Henry Paulson - On the Brink - Inside the Race to Stop the Collapse of the Global Financial System

We discussed taking preventive measures. The Fed was exploring options for flooding the market with liquidity, or, as Tim said, “putting foam on the runway.” But with conditions as fragile as they were, I questioned whether there was much we could do to stabilize the markets if Bear went down suddenly.

We agreed to confer again first thing in the morning. Tim said, “We’ll have our teams working all night.” His staff would drill down on what a Bear failure might mean to the infrastructure—the markets for secured loans, derivatives, and such that constituted the unseen but vital plumbing of finance. It would be the first of many nights during the crisis when teams at the Fed—or Treasury—would work through the night against excruciating deadlines to try to save the system.

I couldn’t sleep. I was hot and agitated. I tossed and turned. I couldn’t stop thinking about the consequences of a Bear failure. I worried about the soundness of balance sheets, the lack of transparency in the CDS market, and the interconnectedness among institutions that lent each other billions each day and how easily the system could unravel if they got spooked. My mind raced through dire scenarios.

All financial institutions depended on borrowed money—and on the confidence of their lenders. If lenders got nervous about a bank’s ability to pay, they could refuse to lend or demand more collateral for their loans. If everyone did that at once, the financial system would shut down and there would be no credit available for companies or consumers. Economic activity would contract, even collapse.

In recent years banks had borrowed more than ever—without increasing their capital enough. Much of the borrowing to support this increase in leverage was done in the market for repurchase agreements, or repos, where banks sold securities to counter-parties for cash and agreed to buy them back later at the same price, plus interest.

While many commercial banks had big pools of federally insured retail deposits to rely on for part of their funding, the investment banks were more heavily dependent on this kind of financing. Dealers used repos to finance their positions in Treasuries, federal agency debt, and mortgage-backed securities, among other things.Financial institutions could arrange the repos directly with one another or through a third-party intermediary, which acted as administrator and custodian of the securities being loaned. Two banks, JPMorgan and Bank of New York Mellon, dominated this triparty repo business.

The market had become enormous—with perhaps $2.75 trillion outstanding in just the triparty repo market at its peak. Most of this money was lent overnight. That meant giant balance sheets filled with all kinds of complex, often illiquid assets were poised on the back of funding that could be pulled at a moment’s notice.

This hadn’t seemed like a problem to most bankers during the good times that we’d enjoyed until the previous year. Repos were considered safe. Technically purchase and sale transactions, they acted just like secured loans. That is to say, repos were considered safe until the times turned tough and market participants lost faith in the collateral or in the creditworthiness of their counterparties—or both. Secured or not, no one wanted to deal with a firm they feared might disappear the next day. But deciding not to deal with a firm could turn that fear into a self-fulfilling prophecy.

A Bear Stearns failure wouldn’t just hurt the owners of its shares and its bonds. Bear had hundreds, maybe thousands, of counterparties—firms that lent it money or with which it traded stocks, bonds, mortgages, and other securities. These firms—other banks and brokerage houses, insurance companies, mutual funds, hedge funds, the pension funds of states, cities, and big companies—all in turn had myriad counterparties of their own. If Bear fell, all these counterparties would be scrambling to collect their loans and collateral. To meet demands for payment, first Bear and then other firms would be forced to sell whatever they could, in any market they could—driving prices down, causing more losses, and triggering more margin and collateral calls.

The firms that had already started to pull their money from Bear were simply trying to get out first. That was how bank runs started these days.

Investment banks understood that if any questions arose about their ability to pay, creditors would flee at wildfire speed. This is why a bank’s liquidity was so critical. At Goldman we had absolutely obsessed over our liquidity position. We didn’t define it just in the traditional sense as the amount of cash on hand plus unencumbered assets that could be sold quickly. We asked how much money, under the most adverse conditions, could disappear on any given day; if everyone who could legally request their money back did so, how short would we be and could we meet our obligations? To be on the safe side, we kept a lockbox at the Bank of New York filled with bonds that we never invested or lent out. When I was CEO at Goldman, we had amassed $60 billion in these cash reserves alone. Knowing we had that cushion helped me fall asleep at night.

Bear had started the week out with something like $18 billion in cash on hand. It now had closer to $2 billion. It couldn’t possibly meet demands for withdrawals. And in the morning, when the markets opened, no counterparties were going to lend to Bear: they’d all be pulling their money out. This would be bad news indeed, not just for Bear Stearns, but for every institution dealing with them.

No wonder I slept no more than a couple of hours that night. I had never had trouble before, but this night was the beginning of a prolonged bout of sleeplessness that would haunt me throughout the crisis, and particularly after September. On tough days, I would fall asleep exhausted around 9:30 p.m. or 10:00 p.m., then wake up several hours later and lie awake for much of the rest of the night. Sometimes I did my clearest thinking during these hours, occasionally getting up to write things down. By the time the newspapers were delivered at 6:00 a.m., I would have already been up for an hour or two, often turning on cable TV to check on overseas markets.

On Friday morning I had just shaved and was about to get in the shower when the phone rang. It was Bob Steel telling me that a conference call would start around 5:00 a.m. Still wearing the boxer shorts and T-shirt I slept in, I jogged up to the third-floor study of our house so I wouldn’t wake Wendy. On the line were Tim Geithner, Ben Bernanke, Kevin Warsh, and Don Kohn from the Fed; Tony Ryan and Bob Steel from Treasury; and Erik Sirri from the SEC. We waited at first for Chris Cox, who was standing by in his office but never came on because of a communications mix-up. For a few minutes, we plugged in Jamie Dimon, CEO of JPMorgan, Bear’s clearing bank. He painted a dark picture, emphasizing that a Bear Stearns failure would be disastrous for the markets, and that the key was to get Bear to the weekend.

Once Jamie got off, Tim reviewed a creative way he and his team had devised to buy time. The Fed would lend money to JPMorgan, which in turn would lend the money to the beleaguered brokerage firm. To make this work, the Fed’s loan would have to be non-recourse: it would be backed by collateral from Bear, but neither JPMorgan nor Bear would be liable for repayment.

By law the Federal Reserve can lend against assets only when the loan is secured to its satisfaction, meaning in practical terms that there is a minimal chance of the Fed’s losing money. But if this loan could not be repaid, for whatever reason, and the Fed had to sell the collateral for less than the value of the loan, the central bank would incur a loss. It would be a bold, unprecedented action for the Fed to make such a deal.