Global Policy Forum

By Joe Nocera

Next week, perhaps as early as Monday, the American International Group is going to report the largest quarterly loss in history. Rumors suggest it will be around $60 billion, which will affirm, yet again, A.I.G.'s sorry status as the most crippled of all the nation's wounded financial institutions. The recent quarterly losses suffered by Merrill Lynch and Citigroup — "only" $15.4 billion and $8.3 billion, respectively — pale by comparison.

At the same time A.I.G. reveals its loss, the federal government is also likely to announce — yet again! — a new plan to save A.I.G., the third since September. So far the government has thrown $150 billion at the company, in loans, investments and equity injections, to keep it afloat. It has softened the terms it set for the original $85 billion loan it made back in September. To ease the pressure even more, the Federal Reserve actually runs a facility that buys toxic assets that A.I.G. had insured. A.I.G. effectively has been nationalized, with the government owning a hair under 80 percent of the stock. Not that it's worth very much; A.I.G. shares closed Friday at 42 cents.

Donn Vickrey, who runs the independent research firm Gradient Analytics, predicts that A.I.G. is going to cost taxpayers at least $100 billion more before it finally stabilizes, by which time the company will almost surely have been broken into pieces, with the government owning large chunks of it. A quarter of a trillion dollars, if it comes to that, is an astounding amount of money to hand over to one company to prevent it from going bust. Yet the government feels it has no choice: because of A.I.G.'s dubious business practices during the housing bubble it pretty much has the world's financial system by the throat.

If we let A.I.G. fail, said Seamus P. McMahon, a banking expert at Booz & Company, other institutions, including pension funds and American and European banks "will face their own capital and liquidity crisis, and we could have a domino effect." A bailout of A.I.G. is really a bailout of its trading partners — which essentially constitutes the entire Western banking system. I don't doubt this bit of conventional wisdom; after the calamity that followed the fall of Lehman Brothers, which was far less enmeshed in the global financial system than A.I.G., who would dare allow the world's biggest insurer to fail? Who would want to take that risk? But that doesn't mean we should feel resigned about what is happening at A.I.G. In fact, we should be furious. More than even Citi or Merrill, A.I.G. is ground zero for the practices that led the financial system to ruin.

"They were the worst of them all," said Frank Partnoy, a law professor at the University of San Diego and a derivatives expert. Mr. Vickrey of Gradient Analytics said, "It was extreme hubris, fueled by greed." Other firms used many of the same shady techniques as A.I.G., but none did them on such a broad scale and with such utter recklessness. And yet — and this is the part that should make your blood boil — the company is being kept alive precisely because it behaved so badly.

When you start asking around about how A.I.G. made money during the housing bubble, you hear the same two phrases again and again: "regulatory arbitrage" and "ratings arbitrage." The word "arbitrage" usually means taking advantage of a price differential between two securities — a bond and stock of the same company, for instance — that are related in some way. When the word is used to describe A.I.G.'s actions, however, it means something entirely different. It means taking advantage of a loophole in the rules. A less polite but perhaps more accurate term would be "scam."

As a huge multinational insurance company, with a storied history and a reputation for being extremely well run, A.I.G. had one of the most precious prizes in all of business: an AAA rating, held by no more than a dozen or so companies in the United States. That meant ratings agencies believed its chance of defaulting was just about zero. It also meant it could borrow more cheaply than other companies with lower ratings. To be sure, most of A.I.G. operated the way it always had, like a normal, regulated insurance company. (Its insurance divisions remain profitable today.) But one division, its "financial practices" unit in London, was filled with go-go financial wizards who devised new and clever ways of taking advantage of Wall Street's insatiable appetite for mortgage-backed securities. Unlike many of the Wall Street investment banks, A.I.G. didn't specialize in pooling subprime mortgages into securities. Instead, it sold credit-default swaps.

These exotic instruments acted as a form of insurance for the securities. In effect, A.I.G. was saying if, by some remote chance (ha!) those mortgage-backed securities suffered losses, the company would be on the hook for the losses. And because A.I.G. had that AAA rating, when it sprinkled its holy water over those mortgage-backed securities, suddenly they had AAA ratings too. That was the ratings arbitrage. "It was a way to exploit the triple A rating," said Robert J. Arvanitis, a former A.I.G. executive who has since become a leading A.I.G. critic. Why would Wall Street and the banks go for this? Because it shifted the risk of default from themselves to A.I.G., and the AAA rating made the securities much easier to market. What was in it for A.I.G.? Lucrative fees, naturally. But it also saw the fees as risk-free money; surely it would never have to actually pay up. Like everyone else on Wall Street, A.I.G. operated on the belief that the underlying assets — housing — could only go up in price.

That foolhardy belief, in turn, led A.I.G. to commit several other stupid mistakes. When a company insures against, say, floods or earthquakes, it has to put money in reserve in case a flood happens. That's why, as a rule, insurance companies are usually overcapitalized, with low debt ratios. But because credit-default swaps were not regulated, and were not even categorized as a traditional insurance product, A.I.G. didn't have to put anything aside for losses. And it didn't. Its leverage was more akin to an investment bank than an insurance company. So when housing prices started falling, and losses started piling up, it had no way to pay them off. Not understanding the real risk, the company grievously mispriced it.

Second, in many of its derivative contracts, A.I.G. included a provision that has since come back to haunt it. It agreed to something called "collateral triggers," meaning that if certain events took place, like a ratings downgrade for either A.I.G. or the securities it was insuring, it would have to put up collateral against those securities. Again, the reasons it agreed to the collateral triggers was pure greed: it could get higher fees by including them. And again, it assumed that the triggers would never actually kick in and the provisions were therefore meaningless. Those collateral triggers have since cost A.I.G. many, many billions of dollars. Or, rather, they've cost American taxpayers billions.

The regulatory arbitrage was even seamier. A huge part of the company's credit-default swap business was devised, quite simply, to allow banks to make their balance sheets look safer than they really were. Under a misguided set of international rules that took hold toward the end of the 1990s, banks were allowed use their own internal risk measurements to set their capital requirements. The less risky the assets, obviously, the lower the regulatory capital requirement. How did banks get their risk measures low? It certainly wasn't by owning less risky assets. Instead, they simply bought A.I.G.'s credit-default swaps. The swaps meant that the risk of loss was transferred to A.I.G., and the collateral triggers made the bank portfolios look absolutely risk-free. Which meant minimal capital requirements, which the banks all wanted so they could increase their leverage and buy yet more "risk-free" assets. This practice became especially rampant in Europe. That lack of capital is one of the reasons the European banks have been in such trouble since the crisis began.

At its peak, the A.I.G. credit-default business had a "notional value" of $450 billion, and as recently as September, it was still over $300 billion. (Notional value is the amount A.I.G. would owe if every one of its bets went to zero.) And unlike most Wall Street firms, it didn't hedge its credit-default swaps; it bore the risk, which is what insurance companies do. It's not as if this was some Enron-esque secret, either. Everybody knew the capital requirements were being gamed, including the regulators. Indeed, A.I.G. openly labeled that part of the business as "regulatory capital." That is how they, and their customers, thought of it.

There's more, believe it or not. A.I.G. sold something called 2a-7 puts, which allowed money market funds to invest in risky bonds even though they are supposed to be holding only the safest commercial paper. How could they do this? A.I.G. agreed to buy back the bonds if they went bad. (Incredibly, the Securities and Exchange Commission went along with this.) A.I.G. had a securities lending program, in which it would lend securities to investors, like short-sellers, in return for cash collateral. What did it do with the money it received? Incredibly, it bought mortgage-backed securities. When the firms wanted their collateral back, it had sunk in value, thanks to A.I.G.'s foolish investment strategy. The practice has cost A.I.G. — oops, I mean American taxpayers — billions.

Here's what is most infuriating: Here we are now, fully aware of how these scams worked. Yet for all practical purposes, the government has to keep them going. Indeed, that may be the single most important reason it can't let A.I.G. fail. If the company defaulted, hundreds of billions of dollars' worth of credit-default swaps would "blow up," and all those European banks whose toxic assets are supposedly insured by A.I.G. would suddenly be sitting on immense losses. Their already shaky capital structures would be destroyed. A.I.G. helped create the illusion of regulatory capital with its swaps, and now the government has to actually back up those contracts with taxpayer money to keep the banks from collapsing. It would be funny if it weren't so awful.

I asked Mr. Arvanitis, the former A.I.G. executive, if the company viewed what it had done during the bubble as a form of gaming the system. "Oh no," he said, "they never thought of it as abuse. They thought of themselves as satisfying their customers." That's either a remarkable example of the power of rationalization, or they were lying to themselves, figuring that when the house of cards finally fell, somebody else would have to clean it up. That would be us, the taxpayers.