Stress Test

  At a time when mortgages were becoming radioactive, Countrywide’s business model was all about mortgages: originating them, servicing them, selling them, packaging them into securities, trading those securities, and using those securities as collateral for borrowing that financed the rest of the operation. And Countrywide had helped lead the erosion in lending standards across the country, selling exotic mortgages to families with dubious credit. In late July, it revealed that its delinquency rates in subprime had doubled in three months; CEO Angelo Mozilo told investors the housing market was the worst since the Great Depression. When investors saw those unexpected losses, Countrywide became an obvious target for a run.

  In my Charlotte speech in March 2007, I had expressed concern about subprime mortgage lenders who sold all their loans and didn’t face the financial consequences of default. Some critics later blamed the crisis on this “originate-to-distribute” model—on lenders with limited incentives to worry about the long-term creditworthiness of borrowers because they had no “skin in the game.” This was a problem, but Countrywide had plenty of skin in the game. It hadn’t distributed enough of the risks it originated. It was like a drug dealer who got high on his own supply. Its main problem was not bad incentives, but bad beliefs—the widespread delusion, profitable for so long, that home prices would defy gravity indefinitely.

  That problem only became an existential crisis because of Countrywide’s reliance on unstable, runnable short-term funding. The firm raised much of its operating cash by issuing “commercial paper,” corporate IOUs that matured in nine months or less. Typically, lenders “rolled over” commercial paper at maturity, renewing the IOUs, but there were no guarantees. Countrywide also financed itself through a complex market known as “tri-party repo,” selling securities while agreeing to repurchase them at a specified time, often as soon as the next day—essentially, borrowing overnight with the securities as collateral. Much of the firm functioned like a thinly capitalized bank without deposit insurance, dependent on the confidence of its funders. They could decide to stop funding at any time.

  In early August, some of Countrywide’s lenders refused to roll over its commercial paper, forcing it to sell assets to pay them back. Those were the “unprecedented disruptions” Countrywide disclosed on the 9th. The next day, I was forwarded a rosy email from Countrywide’s supervisor, OTS Director John Reich, suggesting the firm was not only safe, but poised to benefit from the upheaval in its industry. “The longer term looks positive, as their competition has greatly decreased,” he wrote. Reich added that Washington Mutual, another obviously troubled thrift, was in “very good shape to weather the current conditions.”

  On the 15th, with the price of insuring Countrywide’s debt against default up eightfold in just a month, a previously bullish Merrill Lynch analyst issued a report titled “Liquidity Is the Achilles’ Heel,” warning that cash pressures could force Countrywide into bankruptcy. Mozilo insisted his firm was fine, accusing the analyst of shouting fire in a crowded theater. But when confidence goes, it’s hard to recapture. As Bagehot wrote, “every banker knows that if he has to prove he is worthy of credit, however good may be his arguments, in fact his credit is gone.”

  That evening, I had my second traumatic surprise in a week. My general counsel, Tom Baxter, walked into my office to tell me that Bank of New York Mellon, Countrywide’s tri-party repo clearing bank, was threatening not to “unwind” the firm’s $45 billion repo book the next morning. That had never happened before, and it would have had devastating implications.

  BoNY’s role in tri-party repo was mostly operational—shifting cash and securities back and forth between borrowers and lenders, pricing collateral—but it also provided borrowers with several hours of daytime (or “intraday”) credit while arranging their transactions. Now it was about to tell the world it couldn’t take the risk of Countrywide failing during those hours. Instead of returning the cash that money market funds and other investors had lent to Countrywide overnight, BoNY was threatening to give those lenders the securities that Countrywide had put up as collateral. Countrywide would get tarred with a potentially fatal vote of no confidence. And the money market funds and other investors that had lent the firm short-term cash would get stuck with securities they didn’t want, including some mortgage securities that were already plunging in value. They might start pulling away from other firms that looked similar to Countrywide, triggering a run on the entire $2.3 trillion tri-party-repo market.

  My colleagues and I at the New York Fed made a flurry of calls that night, trying to persuade BoNY to unwind, trying to figure out how markets would react if it didn’t. I participated in thirty-seven calls, the last one at 2:32 a.m., including multiple conversations with Ben, Don, Mozilo, BoNY CEO Gerald Hassell, and Countrywide’s bankers.

  We were all worried about fire-sale dynamics. Money market funds need to remain liquid, and they’re not allowed to hold risky long-term assets. If they got stuck with Countrywide’s securities, they’d have to sell them right away into a falling market, which would exert more downward pressure on prices, produce more problems for firms with mortgage exposures, and exacerbate concerns about the solvency of the financial system. That was scary. But three-fourths of Countrywide’s collateral was in ultra-safe Treasury securities or almost-as-safe Fannie- or Freddie-sponsored “agency” securities with an implicit federal backstop. That made the situation even scarier. One banker observed to me that night that BoNY’s concern over a portfolio dominated by Treasuries was the classic definition of a panic. Nobody questioned the value of Treasuries, but fear was displacing greed, and perception was becoming more important than reality. Institutions didn’t want exposure to Countrywide or even the appearance of exposure to Countrywide, regardless of the quality of its collateral.

  BoNY and Countrywide both urged the Fed to intervene to assume the risk and protect the system, which would become a recurring theme throughout the crisis. BoNY executives said they’d roll over Countrywide’s book if the Fed guaranteed the resulting intraday credit exposure to the firm, indemnifying them against losses if Countrywide failed while they were on the hook. They were basically asking us to stand behind the entire tri-party repo market, because if we backstopped one firm we’d have to backstop them all. At the same time, Countrywide was requesting emergency help from the Fed. Its thrift was eligible for our discount window, but its cash crunch was in its nonbank affiliates. We couldn’t help them unless we allowed them to shift assets out of their troubled nonbanks and into their thrift, or unless we agreed to lend directly to those affiliates, which would have required invoking Section 13(3) of the Federal Reserve Act, the “unusual and exigent circumstances” clause.

  We thought about it, conferred with Washington, and said no. We hadn’t lent to a nonbank since the Depression, and it was too soon for such extraordinary measures. We were still feeling our way, trying to figure out how bad things were. Financial crisis response is, after all, an exercise in triage. The goal is not to save every major firm regardless of its viability; that’s a recipe for the kind of moral hazard that can sustain bloat in the system and increase the risk and intensity of future crises. The goal is to make sure contagion doesn’t get out of control, killing the healthy along with the terminally ill. And the Fed was the lender of last resort; Countrywide wasn’t even at the point where it had exhausted all its other options. It still had access to an $11.5 billion credit line from a consortium of banks. Mozilo was understandably afraid that drawing down the cash would signal weakness, but the point of a credit line is to provide liquidity in a pinch. How could the Fed justify invoking emergency powers to help a nonbank that wouldn’t even help itself?

  I talked to Mozilo for the first time that night, and the call reminded me of my first chat with Thailand’s bewildered finance minister early in the Asian crisis a decade earlier. Mozilo seemed overwhelmed and unclear about what was happening. Like many of the more desperate CEOs I would deal with during the crisis, his ma
in focus was what the government could do to help his firm, and what we could say to get markets to stop fretting about it. He seemed more concerned about the critics pointing out Countrywide’s weaknesses than about those actual weaknesses.

  By midnight, though, we had the outline of a possible solution. BoNY was one of only two clearing banks in tri-party repo—the other was JPMorgan Chase—and we made it clear to BoNY’s executives that refusing to unwind would harm their own interests as well as the system. If money market funds started reevaluating their assumptions about the safety of tri-party repo and pulling out their cash, it would threaten one of BoNY’s core businesses, and could potentially leave BoNY with crippling exposures to other borrowers. Ultimately, BoNY agreed to unwind if Countrywide upgraded its collateral. And Countrywide agreed to draw down its credit line to do so. When the markets opened, Countrywide’s stock price continued to plummet, but a $2 billion investment by Bank of America soon eased fears of a collapse. In January, Bank of America would buy the rest of the firm for another $4 billion.

  The Countrywide episode foreshadowed much of what came later in the crisis. It was a bracing reminder of the limits of our ability to fix problems outside the traditional banking system—problems that could unleash shocks with the power to harm banks as well as nonbanks. It also revealed just how dependent the entire financial system had become on fragile short-term funding arrangements, a central vulnerability in every financial crisis. And it showed that this fragility extended even to the secured funding markets—where lending decisions, in principle at least, were based on the borrower’s collateral rather than just its underlying creditworthiness.

  At the New York Fed, we had spent a lot of time worrying about the risks in tri-party repo, which had expanded tenfold in a decade, but we hadn’t made progress on its worst vulnerabilities. BoNY had operational problems after the September 11 attacks, and we had spent a lot of time trying to make sure the market would be able to function if something physical or financial happened to one of the two clearing banks. But we had never devised a way to make tri-party repo less vulnerable to a run in the event a major borrower lost the confidence of the markets, even though some firms ended up financing more than $400 billion worth of securities at a time.

  We had paid even less attention to the asset-backed commercial paper market, which had nearly doubled during my Fed tenure to $1.2 trillion. In general, we had tended to view secured funding markets as relatively stable, but the collateral that had made them seem stable was now a source of instability. Since many of the assets backing the commercial paper were linked to housing, the entire market was now under pressure. In one week, the spreads between yields on Treasuries and asset-backed commercial paper had soared from 35 to 280 basis points.

  I had also spent very little time familiarizing myself with the risks posed by money market funds, which had accumulated more than $3 trillion in assets, and in many ways functioned like banks. They offered deposit-like accounts to investors, with the ability to withdraw from the accounts at any time at a stable price, and invested that money in securities that weren’t risk-free or perfectly liquid. They did this without the deposit insurance or discount window access enjoyed by banks—and without the supervision and capital requirements that banks must face in return for those benefits. But money market funds were under the regulatory purview of the SEC, and I don’t remember considering them much of a danger to the system before the crisis, even though Paul Volcker and others did. There seemed to be so many other greater dangers, like the investment banks. But money market funds had provided much of the financing for those other sources of systemic risk.

  Now we saw that these funds could be a threat. On an FOMC videoconference the evening after the Countrywide drama, the new head of my markets division, Bill Dudley, warned that mortgage-related losses “could conceivably cause some funds to ‘break the buck,’ ” meaning they wouldn’t be able to redeem deposits at the fixed one-dollar-per-share value their investors took for granted. “In the worst case, this could even result in a run from these funds,” Dudley said.

  Even in those early days of the crisis, the financial system looked much more vulnerable to runs than we had appreciated. Markets were fleeing to safety, looking for cash or cash-like instruments with minimal credit risk, shunning illiquid securities tainted by links to mortgages in general and subprime mortgages in particular. Many of the instruments that were being unloaded had been highly rated and perceived as safe—and once these perceptions of safety changed, investors wanted out. Not everything was runnable, but enough was runnable to make the system extremely fragile.

  I KNEW from the crises of the past that financial turbulence was likely to damage the broader economy. Credit was already getting tighter for families and businesses. In talks with Ben Bernanke and Don Kohn, we all agreed we would need to ease monetary policy soon; the question was whether to cut rates right away or wait until the next FOMC meeting. With unemployment still just 4.6 percent, and most private forecasters expecting continued growth, an immediate rate cut between meetings seemed premature, and might be perceived as panicky. Ben was also concerned that it would be condemned as a bailout for risk-seeking investors, a “put” offsetting their recent losses and signaling a desire to protect them from future losses.

  At the time, our more pressing concern was the rapid erosion of liquidity within the financial system. We decided to try something unusual right away: to reduce the penalty rate that banks paid to borrow from the Fed’s discount window. That “discount rate” was usually set well above the federal funds rate, to limit its use to emergencies. We proposed to cut the discount rate by half a percentage point, to reduce the stigma for banks who feared that using the window would signal distress, while also extending the terms of the loans. We wanted our banks to remain as liquid as possible, so they’d be in a better position to lend. And we wanted the markets to see that the Fed was taking action.

  Ben and Don proposed the changes during that FOMC videoconference on August 16. These still felt like extremely modest responses, but some of our colleagues bristled. At one point, Richmond Fed President Jeffrey Lacker asked a question about how we expected the banks to respond, then turned to me. “Vice Chairman Geithner, did you say that they’re unaware of what we’re considering or what we might be doing with the discount rate?” he asked.

  Yes, I replied.

  “Vice Chairman Geithner, I spoke with Ken Lewis, president and CEO of Bank of America, this afternoon, and he said he appreciated what Tim Geithner was arranging by way of changes in the discount facility.”

  Lacker was basically accusing me of leaking our discount-rate plan. I hadn’t done that, but I had held a series of conversations with bankers and market participants over the previous few days. Some had suggested that we find some way to use the discount window, and I had asked how those ideas might be received in the market. I never would have disclosed an action in advance, but Lacker was understandably irked that he had first heard rumblings about possible use of the discount window from the CEO of a bank in the Richmond Fed district.

  This kind of tension and resentment within the FOMC would be another recurring theme, reminiscent of the emerging-market crises of the 1990s, when many European officials bristled at being presented with rescue plans the United States had shaped with the IMF management. Ben, Don, and I—along with Fed Governor Kevin Warsh, a former Morgan Stanley investment banker who often joined our strategy sessions—sometimes spent weeks developing ideas that we only shared with the rest of the FOMC on the verge of an announcement. That was often unavoidable and probably desirable given the circumstances, but our lack of consultation would add to the divisiveness of our policy debates.

  Lacker also had substantive objections to our approach, an early reflection of a real divide within the FOMC. The divide was partly a disagreement about the extent of the risk that the financial trauma in New York would pose for the overall economy, and partly a difference over whether the benefits
of trying to mitigate liquidity problems in markets were worth the moral hazard risks. At that meeting, Lacker was disturbed that our proposed statement, which warned that the risk of an economic slowdown had risen “appreciably,” did not mention the risk of inflation. “Did I miss something?” he asked derisively.

  Ben drily replied: “No.”

  Lacker was part of a group of hawkish regional Fed presidents—including Fisher of Dallas, Thomas Hoenig of Kansas City, and Charles Plosser of Philadelphia—whose main concerns were preserving the Fed’s inflation-fighting credibility and avoiding moral hazard. They did not expect a downturn, so they generally did not believe the upheaval in mortgages and capital markets justified lower interest rates. And they frequently deployed populist arguments against our lender-of-last-resort initiatives, starting with our plan to cut the discount rate.

  “To some extent, this looks to me like a sort of sham,” Lacker said. “We could easily be portrayed as helping banks make a lot of money on this.”

  Lacker noted that bankers often go running to Fed officials when markets wobble, but he urged us to ignore them. “I recognize that this turbulence in financial markets makes everyone apprehensive,” he said. “I realize the urge to act—to do something or at least be seen as doing something—can be irresistible. But I think we need to avoid the urge to play Mr. Fix It.”

  I don’t think I’m hawkish or dovish by nature. I’ve always been pretty pragmatic, suspicious of ideology in any form, and I took both halves of the Fed’s dual mandate seriously. But I found the more hawkish obsessions with moral hazard and inflation during a credit crunch bizarre and frustrating. Recession seemed like a more plausible threat than inflation. The notion that a slight tweak in the discount rate was too aggressive during an incipient panic just baffled me; in fact, it ended up being largely ineffectual at overcoming the stigma of the window, since only a handful of banks actually came and borrowed. I often joked privately during the crisis that the hawks on the FOMC thought they lived in Dubai at the peak of an oil boom. Sure, Kansas City and Dallas were still relatively calm, so perhaps from those vantage points—far from Wall Street and the global markets, less directly damaged by the housing collapse than the coasts—the risk of recession seemed remote. The inflation hawks seemed confident that the “real economy” could operate just fine no matter what happened in the markets, but given the scale of the credit and housing booms, it seemed more likely that financial fragility would lead to serious economic weakness.