Stress Test

  The eventual losses in subprime went higher than 14 percent, especially in structured products that concentrated the riskiest loans. And mortgage losses spread further than subprime. The prevailing assumption that housing prices would not slump nationwide, though widely shared and backed by seven decades of history, also turned out to be wrong. This is often seen as our worst misjudgment, since so many subsequent problems in the financial system would involve mortgages and mortgage-backed securities. But it was not our most important miss.

  Even though we did not expect a nationwide real estate bust, we did analyze scenarios where house prices fell sharply across the country. Unfortunately, we concluded that the impact on the financial system and the broader economy would be relatively modest. We didn’t foresee that falling housing prices alone would trigger widespread mortgage defaults that could cause significant problems in the banking system, because we didn’t examine the possibility that the initial fears associated with subprime mortgages and the general fall in house prices could trigger a classic financial panic, followed by a crash in household wealth and a collapse of the broader economy. This was the arc of crisis described in Kindleberger’s book: manias, followed by panics, ending in crashes. And this was the death spiral I had seen so often at Treasury during the emerging-market crises of the 1990s.

  We knew the long boom in credit in general and mortgage credit in particular exhibited all the classic signs of mania, including the widespread belief that housing prices would never fall to earth. And we were keenly aware that dangerous levels of leverage were building up throughout the financial system, especially in less closely supervised corners of the system. But we didn’t appreciate the extent to which nonbanks were funding themselves in runnable short-term ways, or how vulnerable the banking system would be to distress in the nonbanks. We knew we were vulnerable to a panic, and I spent years concerned about our vulnerabilities, but ultimately we were even more vulnerable than we realized.

  We also failed to anticipate the savage depth of the Great Recession, or the debilitating feedback loop between problems in the financial system and problems in the broader economy. A mere drop in housing prices could not have triggered mass mortgage defaults or depression-level losses in the banking system, but as unemployment increased and jobless homeowners missed payments, actual and expected mortgage losses increased as well, triggering margin calls and selloffs of mortgage-related assets, making the economy, unemployment, and the housing market worse. We weren’t prepared for that kind of doom loop.

  You could say our failures of foresight were primarily failures of imagination, like the failure to foresee terrorists flying planes into buildings before September 11. But severe financial crises had happened for centuries in multiple countries, in many shapes and forms, always with pretty bad outcomes. For all my talk about tail risk, negative extremes, and stress scenarios, our visions of darkness still weren’t dark enough. When models told us that only a depression-level shock could cause severe distress to banks, we should have done more to consider what would happen in the event of a depression, no matter how unlikely it seemed. And when our analyses told us the financial system was well positioned to shrug off housing losses, we should have thought harder about the possibility that they could trigger panic.

  It’s not as if we were twiddling our thumbs during the gathering storm. The New York Fed was the leading source of initiative in the regulatory community in trying to reduce the financial system’s vulnerability to a crisis. Even at the height of the boom in credit and leverage in mortgage-related markets, we tried to lean against the winds of confidence, to encourage major banks and nonbanks to prepare for the worst. We recognized there was a lot of dry tinder in the system, even if we were late to see how subprime losses could be the spark that ignited it.

  None of this was enough. In the end, we still ended up with a raging inferno. We certainly could have been more prescient, more forceful, more imaginative. But we were human. And our limited authority left us with limited options. For decades, the U.S. government had allowed the financial system to outgrow and outflank its regulatory system. Financial stability cannot depend on omniscient supervisors identifying and preemptively defusing any potential source of crisis; it requires safeguards that can help the system withstand the force of a severe storm, and tools the government can use to limit the damage.

  Even while our storm was gathering, we could see that the safeguards in our financial system were terribly inadequate and needed fundamental reform.

  BY THE summer of 2007, the subprime market was imploding.

  Borrowers with “no-doc loans” (without proof of income) and “liar loans” (with inflated claims about income) and “NINJA loans” (No Income, No Job, No Assets) were defaulting in droves. The second largest subprime lender, New Century Financial, went bankrupt, while the largest, Countrywide Financial, revealed that it was running short on cash. Meanwhile, the price of insuring bonds backed by subprime mortgages against default soared, prompting the rating agency Standard & Poor’s to warn that hundreds of those bonds could be downgraded. And two big hedge funds financed by Bear Stearns—including one called, incredibly, the Enhanced Leverage Fund—collapsed after their subprime investments tanked.

  I didn’t find any of this shocking, or even necessarily unfortunate. The overheated subprime market was overdue for a cooling, and the investors who had bought securities expecting the boom to continue were generally consenting adults. The risk premiums that were now rising across the system had been too low for too long, encouraging too much reaching for yield. It was an absurd sign of the times that “enhanced leverage” had become a selling point for an investment vehicle, instead of a warning; it was like naming a new car model after its faulty brakes. At the Fed’s rate-setting meeting on August 7, 2007, I described the turmoil in subprime as “a necessary adjustment, a generally healthy development.”

  But I also warned the committee that the turmoil had “the potential to cause substantial damage”—through a contraction in credit, through a liquidity crunch, “through the potential failure of more consequential financial institutions; and through a general erosion of confidence among businesses and households.” I noted that investors were losing confidence in their ability to assess risk in mortgage markets, in the market’s ability to value complex mortgage securities, and in the rating agencies that had assured them for years that those securities were safe. I also said we were entering a new realm of uncertainty and risk, and “a lot of that risk has gone to leveraged funds that have much less capacity to absorb this kind of shock.”

  I was more worried than many of my colleagues, but I was not nearly worried enough. Our debate that day was not about whether to cut rates. Unemployment was only 4.7 percent, so the economy didn’t seem to need immediate help. Our debate was about whether we should hint that we might cut rates in the future, and most of the committee preferred to wait for more evidence that market disruptions were damaging the economy first. “If we took that approach,” I warned, “we’d inevitably be too late.” But the group wasn’t ready to signal that a future rate cut was likely. In our official statement, the committee noted that “downside risks to growth have increased somewhat,” but maintained that increased inflation was still a larger concern than an economic downturn.

  I was uneasy. I urged caution. But I didn’t clamor for action. I even conceded it would be tricky to temper our public optimism without “feeding the concern … about underlying strength in the fundamentals of the economy as a whole or in the financial system.” I didn’t want to sound too dark too soon.

  I didn’t have to wait long.

  On August 8, I left to see my family on Cape Cod, the same place I had been a decade earlier when the Thai crisis hit. The next day, my staff called to tell me that French bank BNP Paribas, one of the world’s largest financial institutions, was suspending withdrawals from three of its investment funds with heavy exposure to the U.S. subprime meltdown. BNP had announced that it c
ouldn’t even put a fair price on its subprime-backed securities, citing “the complete evaporation of liquidity” in those markets. Saying the bank had no idea what its subprime assets were worth was much worse than saying the values had declined by 20 or 30 percent. It intensified uncertainty about subprime, as well as other U.S. mortgage-backed securities, and helped to create the larger liquidity crunch I had warned my Fed colleagues about much too gently two days earlier.

  But if the BNP Paribas announcement troubled me, the force of the European Central Bank response really got my attention. Under the leadership of Jean-Claude Trichet, a former French finance official I had known during my time at Treasury, the ECB was considered a solid and conservative central bank, generally reluctant to intervene to support the economy or the financial system. It wasn’t being conservative now. First it announced that it would lend unlimited amounts at a discounted rate to any bank that requested it, an unprecedented effort to pump liquidity into the system. Then it announced that it was fulfilling 95 billion euros in requests for cash, about $130 billion.

  Wow. My first response to the news was a string of expletives. I knew Trichet wouldn’t put up a wall of money like that on a whim. This wasn’t just a little problem on the fringes of the U.S. mortgage market.

  I had a sick feeling in my stomach. I knew what financial crises felt like, and they felt like this.

  FOUR

  Letting It Burn

  Central bankers are supposed to be the brakes on the boom, taking away the punch bowl just as the party gets going. But when there’s panic in the air, and liquidity starts to evaporate, central bankers are supposed to be the accelerator.

  In his 1873 book Lombard Street, the bible of central banking, Walter Bagehot explained that the way to stop a run is to show the world there’s no need to run, to put money in your window, to “lend freely, boldly, and so that the public may feel you mean to go on lending.” The loans should be expensive—Bagehot recommended charging “a penalty rate”—so it would be economically unattractive to borrow from the central bank once the panic subsided. And like any central bank loan, they should be secured by solid collateral. But the goal is to keep the financial system functioning by providing liquidity to solvent institutions when private markets freeze. That’s what the ECB did on August 9, 2007. Creditors were pulling back from European banks, frightened by their heavy exposures to U.S. mortgages, and many of those banks were bleeding liquidity. The ECB told them: We’ve got cash. If you need it, come and get it.

  Chairman Bernanke believed in Bagehot, and he believed that central bankers who fail to act in a crisis could make it exponentially worse. In his academic life, Bernanke had been a leading scholar of the Great Depression. His research at Princeton had demonstrated how the central bankers of the 1930s were too timid, too reluctant to provide liquidity, too faithful to the tight-money orthodoxies of their day, allowing widespread bank failures and clogged credit channels to turn a financial crisis into an economic disaster. Ben is a low-key guy, not prone to hysteria or exaggeration—I called him the Buddha of central banking—but he was determined not to repeat their mistakes. While our personalities and our backgrounds are quite different, we complemented each other. His strengths were my weaknesses, with his patience balancing my impatience. We would work in almost perfect harmony throughout the crisis. We trusted each other.

  On a call that morning, Ben, Don Kohn, and I agreed we needed to make it clear the Fed was willing to inject liquidity into our markets, too. Fear was building across the financial system, especially in the corners that touched housing. Securities backed by “nonconforming” mortgages—the ones that weren’t guaranteed by Fannie or Freddie—were becoming pariah paper; no one knew how much they were worth anymore, so almost everyone was running away from them. That had dangerous implications for financial institutions that were using them as collateral for short-term loans or had calculated how much capital to hold based on optimistic assumptions about their value. Countrywide Financial cited “unprecedented disruptions” to mortgage finance in an SEC filing that day. The bursting of the housing bubble had felt like the end of a mania, and the backlash against housing-related investments now felt like the start of a panic.

  At the start of any crisis, there’s an inevitable fog of diagnosis. You can recognize the kind of vulnerabilities that tend to precede severe crises, such as substantial increases in asset prices and dramatic expansions of leverage, but you can’t be sure whether the initial market turmoil is a healthy adjustment or the start of a systemic meltdown, a precursor to a modest economic slowdown or something much worse. An orderly deleveraging of an overheated market can be a good thing, even if some big companies fail. That’s how capitalism is supposed to work: weak, bloated, and mismanaged firms make way for more dynamic competitors. Creative destruction instills discipline in the survivors. But a healthy correction can spiral out of control if fear and uncertainty gain too much momentum. You don’t want a central bank to underreact, allowing runs by depositors and creditors to unleash a cascade of fire sales, where firms desperate for cash have to sell assets into a depressed market, further depressing asset prices, inducing more cash shortages and more desperate selling. Once this dynamic of runs and fire sales gets started, it can be incredibly hard to contain, with enormous costs to the economy.

  That said, if a central bank treats every disturbance as a potential catastrophe, riding to the rescue at the first sign of trouble, it can create real moral hazard—encouraging reckless risk taking, propping up nonviable “zombie” banks, re-inflating bubbles, and setting up the economy for a fall from an even higher cliff, ultimately magnifying the size of the next crisis. A central bank overreaction can even backfire right away, feeding instead of easing concern. Markets sometimes conclude a situation must be dire if policymakers act like it’s dire. Often it’s better to stand back, as we had done in September 2006, when bad bets on natural gas futures sank a giant hedge fund called Amaranth Advisors. That looked like a one-off, an idiosyncratic casualty of bad risk management, unlikely to light the rest of the financial system on fire. And it was.

  The situation in August 2007 seemed more dangerous and more systemic. We quickly pumped $62 billion into the U.S. financial system—not as much as the ECB poured into Europe, but enough to send a message that we were on the case. It was a fairly modest and completely conventional early-stage escalation by a central bank, injecting some liquidity into markets. Yet some of our colleagues on the FOMC already thought we were being too aggressive, prompting a debate reminiscent of the discussions we used to have about moral hazard during the Asian crises. On a conference call the morning of the 10th, Dallas Fed President Richard Fisher suggested we were helping banks without getting anything in return, “putting our finger in the dike” without extracting promises of more lending now or more responsible behavior in the future. He said he wanted to make sure “we don’t send any signal that we’re just going to be … indiscriminate.”

  “I don’t think that’s the way to think about it,” I said. We needed to signal that cash was available, to help thaw what Ben called “a general freeze-up” in mortgage markets, so that liquidity shortages didn’t create a vicious spiral of forced asset sales and falling asset prices. That meant committing to provide the liquidity that markets needed to function, without conditions, even if it did make us look indiscriminate. “You can’t condition that statement without undermining its basic power,” I said.

  Fisher’s discomfort foreshadowed what we were up against. Even inside the Fed, we were susceptible to Old Testament and moral hazard critiques. One day into the crisis, some already thought we were coddling the perpetrators. That’s the way it always is in a crisis. And I felt that familiar nausea of foreboding.

  I headed back to New York to a familiar blizzard of strategy meetings. I remember Carole calling in the middle of one tense conference call to tell me that bats were flying around our cottage on the Cape, scaring the kids. She had a debilitating case of ca
lcific tendinitis, which made her hand feel like a truck had run over it, an ailment that was not conducive to bat removal or single parenting. I was absent again, just as I had been so often when our kids were babies. I would rarely be fully present over the next several years. I would be the guy coming home late and walking through the door holding cell phones on both ears. My brother Jonathan once said he could always tell from the moment he saw me at family gatherings whether I was really there or not; after August 2007, I usually was not.

  I felt overwhelmed by dread about what lay ahead, worse than I had ever felt while fighting crises at Treasury. I knew our financial system was heavily leveraged and vulnerable to a sudden shift in psychology. I wasn’t confident we had the tools to deal with a panic. And I knew there was a lot we didn’t yet know.

  FOR EXAMPLE, the California lender Countrywide wasn’t on our radar screen until shortly before the night of August 15, when it nearly paralyzed the entire system.

  Countrywide was not only the largest subprime mortgage lender, it was the largest mortgage lender, period, originating nearly half a trillion dollars in loans in 2006. Fortune had dubbed it “the 23,000% Stock,” a tribute to two decades of spectacular growth. It owned a midsize bank with insured deposits, which it had reorganized as a thrift in order to get the lenient OTS as its regulator, but most of its action was in largely unsupervised nonbank affiliates on the wild frontier of the mortgage markets. It was a case study in regulatory balkanization: big enough to matter, but without a regulator responsible for overseeing the institution as a whole, or its potential risks to the system. The New York Fed did not supervise any part of Countrywide, but it was about to become our problem.