’s brilliant.

You know, another thing that we’re looking at
is -MS. NOONAN:

Well, because they’re making

money on something that they’re never going to actually
have to pay out on.
MR. STANTON:
MR. DAS:

Talk about counterparty risk.

I generally like making promises

that I don’t have to honor as well.
MS. NOONAN:

Right.

MR. SEEFER:

Don’t we all?

Another thing we’re looking at, is the whole
amplification within CDS, whether it’s the use of naked
CDS or whether it’s, you know, synthetic CDOs that, you
know, where bonds are referenced in multiple CDOs.
Have you done any work on that or do you have
anything you can help us with on that?
MR. DAS:
things I would say.

Look, I think -- there’s a couple of
Look, the issues about CDSs, if you

are logical, you will deal with as part of the
derivatives issue overall.

Because I think once you

start to basically deal with -- try to segment the
37

FCIC Interview of Satyajit Das, May 7, 2010
derivatives universe, it’s very dangerous.

Okay, let me

explain why I mean that.
If I say it’s a CDS, how is that different
from a derivative on a bond?
the same.

It’s not.

It’s exactly

So you get into definitional debate, which is

what, as you know, the Senate and the House of
Representatives are now in a tangle over, it’s a
definition of standardized derivatives and this is
non-standardized derivatives.
which you need to make.

It’s not a distinction

So I think you need to deal

with them overall.
Now, coming back to the naked CDOs issue, it
goes to the issues we’ve talked about.

If, effectively,

you want these instruments to use as hedges, then
basically you have to have an underlying position.

It’s

as very simple as that.
And so if you own a loan and you’re hedging
it, that should be permitted.
Similarly, if you want to take that risk,
instead of buying a bond, you basically are buying the
risk through a CDS contract, that should be permitted.
But you should have to put up 100 percent of the cash.
Because if you bought the bond, you would have to put up
100 percent of the cash.
And once you put those rules in place, that
38

FCIC Interview of Satyajit Das, May 7, 2010
will effectively deal with one element of the problem.
Now, let me also give you the lobbyists on the
other side who are going to come back to you.

Their

argument is going to be that, effectively, this will
destroy the civilization as we know it, the Parthenon
will crumble, the Capitol Hill will sort of have a crack
open up underneath it, because liquidity in the markets
will disappear.
The financial markets are generally not
liquid, anyway.

Okay, that’s the first comment I would

make.
Secondly, liquidity is not costless.
Liquidity has a huge, huge cost.

And one way to look at

it is, a speculative interest is actually when it blows
up, the cost of that liquidity.
And so the question is, you have to weigh the
two up.

And if you ban naked CDSs or naked derivatives

generally, you’ve got to make up your mind that the cost
and the benefit of that argument is in your favor.
And my view is, derivatives are just too large
at the end of the day.
size.

They have to be brought down in

That’s the bottom line.
MR. STANTON:
MR. DAS:

And --

Let me ask quickly, how --

-- you have to find a way to do it.

MR. STANTON:

I’ve got a question.

I mean, if
39

FCIC Interview of Satyajit Das, May 7, 2010
somebody creates a clearinghouse for derivatives in the
United States, what’s to stop the same trades from going
to London or taking on a different form that hasn’t been
defined in the statute?
MR. DAS:

Did I send you the paper called,

“CCP Trans[unintelligible] Solutions?
MR. STANTON:

I don’t have it, but I’d be

delighted to read it.
MS. NOONAN:

Yes, you sent it.

MS. SHAFER:

Yes, you did.

MR. DAS:

Yes, that goes to the clearinghouse.

MR. STANTON:
MR. DAS:
aspirin.

Thank you.

My only comment is, this is like

It cures everything.
It’s not going to clear anything.
The first thing is, the whole proposal is

deeply flawed.
The second thing is, to make this work, you
have to have one clearinghouse globally.
Now, you’re struggling with “too big to fail.”
This is “too big to survive.”
And there is just no way the Chinese, the
Europeans, the Japanese, the English and you are going
to agree to have one clearinghouse.

So you’re going to

have multiple clearinghouses, which has two problems.
40

FCIC Interview of Satyajit Das, May 7, 2010
One is the problem that you’ve just mentioned,
is that are you going to get migration to the
clearinghouse which suits them.
Or secondly, the problem is going to be, there
will be multiple clearinghouses with what we
call “interoperatability agreements” between them, so
you can clear through them.

And that is a recipe, in my

judgment, for complete disaster.
MR. STANTON:
MR. DAS:

One thing –

MR. STANTON:
MR. DAS:

I didn’t mean to take you --

-- off course.

Thank you.

No, that’s fine.

Just one last comment I would make on that,
which is kind of interesting, is that one of the things
I don’t understand -- and you will pardon my somewhat
bemusement at the attitudes of the United States to
anything which is state-owned.

I think there are

certain things which have to be owned by the state.

And

effectively, one of the things I would have thought is,
the clearinghouse would have to be owned by the state
because it’s the ultimate “too-big-to-fail” backstop
entity.
And if you actually put this into private
hands, it will be a disaster.

The reason it will be a

disaster is because if you look at individual commercial
41

FCIC Interview of Satyajit Das, May 7, 2010
entities running these things, they will want to make a
profit and they will want to compete with each other.
And the only way you can compete in a clearinghouse is
to make the credit terms and the terms of engagement
more loose, not more stringent.
you want in a clearinghouse.
a clearinghouse.

And that is not what

And that is fundamental to

And that’s going to be hugely

problematic.
And the day they put a clearinghouse in
financial markets, I’m basically buying a lot of cans of
baked beans, salmon, and tuna, I’m getting a gun -- and
I’m not great gun fanatic -- and I’m putting in an
electronic barbed-wire fence, and I’m growing my own
vegetables with my own spring and sitting there, waiting
for the people who are the sort of extras from the road
to come down the track.
Where were we again?
MS. NOONAN:

Well, I mean, are there not any

examples of other clearinghouses that work?
MR. DAS:

I mean --

Oh, yes, but they’re very specific.

They’re very narrow.
See, it’s like -- there’s always a problem of
scaling, right?

Something works on a small scale.

But

when we make the scale too large, you have different
problems.
42

FCIC Interview of Satyajit Das, May 7, 2010
We’re trying to make this a central
counterparty work for everything.
small scale.

It would work on a

I think when you make it to the scale

you’re looking at, it becomes very difficult.
For instance, I’ll give you a very simple
example.
If you make a very large central counterparty,
which is now going to basically be netting contracts
which range from basically catastrophe bonds with
derivatives to credit derivatives, to interest-rate and
foreign-exchange derivatives -- potentially the foreign
exchange seems to have fallen off the cliff -- but
basically, if you try to net all those positions, this
is going to be very difficult, and we’re going to go
back to assuming things about correlations between asset
classes which the history will make monkeys of.
I can see a certain level of depression is
setting in on your side.
MR. FELDBERG:
come up with a solution.
happened.

Fortunately, we don’t have to
We just need to explain what

And…
MR. DAS:

That’s right.

I think you’re in a

very -- I had a lovely conversation with one of your
senators’ staff, who at the end basically I don’t think
ever wanted to talk to me again.

Because they were
43

FCIC Interview of Satyajit Das, May 7, 2010
almost 28 years old and bright and bushy I’d about
changing the world.

They discovered the world changes

them, not the other way around.
MR. SEEFER:

Yes, yes.

And they were probably

even younger than 28.
MR. DAS:

They probably were.

But I was

giving them the benefit of the doubt.
MR. SEEFER:

Right.

MS. SHAFER:

Have you paid a lot of attention

to data selection and who has decent data, if at all,
and what data should be looked at?
MR. DAS:

You mean, going backwards or going

forwards?
MS. SHAFER:
MR. DAS:
basic of history.

Well, first, backwards, actually.

Look, the only data -- a little
Originally, nobody collected any

data, okay, because it was all internally within some
institutions.
In the eighties, what happened is a few
institutions that were smart realized that the data gave
them competitive edges, so they started to collect some
data.

And people like JPMorgan, Chase to some extent,

some banks which now don’t exist, like Bankers Trust,
collected a lot of data.
Then what happened was in the
44

FCIC Interview of Satyajit Das, May 7, 2010
nineteen-nineties, there was a lot of data on financial
market prices which were put together, mainly by
JPMorgan, which has now been spun off into Risk Metrics.
And then finally, the last phase was Market
Partners, which is basically a utility firm which is
owned by all the banks, which is, I call it, the plumber
of financial markets.

And they essentially provide a

lot of mark-to-market information and lots of
information.

But it’s mainly price data.

Now, there’s a parallel effort which came in,
in response to regulatory noises, which is the DTC
repository, which collects a lot data, particularly on
CDS contracts and feeds them back to people, which was
really a regulatory information-gathering exercise to
provide more transparency and so forth.
Now, I think that’s all very positive.
think that’s very, very positive.

I

But what you’ve got

to be really, really careful about is, A, the quality of
the data; and the second thing is, the detail and
granularity of the data.

Because I’ll give you why I

think it’s kind of interesting.
You will remember about two years ago, Société
Générale was blown up by Mr. Kerviel.

Mr. Kerviel

reported net positions, and his net position was zero.
The problem is, he was long $50 billion of certain
45

FCIC Interview of Satyajit Das, May 7, 2010
stocks.

And the reason he showed zero was he had these

fictitious trades on the other side showing net, that he
was long and short.

So when he netted the two things

out, he had no position.
So if you start to get net data, it becomes
very difficult.

And since it’s all about the degree

of granularity -- and to some extent, the
information-gathering is very valuable, but you need
to understand what you’re trying to do with the
information, how that will help you achieve whatever
regulatory or control objectives you have, and get that.
And the industry will always say that it’s
competitively fenced in and essentially people will
trade against us if they know the data.
absolutely true.

And that’s

That’s absolutely true.

So balancing

that is not easy.
And also what you have to do is, you can’t
collect fragmented data from one class of institution or
from one geographic location.

So you would have to

collect it for everybody -- every institution which is
active in the market.
And let me explain why that’s important
because people often use filters of size of positions
and trades.

And if you go back to my example of

embedded leverage, I could have a very small trade, but
46

FCIC Interview of Satyajit Das, May 7, 2010
its potential amplification of loss under certain
circumstances could be huge.

And that’s what you’ve got

to be careful about.
MR. FEDLBERG:

So what should we be most

concerned about when it comes to causes of the crisis?
I mean, you know, obviously, the losses to Citi and AIG
and UBS were probably the biggest places where
derivatives were relevant.
MR. DAS:

Sure.

MR. FEDLBERG:

But are there other ways where

it was -- where the market had an impact?
MR. DAS:

I think one of the things it did was

transmitted the virus fairly effectively.

So if you

think of Citigroup and so forth as Patient Zero, they
slept around a fair bit.
MR. FEDLBERG:

So more or less allegorically?

Metaphorically? **
MR. DAS:

They traded with everybody.

And

this is a concentration risk issue, because basically,
there’s only less than 12 dealers in the world to which
everything flows.

And so once you get -- and the

problem is, if Citigroup infects Barclays, then Barclays
can infect thousands of people.
MR. FEDLBERG:

So what do you think of them?

Some people argue to us that the derivatives market
47

FCIC Interview of Satyajit Das, May 7, 2010
actually worked very well; and even after Lehman went
bankrupt, trades were transacted as they were supposed
to; and that the only real problem is that there were
kind of nodes in the system, or dead-ends, like AIG,
where there was no hedges, so people had to take losses.
MR. DAS:

You’re talking about my friends at

ISDA, I think.
MR. FEDLBERG:
MR. DAS:

I think that’s very likely.

That’s right.

Mr. Pickel.

I think -- isn’t it a wonderful name for
somebody to -- who has to be the head of the trade
organization with the name of Pickel?

I think it’s just

got to be one of life’s great, great, surreal pieces of
divine humor.
But, anyway, leave that to one side.
It worked well in the sense that, effectively,
it functioned.

But let me actually now pick up on that

point since you’ve raised it.

A couple of interesting

issues that I would raise with you.

ISDA is a private

body, it is actually, basically, funded by the industry.
It’s a lobby group, yet it plays a very disproportionate
role in financial markets.
by that.

And I’ll explain what I mean

I sent you a series of papers on credit

derivatives, because that seems to be an area of focus.
Particularly, I draw your attention to the one
48

FCIC Interview of Satyajit Das, May 7, 2010
called, “Quantum Mechanics.”

And effectively, what my

argument there is, is that ISDA have created documentary
framework is almost extrajudicial.

They have now what

they call the “Determinations Committee,” which now
determines whether somebody has defaulted or payments
will be made and credit default fault swaps.
MALE SPEAKER:
MR. DAS:

Uh-huh.

Now, to give you some idea of the

extent of that and the problems that it causes, AMBAC
has not filed for Chapter 11 or any other form of
bankruptcy, to the best of my knowledge, yet the ISDA
Determinations Committee has ruled that they actually
have triggered the bankruptcy provision of its CDS
contract, okay, which is quite extraordinary.
And technically, that might be correct, but
it’s kind of an extraordinary step, is the first comment
I would make.
The second thing is, if you actually look at
the paper it goes through, that the mechanism by which
they work out the pay-out on the CDS contracts
themselves have exaggerated the losses enormously, which
suggests that these mechanisms have distorted the market
quite considerably.
And I do not see how that sits particularly
well with the comment that everything has functioned
49

FCIC Interview of Satyajit Das, May 7, 2010
hunky-dory.
The next comment I would make is that you
should look at Lehman’s and the Lehman’s bankruptcy
proceedings, because there is an awful tale of what goes
wrong with derivatives.
And I have to be very careful talking to you
about Lehmans because essentially I have some
involvement in the matter, so I will tell you what I
can.

But there’s certain things I can’t talk about.
The major thing is, when Lehmans failed, all

the derivative contracts with Lehmans would have either
automatically terminated or been able to be terminated.
Now, that would require the following
procedure:

It would close out all the contracts, and

the idea would be that you would value them.

The way

you’re meant to value them is basically value them at
market prices by getting market quotations.

And the

idea is that everybody on the other side would go into
the market and hedge themselves, and they would work out
what they either owe Lehmans or what Lehmans owes them,
and then there would be a settlement.
That process, if you go to Lehman’s
bankruptcy, you will very quickly discover that
basically most of the contracts did not actually settle
in the way that it was intended.

And there are huge,
50

FCIC Interview of Satyajit Das, May 7, 2010
ongoing legal disputes about that.

And this is all on

the public domain.
And if I can make one suggestion:

You might

like to get somebody from the firm of Alvarez & Marsal,
which is actually running the restructuring of Lehmans.
And you might interview them about their experience with
the derivative contracts.
salutary.

I think that would be fairly

And perhaps then, you can calibrate the issue

about whether or not the contracts worked as intended.
MR. STANTON:

Is that in the trustee’s report

as well, this information?
MR. DAS:

That would be in the trustee’s

report, you’re absolutely correct.

That may be in the

trustee’s report or the -- are you referring to the
trustee’s or the examiner’s report?
MR. STANTON:

I’m sorry, the examiner’s

report.
MR. DAS:

I’ll be honest, the examiner’s

report has some of that.

But the examiner got a bit

excited by repos, the repo when it [unintelligible],
they spent about 2,000 pages talking about that.
But the better one to go to would be Alvarez &
Marsal actually have to publish a statement.

And so in

the court filings, you’ll find a very voluminous
document, plus some PowerPoint slides which set out this
51

FCIC Interview of Satyajit Das, May 7, 2010
process.
To give you some idea, probably less than
50 percent -- very much less than 50 percent of the
derivatives contracts at Lehmans had opened, which was
like 1.2 million contracts, have actually been able to
be settled and closed out.
MR. FEDLBERG:
MR. DAS:

I’m sorry, what percent?

But, of course, ISDA doesn’t find it

convenient to talk about that.
MR. STANTON:
MS. NOONAN:

Thank you.
Yes, this is Dixie Noonan.

I

have a question.
You talked some about models.

Do you have any

familiarity with the model that AIG used to price its
credit default-swap contracts?
MR. DAS:

Yes.

It’s a standard model.

There’s nothing unusual about their model.
MS. NOONAN:

Is this the Wharton model or is

this the model that -MR. DAS:

No, no, no, it’s not the Wharton

model.
Basically, what you do -- I mean, it’s not
rocket science at the end of the day.
What do you model these securitizations?

You

put in a bunch of default probabilities for the
52

FCIC Interview of Satyajit Das, May 7, 2010
mortgages, then you say if they’re going to go and not
pay me, how much can I sell the house for?
the mortgage outstanding?
lose?

How much is

So how much am I going to

And you put in a default correlation factor, so

that basically if one bunch of mortgages goes down, one
other bunch of mortgages will simultaneously go down.
That’s all you put into that.
But it’s not the model -- the model itself is
fairly standard, and the Gorton model -- did you say
Wharton or Gorton?
MS. NOONAN:
thing, but yes.

Well, I think it’s the same

I said, “Wharton.”

MR. DAS:

Gorton is the guy -- the poor guy

who’s going to go to his grave being blamed for this,
together with David Lee.
MS. NOONAN:

Is there a reason why he

shouldn’t go to his grave, being blamed for it?
MR. DAS:

Absolutely, because he did what was

standard practice.
He was asked to solve a problem, right?

He

was asked to solve a problem of, you know, how many of
these mortgages are likely to default and how large are
the losses to be?

He solved their problem.

What he did not solve for is what could
happen, even if these mortgages didn’t default in terms
53

FCIC Interview of Satyajit Das, May 7, 2010
of the market-to-market, which is based on credit
spreads blowing out.
And this is quite important to understand.
And since you raised this, it’s worth spending a couple
of minutes talking about that.
I’ll have a bet with you that the AIG
credit-default swaps on which they owed $18 billion, if
they’re held to maturity, the losses will be different,
and probably substantially lower.
MS. NOONAN:

Right.

Now, they have particular

problems that were –MR. DAS:

The [unintelligible] mark to --

MS. NOONAN:

-- I mean, it seems to me -- and

correct me if I’m wrong -- it seems to me as though they
had particular contractual provisions that made AIG post
collateral in various events -MR. DAS:

That’s right.

MS. NOONAN:

-- that were different from other

CDS contracts.
MR. DAS:

No, no, come back a stage.

You were

correct in your first half but not quite in the second
half.

Let’s go back, okay?

Let’s go back to the first

things.
Now, if I sold protection to you on these AAA
and super-senior pieces of mortgages, anytime I really
54

FCIC Interview of Satyajit Das, May 7, 2010
lose money, if the losses get to what’s called
the “attachment point,” which might be like on these
pools at 30 to 40 percent total losses, that doesn’t
mean that’s not going to happen, but it’s going to be
highly unlikely it’s going to happen on every pool of
mortgages.
And remember, AIG sold protection mostly on
pre-2005 and pre-2006 mortgage pools.

That’s the first

thing, okay.
MS. NOONAN:
MR. DAS:

Uh-huh.

So if I can hold these things

through to maturity, then effectively, there is a good
chance that I won’t lose that much money at all.
Now, the next thing is, then come back one
stage and say, “Well, what was the problem?”

The

problem was, when I agree to mark these contracts to
market, how do I do it?
Now, what Gorton was doing at AIG, he’s using
historical data on the mortgage defaults, likely levels
of recoveries, and correlations to work out what he
thought these mortgages were worth.
Unfortunately, the markets don’t use that.
The markets use market data.

And what they do is, they

calculate some of this import from credit spreads.

So

what happened is in late 2007, early 2008, as the
55

FCIC Interview of Satyajit Das, May 7, 2010
subprime losses went up, the credit spreads went up.
Because, as you know, people who trade in financial
markets don’t make a very pretty sight at the first sign
of panic.

They may be very cool at Gorton Gecko’s when

things are going well, but they’re not when things don’t
go well.
So credit spreads went absolutely ballistic.
And under those circumstances, what actually
happened was, when you came to mark these contracts to
market, there were massive losses.

But those losses are

just people’s estimates of what they think they’re going
to be.

Whether they’re actually going to take place or

not is irrelevant.
However, if you have to mark your contracts to
market, firstly, that has an earnings impact.
And this goes back to my earlier comment, that
basically if you’re going to mark this stuff to market,
you have to have a market.
this stuff.

And there is no market for

So basically, you were just marking to

model, and the models would try to be calibrated to
market inputs, which is basically completely nonsensical
because it’s based on three people-who-are-panicking’s
estimate of what’s happening in the world.
Then the next problem becomes AIG had a
provision.
56

FCIC Interview of Satyajit Das, May 7, 2010
Now, you have to go back in history.
AAA-rated.

AIG was

And AIG Financial Products made its name in

doing very long dated transactions, 30-, 40-year
transactions.

And I did some of these with them.

Now, when people used to do these transactions
with them, there was always a point of debate with them.
And the point of debate was simple:

You’re AAA now, but

you may not be AAA in the future, so how do we protect
ourselves?
So what most people did -- which is quite
logical as well -- is to put in place what we call a
contingent collateral provision.

Under the terms of

those provisions, if AIG got downgraded below AA-minus,
then what would happen is, they would have to post
collateral.
So this is actually worse than collateral
normally, which is gradual [unintelligible].
So under those circumstances, what actually
happened was that when the problem started, you had the
combination of two quite toxic events, which is the fact
of the mark-to-market prices on these mortgages was
very, very adverse; and, in my view, overestimated the
risk of default quite considerably.
And then secondly, you have this liquidity
condition being triggered because theoretically AIG
57

FCIC Interview of Satyajit Das, May 7, 2010
had to pay over the collateral based on these
perhaps-not-correct values.
So to go back to Mr. Gorton, Mr. Gorton was
asked to work out whether he would lose money.

And he

may or may not be right.
But the second question is, he was never asked
about what would happen if there was a mark-to-market
problem and there was this collateral provision, which
I suspect he wasn’t even aware of.
MS. NOONAN:

Well, whose fault is that?

I

mean, at some point, if they were allowed a model,
shouldn’t someone who understands those provisions of
the contract work with rather whoever is putting the
assumptions into the model?
MR. DAS:

But if I can take you back to an

earlier discussion, I pointed out that different parts
of the same organization deal with different parts of
the problem.
So Gorton was given a problem in a nice,
little box.
Because generally, with quants, the way you
treat them is you keep them in a little cage, you throw
them some food occasionally, and you get them to do some
party tricks occasionally.

But you train them to do the

party tricks.
58

FCIC Interview of Satyajit Das, May 7, 2010
MR. SEEFER:
MR. DAS:

And you don’t --

Frankly, you give them the stuff.

I’m sorry?
MR. SEEFER:

I was just going to say,

apparently you don’t throw them a copy of the CDS
contract?
MR. DAS:

No.

they’re not lawyers.

No, no, no.

Because, A,

They wouldn’t understand it.

Besides, there’s a selective fee.

You don’t want people

to know too many things because that causes problems.
And part of that is not because you’re trying to hide
things.
Part of it is, you don’t want the person to
decide to try to work out what the contract is when he
may not, which may waste everybody’s time.
MR. SEEFER:
MR. DAS:

Uh-huh.

So essentially, that’s how the whole

system operates.
And so, you know, there would have been
something called a “Peratt” [phonetic] committee meeting
or something.

I have chaired these committee meetings,

I have attended them, I have presented at them for most
of my life.

They’re a joke, because a product like --

if you do it properly, it would take you two or three
days to go through very carefully and work out what the
59

FCIC Interview of Satyajit Das, May 7, 2010
risks were, and then actually deal with this.
But fundamentally, effectively, what would
happen is that you would rely on somebody who gives you
a ten-page report which you don’t read, but saying to
you at the meeting, “No, that’s fine, I think we’ve got
the risk all boiled down and I have a model.”

That’s

what it would boil down to.
And Gorton probably would not have seen the
CDS contract.
done.

He wouldn’t know known exactly how it was

And he was basically answering the question he

was put.
So if you gave him a grade on his exam answer,
it would probably be an “A.”
I think the stunned silence suggests that you
had a much more high-tech version of how banking works.
MR. SEEFER:

Maybe we were just under the

false impression that the people that build models might
know a little bit more about the product they’re
building the model for.
MS. NOONAN:

Well, let me press this a little

bit, and this is based on, you know, what I learned from
reading Michael Lewis’ book, “The Big Short.”

So we’re

trying -- we’re partially sort of, I’m trying to figure
out if what he wrote is right.
But if Gorton is the one who’s responsible for
60

FCIC Interview of Satyajit Das, May 7, 2010
putting the correlation assumptions into the model on
the risk of the rest of the mortgages defaulting, if
some of the mortgages default, in Lewis’ book, he says
that Gary Gorton was asked, “How many of these loans -how many of these bonds do you think have underlying
subprime residential mortgages as their assets?”

And he

said something like, “Oh, maybe 10 percent,” when, in
fact, it was 90 percent.
So if the person -- I just -- I’m trying to
figure out who I’m supposed to go ask, who should have
known that the assumptions were wrong.
MR. DAS:
right?

You have to go back to the layers,

Somebody writes the mortgages.

He’s the only

person who probably knows whether the guy is actually,
basically going to pay you back or not because he’s
looking at him across the table.
Then that’s sold to somebody who then
aggregates at a higher level of data, and somebody who
aggregates it another high level of data.
By the time you get to the end, there’s just
no way, because each one of these mortgage pools would
have between, say, two and half thousand and 5,000
mortgages.
Unless you go through all of them and actually
collect the data, it’s very, very difficult for any
61

FCIC Interview of Satyajit Das, May 7, 2010
person to do it.
And it’s like everything.

As you know, we use

reduced-form models in financial markets because you
can’t deal with the complexity at that level of
granularity.

You just can’t do it, so you simplify it.

And once you do that, people lose interest in the
detail.
I can give you a quote from one of the rating
agencies, which basically the guy who was doing some of
this said that he was an expert in the securitization
process, not in the underlyings.
MS. NOONAN:
MR. DAS:

Right.

And he had no interest in the

underlying.
Look, you know, I’ll be very blunt with you:
If you give me a financial problem, I’ll make it elegant
and simple.

Why?

Because I don’t want to go and root

around in 5,000 mortgages of some idiot in Nevada,
whether he’s going to pay me back or not, and how he’s
related to the next idiot in Nevada.
that.

I like buying loans.

I don’t want to do

I don’t want to go around,

you know -- and these people are big, nasty, ugly, with
tattoos and guns.
That’s all my competitive advantage in life.
MR. SEEFER:

What about -- have you heard
62

FCIC Interview of Satyajit Das, May 7, 2010
any -- I mean, one of the things that we’ve read about
and heard about from some folks that we’ve talked to is
when we get into the spring and summer of ‘07 and the
collateral calls come from Goldman to AIG, is that, lo
and behold, you know, Goldman’s models are resulting in
marks that are substantially lower than any of the other
broker/dealers and, lo and behold, that apparently is
the first big collateral call on AIG, not to mention
apparently having a ripple effect on Bear Stearns at the
time, in terms of how they should be valuing their
assets and perhaps contributing to the BSAM blowups.
Any knowledge about that whole area?
MR. DAS:

Oh, I know a lot about that but

there’s not much I can tell you without having to kill
you, but…
MR. SEEFER:

Well, you’re far away.

I’ll take

that chance.
MR. DAS:

Okay.

No, I take my confidentiality

and conflict of interest very seriously.
I can tell you certain things which are
well-known, and I can’t tell you other things.
MR. SEEFER:
MR. DAS:

Okay.

But just before I do that, coming

back to your question is, I’m a little worried that
you’re relying on Michael Lewis’ “The Big Short” as a
63

FCIC Interview of Satyajit Das, May 7, 2010
deep piece of analysis on this.
MS. NOONAN:
MR. DAS:

It’s just a starting point.

I hope so.

I hope so.

You know how there’s chick-lit?
crunch-lit.
crunch.

There’s now

And I would emphasize the “lit” part of the

There’s quite considerable degrees of factual

inaccuracies, because I happen to know several of the
individuals involved in all of this.

And I also happen

to know some of the episodes he’s describing.
And there is -- and he writes beautifully.
And so there is a little bit of literary license in
that, which you need to be careful about.
But coming back to the margin calls.
raised a couple of issues, okay.
when did this all happen?

You’ve

The first thing is,

People have the wrong idea.

This started to happen at the end of 2006.

I noticed

this probably around the third quarter of 2006, because
basically at that stage, what was happening was the
mortgage brokers were sort of falling over like
nine-pins.

And that gave me quite a degree of

considerable anxiety about what actually was going on.
That’s the first thing.
The second thing was, when the margin calls
started to come, there was a huge problem because, as
you know, when you do margin calls, you have to mark the
64

FCIC Interview of Satyajit Das, May 7, 2010
contract to market.
Now, the first thing is, nobody really knew
where to mark [unintelligible] the market.

And this

also loops back to a question you talked about in
models.
Now, effectively, the models used, what’s
known as a Gaussian copula, to do the correlation and
the creation of the loss distribution.
Now, you have to put the correlation into
that, and you take the correlation across the portfolios
of mortgages.

Now, without getting too geeky about it

or too wonky about it, effectively what was happening is
a little technical problem.

If people sort of said,

“Okay, well, really, the AAA should be priced at X,”
then what we were finding was the correlation had to be
more than one to get to that number, which is, of
course, impossible, which was pointing to me that the
fact that the models are actually deeply flawed.
everybody said this was just an aberration.

But

And since

we didn’t have a better model, so let’s not get too
head-up about this.
So what people decided to do initially -which is the end of 2006, early 2007 -- is keep the
marks on the AAAs at very close to par.
So what you have is a discussion which went
65

FCIC Interview of Satyajit Das, May 7, 2010
something like this:

“Where do you think we should mark

the AAAs?”
Everybody goes, “Well, we know it’s going to
pay back in full.

So we really should be at par or a

hundred.”
“Oh, but all the other spreads have moved and
we’ve got all these defaults.

What should we do?”

“Well, let’s mark it at 95.”
That’s what’s going on.
Now, then what happened was, you will remember
the introduction of the famous ABX index.
MR. SEEFER:
MR. STANTON:
MR. DAS:

Uh-huh.
Uh-huh.

The moment the ABX index was

introduced -- and I have to say, that was not a very
smart thing for people to do.

But, you know, we are all

wise with the benefit of hindsight.
The moment the ABX was introduced -- so people
like Paulson and everybody else wanted to short that.
And then the other thing which also happened at the same
time, was the banks, which had mortgages in their books,
and because the securitization market was shutting down,
and they had all this inventory that they couldn’t move,
they basically now had to go and hedge.

And everybody

started to hit this like there was no tomorrow.

So it
66

FCIC Interview of Satyajit Das, May 7, 2010
was just technical.
And as that index fell to “4,” now everybody
had a problem, because you can see this index trading up
there and you’re marking these things at 95, AAA
tranches, where effectively the market’s saying it’s
trading at well below that to the ABX.
MR. SEEFER:
MR. DAS:
got very tricky.

Uh-huh.

And at that point in time, things

Okay, things got very, very tricky.

And so that’s when people started to -- and
then, you know, people started to put in model
correlations of 140 percent into the AAA tranches to get
to the number they needed to.

And, in fact, the big

problem was, how to make the Gaussian copula work with a
correlation of more than one?
I remember having these discussions with three
very puzzled people with backgrounds in non-financial
disciplines, who said to me, “You want me to do this?”
And I would say, “Yes.”
And they would say, “But this is stupid.”
And I would say, “Yes, it’s stupid; but this
is what I want you to do because this is the number I
want to come out of the other side of the system.
don’t care how you get there.
is the output.

So I

This is the input, this

Just go and code this thing up so it
67

FCIC Interview of Satyajit Das, May 7, 2010
works in this way.”
And they would roll their eyes and look at me
like, “This guy is insane.”
And I’d say, “I’m not insane, but I’m paying
your paychecks.

So can we please get on with life?”

MR. SEEFER:

So let me ask you another thing

that you didn’t mention that we’ve heard from some folks
that also may have contributed to these problems in this
time frame, was at least in the U.S., anyway, there was
a change in SEC reporting requirements, that required
people to disclose Level 1, Level 2, and Level 3 assets.
And a lot of people sort of noticed, “Holy shit, look at
all the Level 2 and Level 3 assets.”
Any comment on that?
MR. DAS:

Yes, I can certainly make a couple

of comments about that.
I don’t think that caused the crisis.

It’s

just one of those wonderful things about transparency.
Everybody thinks transparency is wonderful, and it later
reveals the truth, then it’s not so wonderful.
So, you
MR. SEEFER:
all along, right?
MR. DAS:

Unless it would have been there

I mean -Yes, it’s best not to know the evil.

It’s like, you know, your partner says that
68

FCIC Interview of Satyajit Das, May 7, 2010
she or he is working late at the office, when you think
she is having an affair.

Do you really want to know the

truth?
MR. SEEFER:

There’s been a lot of memorable

phrases from you during this interview.
MR. DAS:

Never mind, never mind.

Anyway, coming back to that point -MR. SEEFER:

As I’m working late tonight, my

wife is wondering where I am, so…
MR. DAS:

That’s right, that’s right.

And I

will tell on you if you don’t pay me a large sum of
money.
I used to trade for a living; so, you know,
don’t ever tell me anything that you don’t want other
people to know.

Never let a trader know anything that

can be made to work against you.

They will use it at

some point.
MR. SEEFER:
MR. DAS:

Of course.

Coming back, just before I come back

to the Tier 1, Tier 2, just to answer the question on
the correlations and the marks, just to mention one part
I just remembered, which is the Goldman Sachs
accusations that they marked these positions at very,
very disadvantageous levels to AIG.
That was the question as well; wasn’t it?
69

FCIC Interview of Satyajit Das, May 7, 2010
MR. SEEFER:
MR. DAS:

Yes.

Okay, look, I’m sure that was

possible, but you will never, ever nail anybody on that
for a very simple reason.

I can just move correlations

and some of the assumptions by one or two percentage
points and get to the result that I want.
You certainly need to understand thing.
Generally, when we use models, the traders and the
people who really know what’s going on, use the model to
get to the number they think should be the number.

And

Goldman’s probably said to themselves -- and I can never
prove this, and you can never prove this, either -- you
know, said that, “Look, this is going to get a lot
worse.

We need to get in more collateral while they

have cash.”

So basically, you would move the number to

a number that you’re comfortable with.
And if you think this process is actually a
scientific process, which has worked from principles of
high science, that’s not how it works, because that’s
what I would have been doing.
What I would have been saying is, “Look, they
have money now?

How much -- how bad do we think -- how

bad do we think we’ll get?

If we need to get in

$5 billion, we’ll tweak the numbers so we get to
$5 billion, for God’s sake.

Let’s not be purists about
70

FCIC Interview of Satyajit Das, May 7, 2010
it.”
MR. STANTON:

Well, why would AIG pay it?

Why

would AIG -MR. DAS:

Read the agreement.

Read the

agreement.
The agreement is very clear:

One party has

the right to call for collateral and they have the right
to do the calculation, and they’re the calculation
agent.
MS. NOONAN:
simple.

Well, I mean, it’s not quite that

I think there was a process where you had to

get five dealers just to bid bids.
MR. DAS:

But, anyway…

Well, that’s easy.

That’s the least

of your problem, because all the dealers have the same
interest because you have a quiet conversation with
them, and the conversation is never conspiratorial, but
it goes something like this:

It goes -- you say to the

guys, “You know, we think we’re having some problems to
actually marking these.
number.

You know, we think that’s the

What do you think?”
MR. STANTON:
MR. DAS:

seems right.
that’s it.”

Got it.

And they go, “Oh, yes.

Yes, that

Which side of the trade are you?,”

“Oh,

“Oh, yes, we think that’s right.”

MS. NOONAN:

So you said you were actually -71

FCIC Interview of Satyajit Das, May 7, 2010
you were actually directing people to do models in this
way.
What work were you doing?

I’m sorry that I

don’t -MR. DAS:

Oh, don’t be sorry.

I consulted

people.
So if somebody came to me and said, you know,
“If the correlation’s more to one, more than one, how do
we make the model work?”
MS. NOONAN:

I see.

So you were just on the

outside consulting basis?
MR. DAS:

Yes.

I don’t work for anybody.

Nobody would employ me.
Not if they have any common sense.
MR. STROEBEL:

So -- I’m sorry, if I’m a

little slow on this; but if you’re trying to raise the
marks, wouldn’t you need the correlation to go down and
not up?
MR. DAS:
looking at.

It depends on which tranche you’re

If you’re looking at the super-senior ones,

you have to put the correlation up.
MR. STANTON:

If the super-seniors are going

to get hit, the model is [unintelligible] go down
together.
MR. DAS:

Yes, that’s right.

If the
72

FCIC Interview of Satyajit Das, May 7, 2010
super-senior has to get hit, you have to have a lot more
defaults.

So you’re saying, with very few defaults,

there’s going to be a huge number of defaults.
why the correlation has to go above “1.”

That’s

That’s just

how it works.
On the lower tranches, the correlation has to
go towards zero.

So, by the way, if you actually go and

check some of these and get the correlations on the
equity and mezzanine tranches versus the senior and
super-senior, we’re using different correlations for
those.
So we were using what’s known as the “skew.”
And the skew was on the low tranches we were using, very
close to zero.

And we were using very, very high

correlations on the super-senior tranches.
MR. STANTON:

May I ask a totally different

question, please?
MR. DAS:

I haven’t answered Tier 1 and

Tier 2.
I’ll come back to Tier 1 and Tier 2 -MR. STANTON:
MR. DAS:

A different question.

MR. STANTON:
MR. DAS:

Oh, go ahead.

Go ahead.

Can you hang on to your question?

MR. STANTON:

Yes.
73

FCIC Interview of Satyajit Das, May 7, 2010
MR. DAS:

Okay, Tier 1 and Tier 2, Tier 3 goes

to the accounting problems that we’ve alluded to, that
you couldn’t mark-to-market.

And it created,

effectively, an illusion that these assets were bad.
Now, they may well be bad; but the point is,
Tier 1, Tier 2, and Tier 3 are very rigorous under the
accounting rules.

Tier 1 is where you have prices;

Tier 2 is you don’t have the price of the instrument but
you have verifiable input, which you can put through a
standard model; and Tier 3 is everything else.
And so people got very nervous about Tiers 2
and 3 under those circumstances.

So in themselves,

Tier 1, Tier 2, and Tier 3 are not meaningful; it’s just
that they created this illusion that all the Tier 3
assets were bad, which may or may not be the case.

And

depending on what the assets are and how well they’re
marked, that would not be problematic.
But the other thing that the Tier 3 assets
created was a concern that people were actually
manipulating the values of those Tier 3 assets.
So you have to disconnect -- the accounting
rules allow you to -- or force you to classify in a
particular way.

But on the other side, you know, you

don’t know how they’re being done.

So while in one

level it looks very transparent, on the other level,
74

FCIC Interview of Satyajit Das, May 7, 2010
you don’t know how transparent it is and how much
manipulation is going on.
So I heard banking analysts, for instance,
basically say, “Well, all the Tier 3 should be written
off against equity.”
And I’m going, “Why?
there.”

I don’t know what’s in

So that created even greater uncertainty.

And

that’s the problem with Tier 1, Tier 2.
Look, I’ll send you after this paper –- after
this conference call, rather, two papers:

One on

embedded leverage, one I published on the accounting
standards, and basically my comments on the accounting
standards and looking at this Tier 1, Tier 2, and Tier 3
issue.
So you might find them moderately entertaining
over a bottle of vodka.
MR. SEEFER:

Okay.

I might just take you up

on that.
MR. STANTON:
MR. DAS:

You’ll send the bottle of vodka.

Next question.

MR. STANTON:

Next question:

Wall Street hoards information.

For good reason,

Because if I’ve got

more than somebody else, I benefit.
What was the incentive for the people that
created the ABX index that you mentioned, and what was
75

FCIC Interview of Satyajit Das, May 7, 2010
the incentive of people to let their trades -- the
prices of their trades be reported on the index?
MR. DAS:

Oh, okay.

The ABX was part of a

whole, sort of movement.
As you know, there’s this whole family of
credit indices which started -- actually started right
about ‘97, so it’s a long time ago.

And at that stage,

the individual investment banks had their own indices.
And eventually, what happened was, they were
consolidated into things like the ABX.
Now, the background to that is, it was a
business necessity.

And it was a business necessity for

some curious reasons.
One of the curious reasons was, if you wanted
to sell anything to a fund manager or an investment
firm, they ask two rather inane questions.
The first question is, is there a benchmark?
Because they need a benchmark to track, effectively,
their performance, and show that they’ve outperformed -as you know, the investment management industry, they
like to outperform indexes.

So if the market’s gone

down 40, you’ve only gone down 30, they actually show a
performance of plus-10 and then claim bonuses on that,
by the way.

But that’s neither here nor there.

So that’s the first people who wanted these
76

FCIC Interview of Satyajit Das, May 7, 2010
indexes.

So it was a business necessity for that

reason.
The other reason it was a business necessity
was, with the previous credit indexes that were created
on CDS contracts -- this is like the iTraxx and the CDS
indexes -- there was quite an active trading market in
that, and, so the dealers hoped to capture some fees
from the bit off the spread of trading those.

So that

was the second reason.
And there was a third reason, which came,
actually, from auditors and accountants.
Now, as you imagine, when we are going around,
putting these correlations into these models, these
correlations are sort of works of fiction, because none
of us know what these correlations are, right?

And if

you’re honest, you know what answer you want to get, and
you work back into correlation.
tell the auditors that.

But you obviously don’t

You basically say, you have to

defend the correlation using history and so forth.
So, generally, I used to go back to the
Napoleonic wars to use the correlation sometimes and
say, “Well, you know, it’s a bit like the Battle of
Waterloo.”

So we just used to make up a piece of

history.
But then the auditors got a little smarter and
77

FCIC Interview of Satyajit Das, May 7, 2010
started to ask questions like, you know, “Is there a
market benchmark that you can point us to?”
And what happened with the CDX and the iTraxx
index is, there are specific tranches on those indexes
which trade.

And we used to back out the correlations

because there are certain techniques by which you can
do that.

And that became very much accepted by the

auditing profession and the accounting profession as
ways to calibrate the inputs.
And so basically, there was some business
drivers to creating the ABX index to enable us to track
some of this information.
And the last element was that some of the
dealers themselves were very keen, rather like the
S & P 500 equity futures; that they were looking for an
instrument which was available to provide them with some
level of hedging.

And so all those factors conspired at

the same time rather disastrously and unfortunately in
the creation of the ABX.
And, obviously, at that point in time the
people who were creating this didn’t realize that it was
going to -- the market was going to tank and, B, that
this was going to become problematic in terms of
effectively showing the actual performance in a way
which was going to create big problems for them.

Nobody
78

FCIC Interview of Satyajit Das, May 7, 2010
realized that.
MR. STANTON:
MR. DAS:

And when --

I’ll be honest with you, I didn’t

realize it at all.
MR. STANTON:
being?

When did the ABX come into

Sorry?
MR. DAS:

You’d have to look at it.

It would

have been around September -- it was in 2007 -- 2006,
2007.

I don’t remember the exact date.
MR. STANTON:
MR. DAS:

Okay.

I think it was 2006.

Just go and have a look at it.

I think it’s

2006 it came in.
MR. STANTON:
MR. DAS:

This is good.

I just forget when it was.

MR. SEEFER:
MR. DAS:

Okay, thank you.

Anybody else?

I’m sorry that I’m disappointing you

that I haven’t given you any new stochastic calculus
equations and pointed to learned Nobel laureates.
MR. SEEFER:

So let me ask you this:

Our last

question -- God knows, we’ve been taking a lot of your
time.

Who would you suggest other -- to use the dreaded

word “experts” again in the field of derivatives that
you would suggest we talk to, to get, you know, more
opinions on how they contributed or did not contribute
79

FCIC Interview of Satyajit Das, May 7, 2010
or propagated the financial crisis?
MR. DAS:

Look, I’m very biased, so I -- the

purpose of my comments with that -- I think that three
or four people you would do well to talk to, to get a
plurality of opinions because I’m sort of out there with
Genghis Khan, I suppose.
But certainly the names that come to mind that
you could talk to Janet Tavakoli, who you may or may not
have spoken to.

She has her views and she has some

particular axes to grind.
I think the other people I would talk to is -there’s two other people, one you know of, I think Frank
Partnoy -MR. SEEFER:
MR. DAS:

Yes.

-- who you probably have spoken to.

Frank is trying to basically get the rating
agencies at the moment, I think, so he’s kind of
preoccupied with that.
There’s another guy called Robert Reoch in
London who would be interesting for you to talk to.
MS. NOONAN:
MR. DAS:

Reoch, R-E-O-C-H.

MR. SEEFER:
MR. DAS:

How do you spell his last name?

Thank you.

He runs a company called Reoch

Consulting.
80

FCIC Interview of Satyajit Das, May 7, 2010
The reason I’m giving you these names, is
these are people who have a fairly good knowledge of the
industry, actually practitioners, but are not allied to
any individual firm.
Because if you go to people in the firms,
you’re going to basically find it’s very difficult to
get a degree of honesty, I suppose, in the opinions.
That’s my comment.
The other people you could talk to, which
would be very, very interesting if you could get them to
talk, is some of the people who used to work at some of
the troubled firms who are no longer there.
First, there’s a guy called Robert Gumerlock.
I don’t know where he is these days.
G-U-M-M-E-R-L-O-C-K.

Robert Gumerlock,

He was the head of risk for

O’Connor Partners and Swiss Bank Corporation for many
years, who left after the merger.
So they would provide some interesting
information and perspective.
If I could give you one piece of unsolicited
advice.
MR. SEEFER:
MR. DAS:

Please.

Try to get an understanding of the

history, because I think there is always a desire to
look at the proximate events.

But look at how this
81

FCIC Interview of Satyajit Das, May 7, 2010
built up over a long period of time.
in a year.

This didn’t happen

This happened over 15 to 20 years.
And, if nothing else, if you can contribute to

some degree of understanding of this may -- and I stress
the “may” part of it -- prevent another one.

Because

whatever is going to happen next time will make whatever
happened this time look like a Sunday-school outing.
MR. STANTON:

Thank you for the reassurance.

MS. NOONAN:

We’ll sleep so well tonight.

MR. SEEFER:

That’s been the course we’ve been

on over the last -MR. DAS:

I’ll be dead.

I’ll be long dead, so

I don’t care.
MR. SEEFER:
MR. DAS:

No, I don’t have kids.

MR. SEEFER:
MR. DAS:

You don’t have kids?

Okay.

Well --

If you have kids, you have a

different agenda.
MR. SEEFER:

Yes, yes.

Or at least I have

kids.
Do you guys have -Me and Tom have kids.

The rest of them are

young’uns here, or at least younger than us.
Das, thank you very much for your time.

It’s

been a very interesting and enlightening conversation.
82

FCIC Interview of Satyajit Das, May 7, 2010
I’ve enjoyed it very much.
You know, as I tell everybody we talk to, I’m
sure after we go over our notes, particularly from this
conversation, we may have some follow-up questions.
Is it all right if we give you a ring back
after we go through them?
MR. DAS:

Yes, yes, yes.

MR. SEEFER:
MR. DAS:
reminiscing.

Okay, great.

I have nothing to do, so I like

I’m like the drunk at the end of the bar.

MS. SHAFER:

Das, there actually are a couple

of papers that you referred to.
MR. DAS:

Yes, yes.

MS. SHAFER:
MR. DAS:
forgotten.

I don’t believe we have --

Which ones do you have?

I’ve

I write a lot.
MS. SHAFER:

We have the Wilmont Magazine

piece, we have Derivatives, Tango, and Tranquilizers.
MR. DAS:

Oh, that’s right, because you

specifically mentioned, if I remember correctly, CDSs,
so I didn’t give you the other stuff.
That’s fine, I will -MS. SHAFER:
MR. DAS:

So I would read it.

You would read it?

I don’t know, if you actually bothered to read
83

FCIC Interview of Satyajit Das, May 7, 2010
the “Traders, Guns & Money.”
MS. SHAFER:

I have bought it, but --

MS. NOONAN:

Part of it.

MR. DAS:

Good.

my 10¢ of royalties.
or not.

As long as you buy it, I get

I don’t care whether you read it

It’s irrelevant.
MS. SHAFER:
MR. DAS:

[Unintelligible] the new one.

The new one, [unintelligible]

because it’s not going to be out until I think about -oh, I don’t know, it’s August or something.

But it’s

not new; it’s just got -- look, I’ll tell you the truth,
it’s not new at all.

It’s just basically got an

appendix, which basically looks at what happened.
Like all good authors, I basically am a
prostitute so basically I have the same book with a new
addendum, basically.
Okay, so I will send you the stuff.
you three or four papers.

I’ll send

So let me get this clear what

I’m sending you.
I’m going to send you a paper on,
“Mark-to-market.”
I will send you a paper on -- what was the
other thing that you -MS. SHAFER:

Leverage.

MS. NOONAN:

Embedded leverage.
84

FCIC Interview of Satyajit Das, May 7, 2010
MR. SEEFER:
MR. DAS:

Embedded risk.

Leverage, leverage, leverage?

I can

send you a page on leverage.
MS. SHAFER:

And Quantum Mechanics.

We have

that one.
MR. DAS:

You have that one.

I think you have

that one.
MS. NOONAN:

Quantum Hedges is what we have.

MS. SHAFER:

Someone’s holding out on me.

MR. DAS:

Sorry, sorry.

And then the other thing I will send you is, I
did a very lengthy paper somewhere on the lead-up to the
credit crisis and my diagnosis of the current crunch,
which shows some of the derivative links.

It might be

amusing for you guys, who don’t seem to have any life
other than to read these tedious pieces of drivel.

So

I will send that off to you.
MR. SEEFER:

Well, again, thank you very much.

MS. SHAFER:

You, sir, are among the most

amusing.
MR. DAS:

What was that?

MS. SHAFER:
amusing.

You, sir, are among the most

We can count on lots of fun quotes.
MR. DAS:

Oh, that’s fine.

That’s fine.

Look, anything you need, I’m happy to help you
85

FCIC Interview of Satyajit Das, May 7, 2010
in any way.
matter.

Because all joking aside, this is a serious

And if you can help along the way, I think I’m

happy to help you because this stuff is going to end up
costing people.

And, you know, people forget there’s a

human side to this.
And I’ll just finish up with a little story
which, you know, is not a very pleasant one; but it’s
important to remember.
About two or three years ago, I was in Tokyo
when the Argentines were doing their debt restructuring.
And, as you know, in countries, rescheduled debt, their
finance minister and so forth goes up and explains this
to the investors.

I happened to be in Tokyo at the same

hotel and they were there.

And I know one of the

investment bankers that were going around with them.
And he said to me, “Look, this is very odd.
They’re going to hold this investors meeting in this big
ballroom.”
And I said, “Why?

You usually have a little

room, a dingy room where three people turn up and yawn.”
But basically, it was -- and I went, because I
said to him, “Can I come?”
And he said, “Yes, of course, you can come.”
So I went along.
with Japanese, mainly women.

And I noticed it was filled
The average age would have
86

FCIC Interview of Satyajit Das, May 7, 2010
been over 70.

And the Argentine minister, who was a

little perturbed by this, got up and gave his usual
standard speech of how this would be good for Argentina
and the world economy and everybody.

And, by the way,

of the 100 cents on the dollar that they owed the rest
of the world, they were going to pay 35, and that wasn’t
going to be paid for six years.

The interest would be

deferred for six years, and so on, and it was wonderful.
At the end of the speech, an old woman, who
she would be well into her eighties, stood up and asked
a question which was translated, and the question went
something like this:

“I invested all my life savings in

Argentine bonds because they paid a higher rate of
return because that’s the only way I could continue to
afford to live, and was there any prospect that I would
get any interest and return of principal in my
lifetime?”
And I will remember that to my dying day.
So this is a serious matter, joking aside.

So

if you can’t do anything to prevent a repeat of this, it
would be because people who have lost their life
savings -- there are people in Hong Kong, several people
I know who committed suicide because of the loss of
their savings through Lehmans.

So this is not a trivial

matter, though I joke about it.
87

FCIC Interview of Satyajit Das, May 7, 2010
So if you can do anything, it would be much
appreciated; and I, for one, would be among your
greatest fans.
So on that cheery note, I will leave you to a
delightful weekend, and I will send you some reading in
the next ten minutes.
MR. SEEFER:

Thank you very much, Das.

MS. SHAFER:

Thank you.

MR. DAS:

My pleasure.

Take care now.
MR. SEEFER:

Bye-bye.

You, too.

Bye-bye.

(End of interview with Satyajit Das)
--o0o--

4822-2563-6870, v.

1

88