As presidents often note in their memoirs, every major decision that arrives at the Oval Office is difficult, filled with imponderables and inconsistencies. Otherwise, it wouldn't hit their desk. But not since Franklin Roosevelt has a president had to face the twin crises, inextricably linked, of the economy's collapse and the rescue of the U.S. financial system.

Despite Obama's clear expression of his will about the primacy of health care reform, most of his senior advisers were in agreement with Orrin Hatch-and not only about someone such as Tim Geithner showing up at the Health Care Summit. Their underlying doubts were about Geithner's boss, the president, and whether he should be conducting such a summit now at all. The fear, growing inside the uppermost reaches of the White House through late February, was not just that this was the wrong ordering of priorities, but that it might ensure that none of the three great battles could possibly be done well, a concern that seemed to be quietly stoked by this long day of performances to launch the health care campaign.

Krueger, like most other senior officials, was happy to get back to what he considered more pressing business. At 5:00 p.m., in the midst of Citigroup's woes, the NEC wrestled with "too big to fail." Taken together, the a.s.sets of the largest six banks, which included Citi, were now a stunning 60 percent of the country's overall GDP, significantly more concentration than before the great panic. The issue of systemic risk, of how these still-fragile inst.i.tutions were linked to one another, was all but impossible to fathom. And no one doubted that if the economy were imperiled by the failure of one of the largest banks, the government would be hard-pressed not to step up for a rescue.

Which is precisely why Tim Geithner came late to the meeting: more trouble at Citi. He was in crisis-management mode. He had to excuse himself early: another call to Bernanke.

Meanwhile, Austan Goolsbee, who was trying to revive some of the spirit of reform that was abandoned after the campaign, pushed a proposal about a tax on bank size: that the big banks start getting taxed on a.s.sets above a certain threshold. If the so-called externality-econ-speak for the side effect of a company practice-"is size," Goolsbee said, "then you tax it, and it shrinks away."

Members of the legislative affairs team, sitting in, were enthusiastic. Congress, they noted, was looking to forcefully engage with the issue of preventing more bailouts and the threats of systemic risk. This proposal also raised revenue, which meant it was doubly saleable!

Summers listened to Goolsbee-no threat in his diminished role-elegantly sketch out the sort of market intervention he tended to oppose, as did Geithner. But Geithner was gone. Romer and Krueger were with Goolsbee, discussing the mechanics of various taxing techniques. The key with any such intervention was to structure it soundly and tightly, so it did what it was intended to-never an easy task. Of course, the political folks, the legislative team and Rahm, would view intended outcomes through a shorter lens: something that could be pushed through Congress and look like sound policy. Most important, there was the president-not in attendance for this meeting, but still arguing gamely through the end of February for bold action of some kind, still pressing his case that would have America acting more like Sweden than like j.a.pan.

Based on their long history together, and general agreement on principle, everyone in the room felt that Summers-even as he kept his cards close-was in Geithner's camp.

Geithner, though, was quietly beginning to worry by early March that it might be otherwise. What he saw gathering through February was what his deputy, Lee Sachs-a former Treasury official under Clinton who'd worked at Bear Stearns and then ran a hedge fund-later called an "unholy alliance between the hedge funds and the academics, who were all now calling for tough measures on the banking system." Sachs, who'd been brought in during the transition to head "crisis management" of the financial meltdown, had created models to show how government intervention would drive down the already low price of the toxic a.s.sets. The fact that the toxic a.s.sets were difficult to value didn't mean that, if pressured to, the market wouldn't come up with a price. In a market with few buyers, it would be a low price, making the "hole" the federal government was looking at even deeper. Several reliable estimates of the amount of toxic a.s.sets across the banking system put the figure above $2 trillion. "We realized early on that a two-trillion-dollar hole was more than we could fill with the $350 billion left in TARP," said Sachs. "We were going to need to draw in private money with incentives and guarantees that we knew would make us look like we were in the pockets of the banks."

While Romer was talking to the academics, Summers was on the phone to the hedge funds, many of which had built up significant short positions on bank stocks. Any federal intervention into banking would drive down bank stocks. The shorts would clean up. Though Summers would surely know this, the more worrisome issue was the case the hedgers were making about how the government could force the kind of efficiency and shakeout that Summers felt the banking industry needed. In other words, the banking industry-like everyone else-should get what it deserves.

A few days after the Health Care Summit, Summers made his move at a briefing with the president. His "first, do no harm" test had been satisfied, he said. He joined Romer in support of the president's belief that a major federal intervention into the banking system was now needed.

Geithner pushed back.

"The confidence in the system is so fragile still," he said. "The trust is gone. One poor earnings report, a disclosure of a fraud, or a loss of faith in the dealings between one large bank and another-a withdrawal of funds or refusal to clear trades-and it could result in a run, just like Lehman."

Geithner tamped down frustration. Romer, Summers, and even the president couldn't understand what he and Bernanke had lived through-the nights of sleepless panic, terrified phone calls from once-unflappable bankers, secretaries standing in the street holding boxes with paperweights and framed photos. He thought it was unwise for the government to pick a troubled bank and dissolve it, a precedent that would create fear and undermine confidence, rather than promote it.

The president, however, seemed undeterred. In fact, he was enlivened: Summers was now on his side. It wasn't consensus, but it was close. "I think it's time to step up and show what government can do," Obama remarked. "I want to deal with these toxic a.s.sets across the entire banking system. Let's do it now, let's do it right."

In certain ways, Obama was reaching for what senior advisers had begun to call that "rare combination" where the president decided that a sound policy was also politically advantageous. When the two came together, Obama acted. His words of anger at Wall Street had not been followed with actions. But now a tough-love approach to the banks-much like what Volcker had talked about with Obama in the months after the Cooper Union speech-could show his words backed up with action.

There were general discussions about how much it might cost: another $500 billion, maybe more. "We'll find the money, somewhere," Obama said. "When you have a crisis, you find the money." Obama mentioned what everyone already knew: that in February they had put a placeholder of an additional $750 billion in the proposed budget for further government interventions into the broken financial system. The Congressional Budget Office had already "scored" the cost of any such allocation at $250 billion (under a calculus that $500 billion of that money would eventually be returned) though no one was antic.i.p.ating that this just-in-case budgetary "placeholder" was slated to be filled.

No one, now, except Obama.

Geithner, meanwhile, said that many of the president's desires for action could find a home, at much lower costs, in his "stress tests," the planning for which were well underway. Geithner's team at Treasury had been working on the structure of the stress tests, in conjunction with Bernanke, since before Geithner mentioned them in his nightmare early-February press conference. They would empower government to a.s.sess the health of the large banks over the next few months, almost the way a rating agency would, and then tell the banks how much more capital they needed to continue as going concerns. The government could then decide whether to give a bank a cash infusion or to take it down, with a ratings system that the markets would consider credible. The question on the table was complex. Should they wait for the results of the stress tests-which Summers and Romer doubted would be credible-and then decide whether, or how, to take down a few banks that were troubled and unable to raise capital? Or should they move more preemptively, taking several large troubled inst.i.tutions through "resolution"-a term that implied a controlled bankruptcy and brief government takeover-sooner rather than later? Either way, the president was interested in thinking creatively about how to take down some of the nation's largest banks.

Obama listened. "Okay, we should work this out," he said. "Why don't people pull together their proposals."

Geithner left the meeting incensed. Larry had no idea what he was doing or whom he was up against. A meeting was set for Sunday, March 15, in the Roosevelt Room. That meant the teams from Treasury and the White House would have just a week to pull together their presentations. Staffers in the two buildings immediately started calling it The Showdown.

On the afternoon of March 9, Sheila Bair girded herself for the next conference call. It was almost one a day-she would be the only woman on the phone with an army of men, many of them with close ties to Wall Street or an unshakeable belief in the miracle of the markets, the freer the better.

And she would be the scourge.

Tension between Bair and the men managing the town's other regulatory warships was rapidly looking like a redux of the battles Brooksley Born fought in the late 1990s with the fraternity of like-minded regulators allied with Wall Street over derivatives regulation.

If there was one difference, it was that Born had been alone. Now there was a small but powerful contingent of the sisterhood, and a gender battle, long simmering just beneath the surface of cordial relations among regulatory colleagues, was finally starting to draw notice. With Born, now a Washington lawyer, as their inspirational hero, a team of women-led by Bair; Mary Schapiro, chairwoman of the SEC; Elizabeth Warren, heading the TARP Oversight Panel; and the irrepressible Maria Cantwell-was a.s.serting its primacy. They had virtually all been right, and right early, about the way America's financial system was drifting toward crisis. All of them had been shooed away or shouted down by the men, both those manning Wall Street and those atop Washington's regulatory or economic policy posts, who quietly a.s.serted that high finance might be the final mountaintop stronghold of "man's work."

While 58 percent of college undergraduates are now women, and many of the most prized professions and skill-based industries are approaching gender equality, virtually all the top posts on Wall Street and at the largest national banks have long been held by men. Though most of the men won't say it, they feel that the nexus of math and risk-and the gaming of both, without flinching-is an area of male inclination. In fact, many of the women agree. They say that's part of the problem.

Few could, at this point, challenge the idea that the country's male-dominated financial industries had powerfully self-destructive impulses. But Geithner was just the latest in a succession of regulatory men, many with a past (or a bright future) in managing money and risk, who felt the town's few female regulators often didn't understand them or the way Wall Street's male Mecca really worked-knowledge that is crucial to being an effective regulator who can alter ruinous behavior.

The women's response, of course, was that they understood the men better than the men understood themselves.

History's judgments, of late, seemed to be bending toward the ladies.

Bair, for one, was not bashful about pointing out precisely where she'd been right, across nearly three decades. A Kansas Republican who spent most of the '80s working both in campaign and senior staff roles for that state's avatar, Senator Bob Dole, Bair was named one of three commissioners of the CFTC in 1991 by the first President Bush. In the deregulatory environment of that period, Bair-who was once a bank teller in Kansas and waxes nostalgic about kids' opening pa.s.sbook savings accounts and the pride people felt in meeting their obligations with each month's mortgage payment-took her first turn as skunk at the garden party. She was especially skeptical of a fast-growing Houston-based firm called Enron, a diversified energy company that was pressing the CFTC to exempt what the firm called its "sophisticated" futures contracts from antifraud provisions, a move that would have shielded Enron's burgeoning exchange-trading business from CFTC oversight. Bair, voted down 2 to 1, offered a scathing dissent: "If we are to rationalize exemptions from antifraud and other components of our regulatory scheme on the basis of the 'sophistication' of market users, we might as well close our doors tomorrow."

When, in 2001, Enron's trading business was exploding into a historic fraud-a harbinger of the derivatives disasters to come-Bair had little time to gloat. At that point, as a.s.sistant Treasury secretary for financial inst.i.tutions, under Bush, she was intensely interested in the growth of "nontraditional lenders," free-floating finance companies, funded by Wall Street speculators, that were offering loans with low "teaser" rates and hidden fees. What struck Bair was that these subprime lenders generally had responsibility for the loans for only ninety days-three months of payments-before the traditional fiduciary bond between lender and borrower dissolved and the loan was "securitized" and sold off to other investors. Bair sensed trouble along many links of this chain, but found that the defaults by borrowers who were encouraged to take out larger mortgages than they could afford were lower than she expected. That's because they were constantly refinancing, at ever lower rates, and often using the proceeds for general purchasing. Her concerns that this couldn't last, and would end badly, were drowned out in the naysaying of Alan Greenspan, his cheap-money policies, and the rising real estate values that were fueling wider consumption.

It wasn't until 2006, though, when Bush unexpectedly selected her to run the FDIC, an independent agency whose director serves a five-year term, that Bair found the freedom to be . . . just Bair. Having spotted early troubles in both the derivatives and subprime markets-and then launching flares that were ignored-she could now be an independent actor. And act she did. She a.n.a.lyzed all the subprime data the FDIC could buy and closed one of the most egregious subprime lenders, the California-based Freemont Investment and Loan, in March 2007. Seeing a wave of defaults on the way, especially as tens of billions in "teaser-rate" loans readjusted upward, she pressed the banking industry to restructure the mortgages, which would make more of them sustainable, even as it shrank the banks' profit margins from the often onerous rates. The banks said they would, but didn't. She unloaded on them at a mortgage industry conference in October 2007: "Moody's recently reported that less than one per cent-less than one per cent-of subprime mortgages that are having problems were being restructured in any meaningful way," she implored them. "We have a huge problem on our hands . . . I think some categorical approaches are needed, and needed urgently."

The fact that no one budged and disaster soon reigned only increased her ire, especially at Citigroup and Bank of America, the industry leaders, which she felt exhibited anything but industry leadership, especially when they should have known better-after all, she herself had warned them of what was ahead.

But by the fall of 2008 she found herself rushing into a place where regulators rightly fear to tread: cutting deals to buy and sell banks, especially in a volatile market where share prices could drop from respectable to abysmal on an errant rumor. The specific case that snagged her was the sale of the failed Wachovia to Citigroup, a transaction, requiring government a.s.sistance, that she and Geithner provisionally approved in late September. But when Wells Fargo arrived with a richer offer in November, and one not needing federal a.s.sistance, she opted for Wells. Citi's stock summarily plummeted, along with its overall capitalization, pushing it into the arms of regulators and summoning the fierce disdain of a vast community of Citi officials, past and present, from Bob Rubin on down. Bair demurred that she couldn't stop the Wells deal-it was better for Wachovia, an appropriately arm's-length transaction that didn't need help from the government. But former Clinton-era regulators with net worth in Citi stock, many of them now cycling back into the Obama administration, were incensed.

Bair, they cried, just didn't get it-didn't understand how the world, resting on projections of confidence, really worked. All she could talk about was tier-one capital, and how things used to be in the sleepy 1970s. In fact, Sheila Bair, who'd been around long enough to have used Paul Volcker as her role model, had little respect for the "we're all in this together" bond built across three decades between Washington and Wall Street, a relationship of shared interests in which Citi, like Goldman, was a central actor. Her positions on key issues such as shrinking banks to make sure they weren't "too big to fail" and curbing Wall Street's excesses were generally aligned with Volcker's, and her criticisms of Vikram Pandit and Citi's current management were specific and pointed. She thought both should be replaced, and said so publicly.

Geithner's response to a deputy at Treasury: "She keeps up that kind of talk, we'll have a run on Citi-then, I suppose, she'd finally be happy."

If Sheila Bair had especially strong feelings about Citigroup, she had her reasons. Bear Stearns had been rescued, Lehman had failed, Goldman had gamed everyone it met, JPMorgan had avoided the worst of mortgage-backed h.e.l.l, as much by good luck as good management, but Citigroup was in its own special category. There was, after all, no bank that embodied past disasters and future risks like Citi. It essentially invented the concept of "too big to fail."

Anyone with a desire to understand banking in America need only follow the two-hundred-year arc of this inst.i.tution, from its start in 1812 in New York to the $2 trillion behemoth that collapsed in 2008, with three hundred thousand employees, two hundred million customer accounts, and operations in one hundred countries.

Citigroup, under its previous name, National City Bank of New York, was the country's largest bank for much of U.S. history, and had been bailed out by the government many times.

Not that the bank didn't pioneer innovations, including checking accounts, negotiable CDs, unsecured loans, compound interest on savings accounts, and, of course, ATMs. It also was an innovator in the 1920s in creating the disastrous investment trusts, that era's CDOs, that were at the center of the 1929 stock market crash and all but prompted Gla.s.s-Steagall so that banks, with depositors' money, would never again operate as investment houses.

While many smaller banks failed, National City Bank was propped up by FDR, as were other large banks, for fear that the overall system would collapse without them. But it was more than that. The bank, which was pilloried, along with its CEO, was always seen as a representative inst.i.tution-what it did, or what was done to it, would serve as an example for others. No doubt other banks were ever attentive, following Citi's lead as it invested in Latin American debt in the early 1990s (another government bailout) and in the late '90s, as it was growing ever larger with acquisitions and mergers, culminating in the 1998 merger of Citicorp and the Travelers Group. That union, orchestrated by Sandy Weill and his deputy, Jamie Dimon, created a huge financial organism that provided virtually every function in the management of money and risk-from insurance to brokerage services, from investment banking to plain vanilla commercial paper, and every conceivable trading activity. Under one umbrella were brand names galore: Primerica, Travelers, Salomon Brothers, Smith Barney, Commercial Credit, with everything stamped "Citi." More than sixty years after National City's behavior helped prompt Gla.s.s-Steagall, Citi's merger mania helped finally kill the already eroded separation between commercial and investment banks. Everyone, then, could be like Citi, and other banks didn't disappoint.

Not that this slaked Citi's acquisitive thirst. Since 2003 the bank bought four credit card lenders and five mortgage lenders, ballooning up to $2.2 trillion in a.s.sets by 2007, roughly even with JPMorgan Chase and Bank of America-each, itself, a buffet of services and functions, if slightly less varied than Citi.

And starting in 2004, Citi did what almost everyone else did: load up on CDOs, holding nearly $60 billion of them on its books by early 2008.

The key was that the bank, after so many bailouts, was always seen as "too big to fail," and took advantage of every feature that this designation provided, from a lower cost of credit to regulatory favoritism to a "might makes right" lat.i.tude in its all but indecipherable web of interlocking businesses. That's what tends to happen, after all, at this size: the ent.i.ty becomes impossible to manage. Sometimes banks end up on the right side of large market shifts, but often not. To be sure, the way the markets glorified the prowess of Jamie Dimon by the spring of 2009 was how they'd once felt about Sandy Weill.

Citigroup, for its part, was haphazardly managed, going through four CEOs in just under a decade, with the last being Vikram Pandit, who oversaw inst.i.tutional investments and trading at Morgan Stanley, and then ran a hedge fund, before becoming Citi's CEO in December 2007.

Pandit accepted a government check from Paulson's Treasury the next fall-for $25 billion, like the other large inst.i.tutions-but distinguished himself and his bank by returning just a month later, on November 24, for another $20 billion. More important, that same day, Treasury guaranteed $306 billion of Citigroup's a.s.sets. It asked little in return for any of this-no management changes or restructuring. Just some warrants and preferred stock. This guarantee, so-called ring-fencing, allowed Citi to keep its enormous pile of nonperforming and illiquid a.s.sets-mortgage-related a.s.sets and a sinking, toxic haul of credit card debts-on its balance sheet and, with this government support, retain the illusion of solvency. Later, when pressed on this, Geithner cited the reason for this government largess, according to a report by TARP's inspector general, "to a.s.sure the world that the Government would never let Citi fail."

Four months later, Sheila Bair was stressing that the government should now be sending the opposite message: destructive behavior could still, in some instances, draw a death sentence.

In the conference call with the country's top financial regulators at 3:00 p.m. on March 9, she stated her case, as she had on several such calls over the past week. Last fall's ring-fencing was insufficient. The $306 billion wasn't enough. In the ensuing months, Citi's credit card defaults were rising, while the value of its toxic mortgage-related a.s.sets continued to drop. What had been on its balance sheets, after all, was not even the entire mess: many of the toxic CDOs were off the balance sheet, held in SIVs, structured investment vehicles, another Citi innovation from the late 1980s that had spread across the industry.

FDIC a.n.a.lysts who'd examined the bank put the toxic load at roughly $600 billion out of a total of $1.6 trillion in a.s.sets. This figure, however, took into account the "intrinsic value" of the mortgage-related fare, rather than the harsher mark-to-market standard that everyone, everywhere, was ducking-and not without some justification. In the wildly oversaturated real estate market, even solid mortgage-backed a.s.sets would have trouble drawing a depressed price. The intrinsic standard accounted for some modest stabilization of the market at some point in the future.

Many of those on the conference call-including John Dugan, the lead banking regulator in his role as head of the Office of the Comptroller of the Currency, or OCC; Bill Dudley, Geithner's replacement as head of the New York Federal Reserve; and Ben Bernanke-were fearful that this "give" on the valuation by FDIC was a trap. Geithner, leading the call, felt "the markets wouldn't respond well" to the intrinsic standard. Bair disagreed: it was eminently defendable, would help inst.i.tutions get out of their bind of not being able to fully recognize the toxic loads on their balance sheets, and, of course, it was from the FDIC, not known for its charitable view on such matters.

If, hypothetically, that $600 billion number were accepted, and Citigroup moved into some form of "resolution," then, Geithner and others a.s.serted, the FDIC would be on the hook for the whole amount. This game of brinksmanship had been another plotline of the conference calls: fine, if the FDIC wanted to take down Citi, it would have to bankrupt its own accounts to do it. The FDIC, which is supported by a t.i.thing from the commercial banks it insures, would be overwhelmed by the cost and have to appeal to Congress for, well, a bailout.

That was not the kind of resolution Bair envisioned. Now rubber was. .h.i.tting highway. The FDIC's specialty, of course, is shutting down banks. It's been at it, over umpteen weekends, since the 1930s. Citi was huge, but its core was still a bank, and should be treated as such. That meant a prepackaged bankruptcy: the bank would be shut down; management thrown out; equity holders wiped out; troubled a.s.sets moved to a "bad," or aggregator, bank; and a smaller but clean "good bank" would emerge, essentially a new ent.i.ty that could accept investments and get on with business. The key to the equation was that, as in all bankruptcies, creditors would take a haircut. In this case it would be for a few hundred billion, which would lighten the amount of FDIC money that would be required.

This was the key to the equation-to so many equations in this period, where debt had become sustenance and its purveyors on Wall Street the richest community in human history. Geithner, on this point, would not budge. Debt was sacrosanct. No creditor would suffer. Bair was equally intransigent. Secured creditors, such as equity holders, of course, wouldn't be wiped out, but they had to face consequences for lending money to an inst.i.tution whose recklessness had led to its demise. They must, she said, "face some discipline." Her point was that ultimately this would be seen as progress-as when someone finally got a needed operation-which would begin to restore long-term, sustainable confidence in the financial system. Debt was underpriced. Once it was priced properly, the healing could begin.

Bernanke said little. He'd spent more than a year opening the Fed's coffers to guarantee anything that moved to ensure that traditional market corrections, corrections in the pricing of risk, would not commence, at least not yet. Not that his efforts were resulting in the desired easing of credit by banks and other financial firms being supported by Fed dollars. They were just making money from the free money offered by the Fed, and sitting on the profits. Why, after all, would anyone lend when demand was zilch and America was overleveraged, stem to stern? Once the economy ticked up-maybe then. Of course, that couldn't happen unless credit began to flow.

Dugan, a holdover from Bush, whose OCC regulators for years visited banks and, generally, did not make a peep as the banks loaded up CDOs, said he was concerned about "how the markets will respond" to any actions against Citi.

Around they went, gridlocked. Finally, Bair said that at the very least Pandit must go. He was essentially a fixed-income trader and a hedge fund manager. "We need a commercial banker at the top of this bank-someone who knows the business of banking. That's one way to maybe get some lending started. It is mostly a bank, after all."

In terms of some accountability for reckless actions, Bair considered this a starting point. Geithner wouldn't entertain even this fallback position. The government exchanged its $45 billion in direct aid and $306 billion in guarantees for a 36 percent ownership of Citi. Geithner, thinking about how that leverage might be used, said, "Maybe we suggest a few new directors and let them decide."

Dugan said he was concerned about how "the markets will react" to pushing out Pandit.

Geithner, who'd chatted with Pandit a few hours before, added that maybe they could suggest that Vikram hire some more commercial bankers underneath him.

Forty-five minutes pa.s.sed-that was all for today. The men hung up. Bair, after she heard the clicks, wondered, as usual, what more she might have said.

Geithner sat at his desk and signed forms allowing for various foreign acquisitions or investment in the United States, a system started thirty years before to review such activities through the lens of national security.

A call had been scheduled at 4:05. His secretary asked if he was ready; it's Vikram Pandit. Geithner told Pandit, as he did most days, where things stood.

The next day, Bair got a call from Summers's office. She was surprised. She didn't talk to Summers all that often. Today, though, Summers was gracious and eager, particularly interested in discussing the basics of how a prepackaged bankruptcy might work on a bank like Citi. Bair ran through it.

Summers was circ.u.mspect. He didn't tell Bair that he and Romer were now, for the most part, in Bair's camp and that they'd be in a "showdown" Sunday with Geithner about the future of big banks like Citi. He asked how deep the hole was-the hole that some funds, from somewhere, would have to fill if the bank were shut down and reopened. She said it was about $600 billion tops, and explained the "intrinsic value" calculus.

He said Treasury seemed to think it was higher, more like $800 billion-a number so big it made bankruptcy more difficult. That higher number, of course, made Citi too big to fail.

Without context, though, Bair couldn't really discuss how these cost estimates could shape options and policy. She just considered it an informational call and told Summers to call anytime.

On March 11, the AIG mess finally caught up with Tim Geithner-and it was ugly.

The insurance giant had been given another $30 billion just several days before, on March 2, to sh.o.r.e up its operations, bringing the government's total contribution to the firm up to a stunning $170 billion. That same day, AIG declared a fourth-quarter 2008 loss of $62 billion, easily the largest loss in U.S. corporate history.

Geithner knew that bad was about to get worse. Soon the firm would be paying out those secret bonuses. Treasury had managed to avoid an incident in early February, by quietly reshaping the bill by Chris Dodd, but now they were coming to an actual payday.

It would be a $165 million bonus dispersal, mostly to AIG's top bra.s.s. Out of a total $450 million in bonuses, $55 million of which had already been paid, $230 million had yet to be paid out to AIG employees in 2009.

But it wasn't just the bonuses. Tucked in the disclosures AIG was due to make was a story that would carry another set of explosive numbers. The firm had used its bailout money to pay not just bonuses, but also much larger obligations on its credit default swaps to Goldman and other banks. Congressional committees, enlivened by Obama's strong words of censure, were closing in on this point. On March 5-while Geithner was trying to handle Citigroup's woes, and Obama was running his Health Care Summit-the Federal Reserve's Donald Kohn, Bernanke's number two, had testified before Dodd's committee to the effect that he didn't want to release which counterparties were being paid what with the AIG money because it would undermine "confidence" in the markets.

The Fed in fact was pushing Geithner to keep his proposed stress tests confidential, so no one, except Treasury, would know how various banks had rated. But the dome of silence, central to the confidence game constructed between New York and Washington, was cracking. Bloomberg News reporters were hot on the trail of the AIG bonuses. Reporters from the Washington Post were not far behind.

On Wednesday, May 11, Tim Geithner called Ed Liddy, whom the government, in consultation with Goldman Sachs, had placed atop AIG the previous fall. Liddy, a former Goldman executive, told Geithner that there was nothing to be done. The bonuses had to be paid.

"We have to do something, Ed," Geithner said.

"They're contracts, Mr. Secretary," Liddy responded. "You can't violate a contract."

Liddy said he would write a letter expressing why AIG needed to pay the bonuses, no matter how distasteful this seemed.

Geithner hung up the phone.

He knew he'd be drawn deeply into all this. He was the New York Fed chairman on watch when they had approved the AIG bonuses, the counterparty payments-all of it. The question of why he had let it get to this point, nearly six months after the arrangement was struck, and why he hadn't alerted the president, still hung in the air.

His schedule was also a problem. The stress-test proposals were proceeding apace. His whole team was working on them, led by Lee Sachs. But Geithner had to leave town. He had meetings over the weekend in Suss.e.x, England, with finance ministers from the G20, twenty of the world's strongest economic powers, in preparation of the full meeting with Obama and the others, scheduled for April 2.

No one was more delighted about this than Larry Summers. He booked time with the president on Friday. His tough-love proposals were taking shape. Now he could sell them to the president while Geithner was far away, across the ocean.

On Sat.u.r.day, Romer's team from the Council of Economic Advisers and Summers's team from NEC met in the latter's office. News of the AIG bonuses was now all over the papers, and it was Armageddon. The outrage at paying out what looked like a king's ransom to executives of the companies whose "irresponsible" and "shameful" behavior, in Obama's parlance, had wrecked the economy-and who had been saved only by taxpayer money-bled in every direction. Citigroup had gotten $50 billion in federal funds; Bank of America, $45 billion; JPMorgan and Wells Fargo, $25 billion each; and there were plenty more.

If anything, the news storm drove the Summers-Romer team even harder. This was precisely the problem, they said, when government forgot that its role was not to support failing businesses or use government funds to create private profits. It was time for a clean break. They spent the afternoon working through their proposal for government's intervention in the financial system.

In this area, Romer had particular strengths. Like Bernanke, she was an expert on the Depression, especially on the ways the Roosevelt administration had restructured the American financial system. The restructuring had yielded a kind of defining clarity to the managing of money and risk across four decades. But now, after thirty subsequent years of drift without this clarity, Obama had a chance to be Roosevelt. With Romer and Summers working in concert, matching her expertise with his rhetorical gifts, this might be the moment.

The tension at this point was, at any rate, acute. Geithner felt that what Summers and Romer were doing was nothing short of reckless and that they were leading Obama down a path to disaster.

Team Summers-Romer dialed up Team Geithner on an overseas call.

Geithner and his deputies were on a plane back from Suss.e.x. The call started cordially, but descended quickly into angry exchanges.

As one of Geithner's deputies told Romer, "Mommy and Daddy are fighting-can't someone help us."

Apparently not.

On Sunday morning, March 15, Alan Krueger ducked his head into Geithner's office.

"Tennis today? It's nice out."

Geithner was a good player-a great athlete in fact, with a New York magazine article describing him as a " 'dauntingly fit' stud on the tennis court."

While Obama favored golf, for the small group that managed U.S. economic policy, it was all about tennis. They'd all played high school tennis, some junior tournaments. Summers was a strong player. Gene Sperling, from Treasury, had played at the University of Michigan. The four often played doubles. Krueger was the best, still able to take games off of top college players.

"I can't tell you how much I'd love to play tennis, Alan," Geithner said wearily. "I'd give anything."

Geithner had a much more important match to prepare for, against Summers, for high stakes. That afternoon Geithner's team gathered to finalize their battle plan. His proposed stress tests had evolved nicely. If the president could see them as forceful action to repair the banking system, and not just one more delay in dealing with this th.o.r.n.i.e.s.t of issues, the stress tests would become the undisputed government policy.

But it would be a matter of both offense and defense, defend and attack. In his mind, Summers was just marshaling the arguments of amateurs, pundits, and politicians to cozy up to the president. He had no idea what the takedown of a bank looked like, and Bernanke, who did, was with Geithner. They, after all, would have to play an instrumental role in either the stress tests or any wider intervention.

Yet Geithner had other things to worry about. He'd been mentioned in every story on the AIG bonuses. He would have to face the Watergate question: What did he know and when did he know it? Geithner, himself, had become toxic.

And it was none other than Larry Summers who was on several of the Sunday morning talk shows essentially criticizing what Geithner had done, or not done, on the controversial bonuses.

"There are a lot of terrible things that have happened in the last eighteen months, but what's happened at AIG is the most outrageous," Summers said on ABC's This Week. He reiterated a similar position that same day on another show, CBS's Face the Nation, saying that the administration's priority was safeguarding the American taxpayer. "No one cares about the shareholders of AIG. No one feels the slightest obligation to people who led us into these difficulties."

Some felt that those "people" included Tim Geithner.

By late afternoon, the Roosevelt Room was already crowding up. People had come early to get a good seat. Everyone was there: Summers, Romer, Emanuel, Biden, Geithner, Axelrod, Jarrett, and the political team, folks from Treasury, teams from the CEA and the NEC, a.s.sistant to the president Phil Schiliro and his legislative affairs staff.