So, the financial questions keep coming. I'm avoiding a lot of them, because
(A) they bore me, and (B) I'm really not the right person to ask. I try to stay
out of this stuff unless I have some clue of what I'm talking about. Rest assured, I'm not spending all of my blogging time on this; I've got a post on cryptographic modes of operation in progress, which I hope to have time to finish after work this evening.
But there's one question that keeps coming in, involving the nature of things
like so-called "Credit Default Swaps", which I thought I'd explained, but
apparently my explanation wasn't particularly clear. So I thought I should fill
in that gap, and strengthen the main weakness in my earlier explanations.
The basic question is: "What's a credit default swap?"; I think what people
really want to know is both what, specifically, a credit default swap is, and how
the system surrounding credit default swaps and related monstrosities work.
Credit default swaps are interesting - in the same way that a Rube Goldberg
device is interesting. They are in a fundamental sense very simple, but the
structure that's built up around them is so bizarre, so ridiculous on the face of
it, that when you look at it in retrospect, it's hard to believe that anyone
actually thought that it was a good idea, or that it could ever work.
A credit default swap is something in between a gambling chit and
an insurance policy. To understand that, it's easiest to start by looking
at how these things come about. (The numbers in the explanation are made up, but they're not unreasonably far afield from reality.)
Suppose that I'm running the BigNosedGeek pension fund. It's incredibly
important that I keep the money in the fund safe. At the same time, it's important
that I invest the money intelligently, so that the money in the fund grows over
time, to ensure that I'll be able to pay the benefits that I've promised to all
the big-nosed geeks. I've got two opposing goals: I want to be safe, and the
safest investments tend to earn very little money; I want to make a decent return
on the money, but the investments with good returns tend to involve some risk. In
fact, that's deliberate: the reason that risky investments pay more is that they
need to justify their risk - to provide something to the investor to entice them
to buy the riskier investment instead of the safe one. In effect, you pay for
Now, suppose that this year, BNG pension has a billion dollars to invest. I
can invest it in federal government bonds for a return of 2%/year. That's a pretty
lousy return for a billion dollar investment. Instead, I decide that I want to
invest in mortgages, which will earn me 8%. That's much better. And
historically, that's a a very safe thing! Of course, while it's reasonably safe,
it's not guaranteed to be safe. So I want to protect my investment. I
want to buy insurance to cover my investment - so that if it goes bad, I
won't lose any of the money that's supposed to cover retired BNGs.
How can I insure a billion dollar investment in mortgages?
This is where the credit default swaps come in. In a CDS, you pay someone to
take on the risk of the investment failing. You're swapping the risk of credit
default with someone, in exchange for a payment. What you basically do is go to a
financial market, and say "I'll pay someone 2% per year if they'll cover my $1
billion investment in mortgages." If someone takes you up on that, you pay them
$20 million per year - and in exchange, they promise to replace your money if the
investment goes sour.
Everyone's happy. You've invested your $1 billion in mortgages, which will
earn a nice healthy return. Even after paying someone to take on the risk, you've
doubled the return on your investment. The guy who took you up on your offer is
making $20 million per year, for doing nothing, so long as the mortgages
are good. He's taking on a risk - he's going to be on the hook for a lot
of money if something goes wrong. But he's getting a lot of money for a minimal
risk. He doesn't even need to have the $1 billion on hand; he's just promised to
pay it if something goes wrong. He's got no money tied up in it - he's
just being paid! Everything is wonderful.
That's the idea of the credit default swap. Sell the risk of an investment to someone else.
In gambling terms, you can look at the CDS as a bet that the a loan is going
to default. You want to be covered in case the loan defaults; so to
protect yourself, you make a high-odds bet against your investment.
Then if the investment goes bad, the bet pays off.
A bookie wants to make money no matter what happens. If he's taking
bets on a boxing match, then he'll offer odds to different betters in
a way that producing a balance. Suppose you're looking at a fight between
Freddie the Fighter and Charlie the Challenger. How does a bookie set the odds
of the fight of Charlie verses Freddie? He looks at who's betting which way,
and sets up the odds so that no matter who wins, he'll have enough money to
pay the people who bet on the winner, with some left over for himself. If
the fighters are very evenly matched, and the bets are running even, then
he'll give equal odds: bet 1 dollar for charlie, and if you win, you'll get
back an extra dollar. If Charlie is really a hobo who Freddie's promoter hired
to take a fall to extend Freddie's undefeated record, then the odds are
going to run very heavily against Charlie: betting for Freddie to beat Charlie
could only pay off $0.001 for each dollar bet, whereas betting $1 for Charlie
would win $1000 if he won. The bookie is going to manage the odds that he's
giving betters to mantain the balance.
Credit default swaps are a form of bet, and they're working inside of a
market, which acts as a sort of headless bookie. The market effectively sets the
odds in a way that strikes a balance. A credit default swap is, basically, a bet
that a particular loan will default. So if I'm bank A, and I'm loaning a billion
dollars to bank B, I want to be sure that I'll get my money back. One way of doing
that is, basically, to place a bet. I bet that bank B is going to default on the
loan. Since B is very unlikely to default, the odds on that bet are huge - I can
bet one million dollars that they'll default, and if they do, the bet will pay off
all $1 billion.
The problem with this is that no one is going to take the other down side of
the bet. On one side, you've got someone who wants insurance against an exceeding
unlikely event - so they're willing to put up some money, effectively betting on a
very unlikely outcome, but with huge payoff odds - things like better $1 million
that a bank will default, expecting to lose the million, but with the proviso that
if the bank defaults, they'll get a $1000 to 1 payoff. But on the other
side, you're talking about putting $1 billion to win a paltry $1 million. That
makes no sense at all - no one's going to put up a billion dollars for a return of
one tenth of one percent! To try to get around this, the people who set up the
market sweeten the deal in two ways.
First, if you take the bet against the bank, the money that gets bet is yours.
So if the other guy bets $1 million that the bank will default, you get a million
dollars up front.
Second, if you take the bet against the bank, you don't have to put up the
money up front. By taking the money bet by the other guy, you're making a
commitment to pay up if the unlikely event occurs, but you don't need to pay up
front. So, you get to take the money bet by the other guy, and you don't need to
tie up your own money. Odds are, that's a damned good deal. You're getting money
for doing nothing.
Looked at in gambling terms, the CDS looks pretty much like a scenario where
you go to the bookie, and say "I want to bet on the champ defeating the hobo"; and
the bookie just gives you the money bet on the hobo winning, and takes
himself out of the picture. If the hobo does win, you're holding the bag to to pay
off everyone who bet on the hobo at huge odds. The bookie didn't make sure that
you had enough money to pay off the bets if the hobo won. The bookie doesn't
really care; he's not losing anything. If the bettor can't come up with the money,
it's the other bettors who won't get paid. The bookie is just an agent
for connecting up the betters on the two sides; the question of who's going to be stuck paying the betters isn't his problem.
Just going this far, it should be obvious what can go wrong. What if the
investment goes bad, and the guy who took your swaps can't pay up? There's no real
guarantee here: just an agreement. The guy who took the swaps doesn't have to
prove that he's got some plausible way to come up with the money! The swaps
market is private, with no regulation. People in the market trust each other
because it's profitable (in the short run) to trust each other. But there are no guarantees. All it takes to buy up a collection of CDSs is enough money
to buy into the market. Not necessarily enough money to pay off the CDSs you buy - no one checks that. All you need is the money to buy a position
in the market.
But as is all too common in situations like these, it gets a lot worse.
What if you believe that the mortgages are going to go bad, and you want to
make money on it?
Just like there's no guarantee that the guy who accepts the credit default
swap will be able to pay up, there's no guarantee that the person
offering the swap actually has the investment that it covers! I don't
have to have $1 billion worth of mortgages to make a deal to buy $1 billion of
credit default swaps. The swaps are completely decoupled from the instruments that
they purportedly were created to ensure.
So you wind up with things really degenerating down to the gambling scenario. The CDS isn't just an insurance policy to protect an investment. That might have been the intention when it was invented, but that's no longer true. Now it's
a true bet: anyone who thinks that a loan might default can place a bet on it
happening. Anyone who thinks that a loan won't default can place a bet on it. And this gambling system goes beyond just loan defaults. There's a whole system of
contracts which started as pseudo-insurance, and turned into gambles - for example, if you're worried about the US dollar decreasing in value, there are
derivatives which are similar to CDSs which are tied to changes in the exchange rate of the dollar relative to other currencies. So you've got a very complex
system which is really a gigantic, unregulated, unverified gambling arena.
To make matters worse, bets in the CDS market are treated as assets. That is,
if you've got a chit showing that you took a CDS that paid $1 million/year, that's
treated as a real asset worth $1 million. So you can use the bet as collateral for
loans outside of the CDS market - and then those loans are guaranteed by
other CDSs, which then become assets, which can be used...
And of course, no one could have predicted that that would be a disaster,
The other question that people keep sending me relates to the fact that I've made it clear that I'm in favor of regulation. The question is "How could regulation have prevented any of this without totally gumming up the market?"
To some extent, you can't. But that's not necessarily a bad thing. One way of
looking at the current chaos is that there's a huge amount of stuff in the market
that's built on sand. There's a whole lot of fake wealth - stuff that's created by
piling up levels of "wealth" that are based on absolutely nothing at the bottom.
Being paid for a CDS which is based on a loan collateralized with another CDS,
which is based on a loan collateralized by another CDS - that's not really creating wealth. That's just creating an illusion of wealth, which can be used as a tool for tricking people into thinking that you've got something
really valuable. If the CDS market were regulated, and you couldn't take
on a swap without demonstrating that you had a genuine ability to pay it off
if it went bad, then a lot of the speculative stuff that drove a huge amount
of economic activity would never have occurred. I would argue that
it was fraudulent economic activity, and that the fraud that drove the markets
in things like CDS shouldn't have been permitted. If you exclude that kind
of massive fraud, then a lot of economic activity goes away - and with it,
you would see a reduction in money available for borrowing (both by individuals and by businesses), you would see a reduction in business revenue growth, and a reduction in government revenues. But in the long run, the fact that worthless
stuff is worthless is going to come back to haunt you - if you tolerate
the fraud, then you'll see nice apparent growth for some period of time, but
eventually, it's going to crash and burn.