Stress Test

  So this seemed like a perfectly appropriate time for the central bank to do something. Ben was not a creature of Wall Street; he had literally grown up on Main Street in a small South Carolina town. When I had hosted dinners for him at the New York Fed, he had mingled awkwardly with Wall Street’s tycoons. They worried he was too academic, too short of experience in markets. But while Ben was the understated opposite of a classic financial CEO, he was not afraid to act to protect the financial system. He understood the danger that trauma in the markets could cause for ordinary Americans, that, as he put it, “the grass gets trampled when elephants fall.”

  I shared his fear, if not his quiet way of expressing it. A full-blown financial crisis would threaten jobs, homes, retirement savings, car loans, student loans, small business loans, and international trade, not just exorbitant Wall Street bonuses. Perhaps I had a bias toward action, but the Mr. Fix It criticisms felt like excuses for inaction. In the emerging-market crises, I had attended countless international meetings where similar arguments about moral hazard, inflation risk, and central bank credibility had been invoked to justify delay. Well, delay could be risky, too. Mervyn King, governor of the Bank of England, had criticized the ECB and the Fed for overreacting after we pumped liquidity into the markets in early August, warning that we were creating dangerous moral hazard. He was less critical after September 14, when he had to provide similar assistance to Northern Rock, a mortgage lender that had become the target of England’s first bank run in nearly 150 years.

  On September 18, we cut the federal funds rate target by a half point, and the markets rallied. Financial tensions in Europe and the United States eased a bit. We cut another quarter point in October, and the markets continued to show signs of calm. I was still worried we faced a long war, so I assigned my staff to investigate new ways to inject liquidity in times of stress, including potential aid to nonbanks. I also tried to bring more force to our broader efforts to enhance the resilience of the financial system. We convened an international “Senior Supervisors Group,” bringing together U.S. and foreign regulators of the largest commercial banks as well as investment banks to encourage more prudence in risk management. Using the model of our horizontal risk reviews, we wanted to push major institutions to improve their stress testing and prepare for a severe crisis.

  By October 17, things seemed calm enough that I referred to the crisis in the past tense in private remarks at a seminar with international financial officials in Washington. “In important respects, the system worked,” I said. “The capital cushions at the largest banks proved strong enough to withstand the shock.”

  I did point out that the events of August had exposed significant fragility in the system. Market assumptions about limitless liquidity had been proven wrong. Leverage had piled up outside the traditional banking sector. “The very substantial improvements in risk management since the last crisis did not capture some risks that inevitably look more obvious in retrospect than they did left of the boom,” I said.

  Left of the boom was a phrase I had picked up from analysts studying improvised explosive devices at the RAND Corporation think tank, where I served on the advisory board. The phrase referred to the time before an IED explodes. In October 2007, I thought we had survived a major financial explosion.

  In fact, we were still left of the boom.

  A WEEK later, the investment bank Merrill Lynch announced $7.9 billion in mortgage-related losses, the largest write-down in Wall Street history. That was almost twice as large a write-down as Merrill had predicted three weeks earlier, leaving the impression that losses were exploding and more unpleasant surprises lay ahead. Merrill CEO Stan O’Neal was forced out, although he did receive a $161.5 million severance to ease the blow.

  The bulk of Merrill’s losses came from “collateralized debt obligations,” piles of mortgage-backed securities where the income streams had been sliced up and repackaged into smaller streams known as “tranches.” Merrill was a leading manufacturer of CDOs, and it had made billions selling them to investors around the world. But the investors, reaching for yield, had shown little interest in the safest tranches, the “super-senior” CDOs that would pay out in full unless mortgage losses were so severe that investors in every tranche below them were wiped out. That seemed highly unlikely, so Merrill usually kept the super-seniors on its balance sheet. Their modest returns were still more than the cost of financing them, and they seemed almost bulletproof. Standard & Poor’s estimated a mere 0.12 percent chance that one of its AAA-rated CDOs would fail to pay out over five years—and super-seniors were considered safer than typical AAAs.

  But as Nate Silver noted in The Signal and the Noise, his excellent book about why many predictions fail, the actual default rate for AAA-rated tranches of CDOs would be 28 percent, more than two hundred times higher than S&P had predicted. Their perceived safety rested on all kinds of flawed assumptions, starting with the notion that housing prices would never fall simultaneously across the country. CDOs were often spliced together from geographically diverse piles of subprime mortgages, which was supposed to mitigate the effects of a housing slump in any one region. But geographic diversity didn’t help much when borrowers started defaulting nationwide.

  The forecasts also failed to anticipate the financial-death-spiral effects of the panic, the danger that market psychology could amplify problems in the real estate market. Regardless of the actual safety of the super-seniors, once it became clear that the AAA-rated pieces of CDOs were less safe than previously believed, investors who had bought them (or accepted them as collateral) without examining what was in them began to sell them (or reject them as collateral) with just as little thought. Panic is the flip side of overconfidence; once markets start racing to safety, investors rarely bother to try to distinguish truly overvalued assets from assets that seem similar but might be getting a bad rap. Once a stampede begins, everyone wants to get out the door, and hesitation can be deadly.

  The mortgage contagion soon infected one of the world’s biggest banks, one of the New York Fed’s banks. On November 4, Citigroup broke Merrill’s new record, warning that it might take as much as $11 billion in write-downs, seven times what it had projected on an earnings call three weeks earlier. Citi also revealed it had $55 billion worth of subprime exposure, four times what it had said on that mid-October call. Citi’s last piece of news was that Chuck Prince, the CEO who had said banks needed to keep dancing while the music kept playing, was out of a job.

  Clearly, the music had stopped.

  WE NEVER thought of Citigroup as a model of caution. It had been at the center of the Latin American debt crisis of the 1980s. The New York Fed had cracked down on it for shenanigans related to the Enron scandal shortly before I arrived, and we hit it with the subprime lending fine that pleased Paul Volcker shortly after I arrived. In 2005, after Citi was forced to shut down its private banking operations in Japan because of illegal activity, we banned the company from major acquisitions until it fixed its internal controls and other overseas governance issues. The British banker Deryck Maughan, whom I knew from his days running Salomon Brothers in Japan, came to see me after he was forced out of his job as chairman of Citigroup’s international operations. His message was that Citi was out of control.

  Citi had more than three hundred thousand employees in more than one hundred countries, and frequent drama in its executive suites. Our supervisors always considered it a laggard in risk management, unwilling to imagine ugly states of the world, unsure how to evaluate exposures across its far-flung businesses. These weaknesses were troubling, but the firm did not appear nearly as vulnerable as many other institutions, certainly not compared to the investment banks, the GSEs, or other nonbanks that didn’t have insured deposits and didn’t have to hold as much capital. Bob Rubin’s presence at Citi surely tempered my skepticism as well. Even though he had no management responsibility or authority, he sat on the board, and he probably gave Citi an undeserved aura of competence in my mind.<
br />
  Citi wasn’t as well capitalized as we thought—partly because a very small share of its capital was common equity, the strongest bulwark against future losses, partly because we vastly underestimated the riskiness of its AAA-rated mortgage assets. Like Merrill, Citi was a leading manufacturer of CDOs, and like Merrill, it kept many of its super-senior tranches. The AAA label ended up being very misleading. The rating agencies were not exceedingly competent. Their ratings typically lagged cycles in finance, staying too optimistic too long. Since the issuers rather than the purchasers of securities paid them, they had some incentive to give generous ratings that kept issuers happy. Moody’s revenue from rating structured products such as CDOs had risen 800 percent in a decade.

  By 2007, Citi’s $2 trillion balance sheet was much riskier than it looked. And Citi had stashed another $1.2 trillion in assets off its balance sheet in ways that allowed it to hold virtually no capital against losses in those assets. Citi didn’t expect funding for these off-balancesheet vehicles to dry up, because of the same delusions that made risky securities seem safe to investors and rating agencies. They didn’t understand how quickly losses could boomerang back onto its balance sheet, and neither did we.

  For example, Citi financed some of its off-balance-sheet securities by issuing asset-backed commercial paper, while providing assurances that it would buy the paper if no one else wanted it. When markets ran from asset-backed commercial paper—the market shrank 30 percent in the second half of 2007—Citi had to shell out $25 billion to make good on its assurances. Citi also had exposure to securities in off-balance-sheet “structured investment vehicles,” a popular financing mechanism that swiftly became unpopular after a few SIVs stocked with subprime mortgages failed. Markets began to flee other SIVs, even those with less subprime mortgage exposure, forcing them to liquidate assets in a hostile market. Some SIVs were wiped out, while Citi and other sponsors had to bring assets from SIVs back onto their balance sheets. That’s how the company’s perceived exposure to subprime quadrupled in four weeks.

  The international Senior Supervisors Group concluded in November that Citi’s managers had no idea how close they were flying to the sun, that the company “did not have an adequate, firm-wide consolidated understanding of its risk factor sensitivities,” that its stress tests “were not designed for this type of extreme market event.” Citi wasn’t as weak as many other banks—not to mention the thrifts, Fannie and Freddie, AIG, and several investment banks—but its unexpected losses were very damaging to overall confidence. As Citi’s supervisors and regulators, the New York Fed, the OCC, and the SEC had missed what Citi’s leadership, creditors, and shareholders also missed: its dramatic exposure to an increase in mortgage defaults and even more dramatic declines in the price of mortgage securities. We weren’t expecting default levels high enough to destabilize the entire financial system. We didn’t realize how panic-induced fire sales and radically diminished expectations could cause the kind of losses we thought would happen only in a full-blown economic depression.

  A few days after Prince stepped down, Sandy Weill, the former CEO who had built Citi into a mega-firm in the 1990s, told me he was going to propose to its board that I should succeed Prince as CEO. I thought this was a crazy idea, for Citi and for me. Carole was appalled by the thought. She thought it sounded unseemly, and would put me in a hornet’s nest. I didn’t yearn for an eight-figure income. And I couldn’t imagine abandoning a job I loved, particularly in the middle of a crisis. I asked Weill not to put my name forward.

  “I’m not the right choice,” I told him.

  A week later, after we met for lunch, Rubin made it clear he didn’t think Weill’s idea made sense, either. He asked me: Would you really want to spend your days arbitrating fights over compensation? He thought my challenges at the Fed were more interesting and meaningful, and so did I. Citi’s board gave the CEO job to Vikram Pandit, the former Morgan Stanley executive who had once warned me that Wall Street was broken.

  WE HADN’T pushed our banks to raise more capital in good times, because they were comfortably above the required regulatory ratios. As their losses mounted, we did push the weaker banks to raise more capital, but by now it was harder for them to raise money on a major scale. We considered forcing banks as a group to stop paying dividends in order to conserve capital, but we were concerned, perhaps mistakenly, that doing so might do more harm than good. It would be unfair to the relatively strong, and it would make investments in financial firms less attractive, which could make it harder for the entire system to raise more capital if conditions deteriorated.

  However, we did force Citi to reduce its dividend, which it had pledged not to do, and we told it to raise new capital. The bank managed to raise $20 billion over the next few months, mostly from sovereign wealth funds in the Middle East and Asia. Those funds, the last big sources of liquidity left in the markets, also injected capital into other struggling firms, including Merrill and Morgan Stanley.

  I believed then and still believe now that forcing banks to hold enough capital and liquidity to absorb significant losses is the best defense against future crises, the ultimate shock absorber. My financial reform mantra after the crisis would be “capital, capital, capital.” The question, of course, is how much is enough. At the Fed, we hadn’t required Citi to hold enough capital because we hadn’t fully understood the extent of the risks it was taking. We also let banks, including Citi, count lower-quality capital, such as preferred stock and certain forms of subordinated debt, that technically complied with the rules but didn’t absorb losses as well as common equity. And I hadn’t pushed the Federal Reserve in Washington to change the rules to raise regulatory capital requirements, because those rules were then mired in protracted international negotiations. The United States could have toughened the existing rules on our banks unilaterally; I didn’t think there was much hope of swift and meaningful reform, but knowing what we know today, we should have tried to do more. Even though banks were better capitalized than much of the rest of the financial system, they were not strong enough to absorb the damage that was still to come.

  As the crisis escalated, markets continued to run from mortgage assets that looked toxic, and as investors shunned them, they became toxic, adding to concerns that the financial system had inadequate capital to absorb losses. Firms with too many toxic assets began to lose their ability to borrow to finance the assets, leading to fire sales that further depressed the prices of the assets and forced the firms to take additional losses, intensifying the vicious cycle and exacerbating the concerns about inadequate capital. In that August videoconference, one Fed governor described the toxic assets as “dreck.” Fisher called them “flotsam and jetsam.” By any name, they were poisoning the system.

  One analogy for the flight from toxic assets, from the economist Gary Gorton, would be the kind of understandable hysteria that breaks out after E. coli turns up in someone’s hamburger. The problem might be poorly inspected ground beef from a single factory, but individual consumers have no way of knowing if the burger or even the steak they want is bacteria free. They just know that meat advertised to be safe turned out not to be, so they respond by distrusting advertisements trumpeting meat’s safety and avoiding meat altogether. But the financial version of an E. coli hysteria was even worse, because once consumers started to avoid certain financial products, they dropped in value regardless of their level of contamination, which made them even more dangerous to consumers and ramped up the hysteria. It was as if avoiding meat actually caused E. coli to spread.

  In a climate this tense, uncertainty about the quality of mortgage-related assets mattered more than the actual quality of the assets. Rationally, not every mortgage was going to end up in default, and even mortgages that did go into default would ultimately recover some value as the foreclosed homes were sold. But uncertainty overwhelmed that kind of analysis, so irrational valuations became rational for investors. Nobody knew when more bad earnings news could cause anoth
er paroxysm of asset selling, which could cause more margin calls and collateral demands, followed by more write-downs and capital shortfalls, followed by yet another paroxysm of asset selling. In theory, the mortgage securities would all be worth something someday, but in reality, no one wanted to buy them because everyone was trying to get rid of them, which meant their real-time “distressed” price was next to nothing. As Keynes supposedly said, the market can stay irrational longer than you can stay solvent.

  This dynamic would recur throughout the crisis, raising tough questions about what assets were really worth. If a security sold at its $100 “par” value a month ago, and you couldn’t sell it for $30 today, but it might be worth $80 in five years, what was its true value? And what kind of write-down should you take? Those kinds of questions would help determine whether banks were insolvent or merely illiquid, and how to remove toxic assets from their balance sheets.

  My colleagues and I talked a lot about whether and how government could help provide a backstop for those assets to avoid fire sales. At that point, Treasury had no such authority, and we didn’t see a strong case for a Fed role. Treasury Secretary Hank Paulson wanted to support a private-sector solution, a massive “Super SIV” financed by big banks that would issue commercial paper to buy toxic assets from existing SIVs. The banks initially pledged to participate, but the effort quickly fizzled, because they needed their cash to absorb losses from their own toxic assets. Without government backing, the case for the Super SIV relied on what felt like circular reasoning; it required the private sector to do what the private sector was already refusing to do.