The chart goes in about two-thirds down this blog.
Sunday, April 12, 2009
9:52 AM
Robert Merton talk at MIT:
It's best if viewed and played back over and over. Even then it is confusing in part because he is using a light-pointer with his slides which does not show up:
Posted April 5, 2009 on
paul.kedrosky.com
Here are my notes from listening over and over to some extent:
[no comments are mine]
9:52 AM
Robert Merton talk at MIT:
It's best if viewed and played back over and over. Even then it is confusing in part because he is using a light-pointer with his slides which does not show up:
Posted April 5, 2009 on
paul.kedrosky.com
Here are my notes from listening over and over to some extent:
[no comments are mine]
Tools of financial science.
How does risk propagate so rapidly?
What are the implications to the inevidable incompleteness of all models?
Risk transfer technologies.
Pursuing comparative advantage.
Pursuing risk diversification.
Allows significant wealth particularly for smaller companies and countries.
Cause of finan crisis:
Refers to analogy: “if there were a body lying here.”
One person says: “heart attack”
Another says: “poison”
many possibilities as to why the man died
Will not know the answer until we do the pathology
We’re far from getting through the first stages of it (pathology)
coincidence of things, I think.
Not like Fineman and his O-ring (Challenger Disaster)
Decline in residential housing, began in 2006, before financial rumblings
Very big and very complicated.
I will take a little slice of something I know about.
Put options:
Started in 17th century Amsterdam;
Have been trading them on exchanges for 35 years.
Put option=Asset value insurance.
Plot a picture
Hockey stick
[graph is incomprehensible without seeing where his pointer is]
credit…is about loans, guarantees of loans
credit risk [not talking about interest rate risk]
If you attach a "guarantee" to risky debt = risk-free debt, which is just "time value of money."
What is risky debt?
Risk-free loan (us govt instru); writing a guarantee
When you lend to a risky counterparty, you are
Carrying out two activities:
1. Lending $ thru time
2. Self-guaranteeing (or writing a put option on it)
If volatility of the assets go up, the value of the put option goes up.
Credit default swaps:
CDS, CDO, CMO
AIG is not alone
Guarantee of debt of the issuer. Is all it is.
How is it that things go along smoothly.
Why is that they can come out with more and more losses? Shouldn’t it stop? Where’s it coming from? Hiding? Maybe, but not likely.
How it works:
Back to put option shape of the curve.
Banks make loans:
Which means:
· Lending $ thru time
· Self-guaranteeing
Which means:
1. lending $
2. writing put options on the assets we’re referring to
say, value of house;
value of guarantee bank has written
value of house falls, say
guarantee goes up; it’s worth more
“how sensitive is the value of the guarantee to the movement of the underlying asset, value of the house?” the slope of the curve; if .30 then it’s for each dollar in move of the dollar of the asset it’s 30 cents; .03 then it’s 3 cents
if this goes up what else goes up
for same movement of the asset it gets steeper; gets steeper and steeper;
most people don’t think about just a standard mortgage as having this property embedded in it.
when value of assets goes down their volatility goes up. The effect on the put option is even greater than what I showed you on that graph. In real world the risk is even greater.
As things go down the risk goes up, but most of the models out there tend to be local models that are multivariant regressions. But the problem is that they are local.
How does risk propagate so rapidly?
What are the implications to the inevidable incompleteness of all models?
Risk transfer technologies.
Pursuing comparative advantage.
Pursuing risk diversification.
Allows significant wealth particularly for smaller companies and countries.
Cause of finan crisis:
Refers to analogy: “if there were a body lying here.”
One person says: “heart attack”
Another says: “poison”
many possibilities as to why the man died
Will not know the answer until we do the pathology
We’re far from getting through the first stages of it (pathology)
coincidence of things, I think.
Not like Fineman and his O-ring (Challenger Disaster)
Decline in residential housing, began in 2006, before financial rumblings
Very big and very complicated.
I will take a little slice of something I know about.
Put options:
Started in 17th century Amsterdam;
Have been trading them on exchanges for 35 years.
Put option=Asset value insurance.
Plot a picture
Hockey stick
[graph is incomprehensible without seeing where his pointer is]
credit…is about loans, guarantees of loans
credit risk [not talking about interest rate risk]
If you attach a "guarantee" to risky debt = risk-free debt, which is just "time value of money."
What is risky debt?
Risk-free loan (us govt instru); writing a guarantee
When you lend to a risky counterparty, you are
Carrying out two activities:
1. Lending $ thru time
2. Self-guaranteeing (or writing a put option on it)
If volatility of the assets go up, the value of the put option goes up.
Credit default swaps:
CDS, CDO, CMO
AIG is not alone
Guarantee of debt of the issuer. Is all it is.
How is it that things go along smoothly.
Why is that they can come out with more and more losses? Shouldn’t it stop? Where’s it coming from? Hiding? Maybe, but not likely.
How it works:
Back to put option shape of the curve.
Banks make loans:
Which means:
· Lending $ thru time
· Self-guaranteeing
Which means:
1. lending $
2. writing put options on the assets we’re referring to
say, value of house;
value of guarantee bank has written
value of house falls, say
guarantee goes up; it’s worth more
“how sensitive is the value of the guarantee to the movement of the underlying asset, value of the house?” the slope of the curve; if .30 then it’s for each dollar in move of the dollar of the asset it’s 30 cents; .03 then it’s 3 cents
if this goes up what else goes up
for same movement of the asset it gets steeper; gets steeper and steeper;
most people don’t think about just a standard mortgage as having this property embedded in it.
when value of assets goes down their volatility goes up. The effect on the put option is even greater than what I showed you on that graph. In real world the risk is even greater.
As things go down the risk goes up, but most of the models out there tend to be local models that are multivariant regressions. But the problem is that they are local.
AIG
What’s happening: wrote a whole lot of credit default swaps. The assets underlying them have all been going down. Whoever’s written these guarantees in pure form. Not only are they losing money, but in increasing …
Things are not conceptually out of control. Not some mystery Black Swan.
People are not recognizing that non-linear curve. What they measure as a 10-sigma event is really a two-sigma event.
Govt writes a guarantee of…the bank’s liability…your deposits. They are writing a put option on them. Govt is writing a put option on a put option. Deposit insur on a classic bank.
So plot curve… starts with almost no slope at all. When that starts to go, the degree of the risk for the govt rises rapidly.
Especially problematic for small countries.
I’ve been talking about plain vanilla loans, vanilla guarantees. Nothing complex.
The good news [then he goes on without the good news; see infra for the good news]:
Innovation and crisis.
Plenty of bad people; plenty of incompetent people.
But there are also well-meaning and ethical people…still a problem.
Structual issue.
100 innovations; lucky if two are successful. Cannot be sure in advance.
Design of a high-speed train. If you don’t have a track you’d be crazy.
Trade-off: you never can go faster than the track. Not satisfactory.
We tend to feel more secure about the familiar than the new even though the risks are the same.
Defined benefit pension plan. Decision to invest in equities. Is equivalent in risk to entering into a total return swap.
GM would be $75 billion. One std devia = $15 billion.
Financed 100% by 30-yrs fixed rate debt. You’d sound crazy.
If you have housing come down you will these effects in the most vanilla of securities.
Only a crazy person would use a math model and then go off fishing (Clark?)
What is not going to change is that models are incomplete in describing complex reality.
Satisfy every constraint
Be totally ethical
Put that on with an incomplete model…will not be random. Assets will be highly pro-cyclical. Bias will be there.
Complexity is “so called” only. Special purpose vehicles (CDO) take… middle risky debt became CDO squared;
Take a garden variety corporation:
Mortgages, never change; stay there for the whole life
Corp:
Subsid
Intangibles
Etc.
On right side:
St debt pension lias
Etc.
So complexity is not at the core of the problem (which we are used to).
Wonder whether the govt would rather not see these valued.
What can we do about this bank problem?
Very smart, hard-working new administration.
They say “to be determined.”
How do anything without valuing the assets. If you take over the banks as in s&l you do not have to value them.
I see no other way to avoid “that.” What is the “that” that he’s referring to?
We’ve lost $15 trillion just in stock and residential mkts
87 crash similar perhaps; but for full yr came back; no internalized that wealth into their lifestyles.
I’ve relabled all my retirement as “supplemental income.”
Lehman was fully collateralized.
AIG was not.
Doing it on old-fashioned credit rating.
Accounting based rules; misaligned incentives.
Finan engineering got its start at MIT. Need has gone up. Need more who undertand what’s going on.
700 trillion of notional derivatives. It’s not going away.
“Regulators can’t understand.” Either learn how to use the technology …
Set up “Nat Trans Safety Board” just for the capital markets.
Sensible.
“Bad banks” maybe create a sovereign wealth fund.
Putting these assets into that fund to be run for the people of US. If it 20 years, so be it.
We don’t even know the assets that will end up there. Automobile instruments, etc.
Something positive:
We will get through this. One application of how you can use mkt proven. Taiwan as example. Computer chips. Comparative advantage. Return as an economy: part is result of world chip industry. Will he hurt no matter how good they are.
Taiwan-specific characteristics. Their thing. They control that.
What about the risk of the world chop industry.
Can get a portfolio that covers that.
Bec they are focused on one industry…
Contract: a swap
Two parties:
On a given amt we will swap returns;
Ttl rtn on world chip industry
I will get you rtn on all industries
Totally non-invasive
Taiwan now gets far better diversification but still doing just the chip industry.
Emerging markt on graph
6% difference extra; rule of 72.
How long it
Int rate/72
12 yrs to double your money
6 or 5 times bigger after 30 yrs; let’s halve it; pure risk management transfer efficiency
this works for small countries
it’s reversible; Taiwan enters into the opposite swap.
