NPR's Morning Edition this morning reported on the causes of insurance giant A.I.G.'s downfall, letting us know that credit default swaps were "a $70 trillion business."

Reporter Adam Davidson says the credit-default swaps business soured and forced A.I.G. into an effective U.S. government receivership, which was announced late Tuesday night.

But consider: Does $70 trillion even exist?

Let's see if we can find out. We'll start by, roughly, dividing 70 trillion equally among the world's people. Estimates of world population range up from 6 billion to about 6.7 billion. Let's round up and call it 7 billion.

Written out that's:

70 trillion dollars: $70 000 000 000 000

divided by

seven billion people: 7 000 000 000

That amounts to $10,000 for every man, woman, and child on planet Earth.

As of 2003, the world's total income divided by the world's population yielded a per-capita total of some $5,500. (this web site explains the calculations--there are a lot of caveats here, but the principle of the thing holds).

Even assuming a few years' increase in economic growth, what this means is that the market in credit-default swaps alone--a single obscure corner of the world of high finance--accounted for nearly double the total income of the entire globe.

(cue Austin Powers clip--trying to ransom the world for $100 billion in 1969--"But, Dr. Evil, there isn't that much money in the whole world!")

How is this possible?

Obviously it's not possible, in reality. But in the world of high finance, "reality" is less important than the model one is using to manage risk. If realty disagrees with the model, reality is considered mere anecdote.

The $70 trillion figure is the "notional" number. That is, $70 trillion is the total value of all the contracts in aggregate. You may have a $200,000 mortgage, for instance, even though you earn only $60,000 per year. The credit default swap contract on your mortgage--(to over-simplify) the insurance policy your lender may have taken out against it--would carry a payoff of $200,000 in case you don't pay.

But what if someone besides your lender wanted to bet that you would not pay back your home mortgage?

Turns out they could do that, and companies like AIG would sell them that insurance policy. Since the odds of your default were long, the policy was cheap--a few thousand dollars. In this way, a contrarian investor could, with just a few grand, place a wager that might pay off $200,000. His contract was also valued at $200,000--even if he only put up $5,000 and even though your $200,000 house was already covered in the event of your default.

Now take the game one step further. Imagine that someone wants to bet against a lot of mortgage payers. They've figured out that the chances of things going bad are better than the insurance companies think. Buying a single credit-default swap wouldn't make much sense--the stakes are too small, the odds too long. So they would borrow billions of dollars and use it to buy millions of swaps.

(If you're an average Joe or Jane, borrowing money to gamble is called "pathological" and is bad. If you are a bond trader or a hedge-fund manager, it is called "leverage," and it's so good that if you don't do it, you cannot call yourself a financier).

Now they've got trillions of dollars in "notional assets" in play, even though the real underlying assets are worth a fraction of that. If they're right, they should be able to pay down their loans and walk away with billions in profit--or so they think.

But what if the insurance company--the House--can't pay off?

That's what's happening to the world economy today. A.I.G.--and lots of other companies, including banks, insurers, hedge funds, and brokerages--opened up a sort of betting window. But they didn't adjust the odds when the betting went against them.

Compare it to what a horse track does.

In 1991 I found myself on the infield of Churchill Downs at the Kentucky Derby, slurping mint juleps from commemorative glasses. Somebody named their horse "Corporate Report," and in the morning it was a field horse, carrying odds of about 100-to-one. But the Kentucky Derby attracts huge numbers of unsophisticated bettors, often with corporate-expense accounts. These drunken suits liked the name, so by post time Corporate Report's odds were something like 8-to-1.

That's how gambling usually works: The more people bet on a particular outcome, the more the house shifts the odds of that outcome against them. That way, the potential payoff never exceeds the amount bet.

But in the credit-default swap market, the odds never changed--or they changed too late.

So many clever people borrowed so much money to bet against people being able to pay their mundane little loans that the players overwhelmed the House's ability to pay off the bettors.

In a basic and real sense, we've broken the bank.