A Crack in The System

But in the end, neither the buyers nor sellers truly understood the enormous risks they were creating. Anyone could sell such a swap, and anyone could buy one, even if he had no stake in the transaction. Some buyers used them to bet against failing companies, prompting a debate among state regulators about whether this type of swap was a form of gambling.

The very nature of credit-default swaps put Financial Products at odds with itself, requiring it to deviate from the disciplined system that had made it a pathbreaker. Everything about the company -- its technology, its people, its rigorous culture of transparency and caution -- was designed to minimize the various risks that it shouldered while solving problems for clients.

That meant hedging whenever possible, a Wall Street term for making offsetting trades to balance risk. For transactions involving credit and loans, it also meant building an escape route so that the firm could get out early if it saw a deal going bad.

With credit-default swaps, there was no way out, and the risk was so minute that hedging was considered unnecessary, as well as problematic. Savage remembers discussions about whether the firm's vaunted computer system could even come up with the proper values needed for the trades that hedging relied on.

All of that made Savage and the others wary. Skepticism was hard-wired into the company's culture, part of its mantra: Hedge if you can. Don't make speculative trades. Above all, protect AIG's reputation and its top-drawer Triple A credit rating, which gave Financial Products credibility and the ability to borrow money at the cheapest rates. The rating was the fuel for Financial Products' innovation and success.

AIG's chairman, Maurice "Hank" Greenberg, had once warned Savage that he would come after him "with a pitchfork" if Financial Products did anything to harm AIG's AAA rating. No one saw credit-default swaps as anything on that scale. After conversations that included AIG executives, Greenberg blessed the new line of business. "There was a long discussion about it," Savage recalled recently, "and he said it was fine."

Greenberg said recently, "I don't think going into it in '98 was wrong." During his tenure, he said, he and his risk managers kept close watch on the swaps and the exposure they created.

Savage retired from Financial Products in 2001. When he left, credit-default swaps were still a small portion of the firm's business. Not long ago, in the dining room of his golf club in Florida, he reflected on the significance of the decision that he and his colleagues made in 1998.

Like his bosses at AIG, he still thinks it made perfect sense to give swaps a try. "The credit derivative business had just begun and because of our role in the derivatives business, it was very natural for us to have some minimal participation," he said.

Savage says he now sees that the decision sent Financial Products down a path at odds with its guiding principles. The firm's success had been built on assessing data daily, recalibrating assumptions constantly, counterbalancing one risk against another and making the hedges. The credit-default swaps didn't require that sort of attention.

"The different nature of those trades from any other trades that FP had done," Savage said, "opened the door to all the problems that came about."

He added later: "In retrospect, perhaps those deals should never have been done."