The following excerpt is from Fatal Risk: A Cautionary Tale Of AIG's Corporate Suicide, a new book by investigative reporter Roddy Boyd.

The role of Goldman Sachs in AIG’s saga had its roots in a little- remarked-upon series of promotions involving a pair of managers known as the “J. Aron guys” taking control of Goldman’s Fixed- Income, Commodities, and Currency unit in the late 1990s. Gary Cohn and Lloyd Blankfein, veterans of Goldman’s sharp-elbowed commodities trading operation, saw a need to do things differently.

As they’d lay it out to the unit’s better producers in twos and threes, the firm was on the horns of a dilemma. In the looming post-Glass-Steagall landscape, “mega banks” like Citigroup and Bank of America (which had spent much of the past few years gobbling up Goldman’s competitors, big and small) would have the ability to throw around capital that Goldman would never have. Spreads were compressing, with margins immediately following, and whatever their vaunted relationships with clients had once been, they wouldn’t hold up in the face of Citi’s consistently being able to absorb a loss when it trades $5 billion of five- year Treasuries cheaper than anyone else.

Goldman didn’t need to change its business model; the marketplace had changed it for the firm. All that was required was to acknowledge it. So Cohn and Blankfein said Goldman should discard the way it had traditionally done its bond business, with an extensive focus on the largest 100 accounts according to assets managed and trading-revenue generation. From now on, Goldman’s bread-and-butter business plan was to compete to do every trade with everyone who rang up and then, after that, they would beat the bushes for more customers so they could do more trades. The bigger the better, of course, but every order was going to be fought for.

A certain group of longtime Goldman trading and sales staff were disgusted at the idea of becoming a glorified PaineWebber, in wasting time to give narrow bid and offer spreads on $1 million bond trades for a midwestern savings-and-loan or some $20-million-in-assets new hedge fund, many of whom might never call the firm again. Blankfein and Cohn would patiently meet with these people and reexplain themselves and lay out their reasons. A few weeks later, when various trading floor snitches reported back that the grumbling and politicking hadn’t stopped, these traders and sales staff found themselves having long midday lunches with Wall Street’s executive recruiters, exploring options at other firms, spinning tales of how they were happily leaving an ugly situation before it got much worse.

People like that, Blankfein and Cohn said, were just hard to reeducate. Another group of traders and salesmen who had joined the firm in the 1990s proved more willing to adapt. With only a passing connection to Goldman’s patrician past, they picked up much more quickly on what Blankfein and Cohn were trying to do. This wasn’t a bid to compete to get every trade per se, but a bid to get what every trade was telling you.

Who was buying what? What bonds were not moving and why? Where were people offsides? Who had conviction and who was sitting on the sidelines? And above all: why, why, why?

To get that information, you had to pay for it. The way you paid for it was in bidding or offering tons of bonds to customers you ordinarily could care less about. If the customer wouldn’t tell you directly, then you could piece it together based on what your desk and perhaps others were doing with them, or other customers like them. Then, armed with that information, they would be able to take the firm’s own capital and make some informed bets on a moment’s notice and make the real money.

In this formulation, Goldman was not to be the biggest trader, have the smartest people, or dominate any one market. But when it came time to take advantage of market moves, they would be there first and with their own money. Other firms might have bigger years and more dominant franchises, but no one would have a better return on equity. Since this was happening in the late 1990s, when Goldman was still a partnership, this was their own money at stake and return on equity was a key measurement. Blankfein and Cohn (and dozens more newly minted partners from the 1990s) were not terribly inclined to maintain a partnership in an era where even their longtime rivals at Salomon Brothers had sold out to Citibank to secure a more solid balance sheet. No, Blankfein and Cohn would push for a public offering and bring in some additional capital.8 Because, from where they sat, just about the entire bond world was evolving away from everything Goldman Sachs was, namely relationship- and client-driven and capital-at-risk averse.

Every day, in meetings in offices and on the trading floor, they drilled it home: margins were gone and they were not going to come back. Everything they did would have to become integrated: the growing prime-brokerage unit would open trading accounts for the growing number of hedge funds out there. Because they provided their own capital to these hedge funds, they had an instant customer base. Sales staff would have more clients to call, analysts would have more people to peddle ideas to, and traders could execute trades. What they lost in higher-margin business they would make up by “touching” the customer a dozen different ways. Things they had long hesitated to do—peddle derivatives en masse—they would do as markets became more integrated and ways to mitigate risk became more accessible. Above all, everyone was to hustle for that idea that had the big payday attached.

The way Blankfein and Cohn saw it, Goldman’s reputation as a repository of old-time investment-banking mores was a helpful asset that existed intellectually, in some vague, public relations type of way. In the world they had to live in, their customers were years removed from the white-shoe image of its past; all they honestly cared about was price and liquidity. To do that, with a competitive landscape that was getting more steep every quarter, was going to require a safecracker’s touch if they wanted to remain independent.