Professor Larry Ball, our friend and colleague, has written a fascinating monograph reexamining the September 14, 2008 failure of Lehman Brothers. Following an exhaustive study of documents from a variety of sources, Professor Ball concludes that the Fed could have rescued Lehman. The firm had sufficient collateral to meet its liquidity needs, and may have been solvent. The implication is that the worst phase of the financial crisis was preventable. (A short summary is available here.)
We are skeptical on several fronts—that Lehman was solvent, that policymakers had authority to lend to an insolvent institution, and that doing so would have limited the financial crisis.
Ball’s analysis can be divided into two parts. First, could the Federal Reserve legally have lent to Lehman? And second, would doing so have been wise? We lack the expertise to evaluate strict claims of legality. That said, our sense is that the answer to these questions turns on whether Lehman was solvent. If Lehman was not solvent, then any credit provision lies in the province of the fiscal authority. Indeed, even if the central bank were to have the authority to make a “loan” to an insolvent institution, doing so is likely to have horrible consequences.
To understand why, we turn to Paul Tucker, who in May 2009, in the midst of the crisis, said:
“Bagehot’s famous dictum, in Lombard Street, was that, to avert panic, central banks should lend early and freely (i.e. without limit) to solvent firms, against good collateral, and at ‘high rates.’” (our emphasis)
In Tucker’s formulation, central bank lending has two requirements: sufficient collateral and solvency. The first (as Bagehot indicated) protects the lender, but is only an imperfect proxy for the second, which protects the integrity of the financial system as a whole.
There are three big reasons that a central bank should not lend to a bankrupt institution. The first is that, by lending secured to an insolvent commercial bank, the central bank further subordinates bondholders. It does this both by allowing short-term depositors to run and by inserting itself ahead of others in the queue for claiming repayment when failure inevitably comes. Such an action picks winners and losers. In democracies, this is the prerogative of elected officials, not central bankers.
Second, lending to an insolvent bank by itself does not put an end to its fragility. Ultimately, the institution must be liquidated or recapitalized. Postponing this resolution is usually costly. The capital hole itself tends to widen when an impaired institution freezes up in a financial crisis. More broadly, when resolution is delayed, the resulting mix of uncertainty and poor incentives damages both the financial system and economy. As we have recently argued, weakly capitalized banks don’t lend to healthy borrowers.
Third, when people find out that the central bank is willing to lend to insolvent banks—and they will find out—then any bank that borrows will be suspected of being bankrupt. The resulting stigma will undermine the critical function of the lender of last resort as a lender to solvent, but illiquid banks. In the end, only those that are bankrupt will borrow and the central bank’s lending facility will become useless. (See Tucker’s more recent discussion here as well.)
So, was Lehman solvent? Having sufficient collateral to meet its immediate liquidity needs, as Professor Ball argues, does not answer this question, which he addresses separately (pp. 53 to 78 of the monograph). Ball concludes that, at the time of the bankruptcy, Lehman’s equity was between -$2 and +$13 billion (pg. 65). This estimate appears very hopeful considering that, in the months prior to the bankruptcy, Lehman had unsuccessfully sought suitors from around the world. This search for new capital required showing the firm’s books to potential investors. As the September failure approached, U.S. government officials urged other intermediaries to evaluate Lehman’s assets and liabilities with an eye toward taking a stake. Leaving aside the possible bid by Barclays (which British regulators blocked), there were no buyers. This market outcome is consistent with the 2013 Cline and Gagnon judgment that the firm was deeply insolvent: they estimate that Lehman, which had assets in excess of $600 billion on the eve of bankruptcy, had net worth between -$100 billion and -$200 billion.
There is certainly room for debate, but we see the question of whether Lehman’s net worth was negligible or sharply negative as ancillary to the real issue. In important ways, lending to a bank of doubtful solvency is little different from lending to one that is certainly so. It will continue to put other institutions, and therefore the system, at risk. In this case, central bank lending to Lehman, an institution widely thought to be bankrupt, would have tarnished everyone else.
It is tempting to point fingers at specific policymakers who failed in their effort to pilot a leaking boat through a hurricane. Aside perhaps from the SEC, which claimed that Bear Stearns was well capitalized even after its March 2008 failure, we are loathe to conclude that there is a villain in the story by the time that Lehman failed in September 2008. At that advanced stage of the crisis, neither policymakers nor the courts had the tools to quickly resolve an insolvent intermediary of Lehman’s legal form, scale and complexity.
When then-Secretary of the Treasury Henry Paulson said he didn’t wish to be “Mr. Bailout,” could he have done any differently? Did the Treasury have the authority to save Lehman? For example, could the mechanism employed to support the JP Morgan Chase purchase of Bear Stearns have been used again? In the Bear episode, the Treasury was perceived as having indemnified the Fed against any potential losses. Specifically, on March 17, 2008, a letter from Secretary Paulsen to then-New York Fed President Timothy Geithner included the following:
“The Federal Reserve Bank of New York (''FRBNY") has arranged a special credit facility with J.P. Morgan Chase Bank ("JPMCB") to assist with the acquisition of The Bear Stearns Companies. On behalf of the Department of the Treasury !support [SIC] this action as appropriate and in the government's interest, and acknowledge that if any loss arises out of the special facility extended by the FRBNY to JPMCB, the loss will be treated by the FRBNY as an expense that may reduce the net earnings transferredby the FRBNY to the Treasury general fund.”
That is, the Treasury admitted that any losses arising from FRBNY lending made to support the acquisition of Bear Stearns assets would be borne by the federal government, rather than the Federal Reserve. (Operationally, this would show up as a reduction in the interest income that the Fed turns over the Treasury on a regular basis.)
But, in March 2008, observers appeared to believe that Bear Stearns was solvent. Had they thought otherwise, under what authority could government officials have committed federal resources without the authorization of Congress? We don’t know of any. From this perspective, to the extent that officials failed in the Lehman episode, it was in their inability to persuade Congress to create an effective crisis resolution mechanism. Lacking such a mechanism, the politics of seeking congressional approval for ad hoc bailouts are simply awful. To see the point, recall Congressional resistance to the workout of the government-sponsored enterprises, which entered into federal conservatorship just days before Lehman failed. Even after Lehman, Congress initially rejected the Troubled Asset Relief Program, until the ensuing stock market collapse prompted a re-vote.
More important, it is doubtful that rescuing Lehman would have done more than postpone an intensification of the crisis. To be sure, a Lehman rescue probably would have buoyed some asset prices. However, by September 2008, the financial system as a whole lacked adequate capital to withstand even minor disturbances. As the chart below shows, the NYU Stern School Volatility Lab’s measure of the capital hole in the U.S. financial system (SRISK) peaked at more than $900 billion at the end of July 2008, more than a month before Lehman’s failure. This exceeded the aggregate capital in the top 25 domestically chartered banks in the United States. At that stage, Lehman wasn’t even in the top 10 of U.S. intermediaries as ranked by SRISK. Consequently, if Lehman had been saved, AIG, Bank of America, Citigroup, Goldman Sachs, JP Morgan Chase, Merrill Lynch, Morgan Stanley, and others would likely have needed assistance in relatively short order. Put another way, by July 2008, the U.S. financial system was a tinderbox waiting for a spark—there was no need for anyone to pour gasoline on the kindling to assure a conflagration would ensue.
U.S. SRISK (Billions of dollars), 2007-2009
Of course, we cannot rule out that on September 14, 2008 Lehman was barely solvent, as Professor Ball argues. And we share his conviction that looking back at this episode is very useful. But, from our perspective, the question is not whether officials could have done a far better job at the moment of Lehman’s failure. As we wrote two years ago, the lesson is about the flaws in the system that made it so vulnerable and so difficult to stabilize and repair.
Addressing these design flaws required changes in two broad categories. First, the system lacked an effective mechanism for early intervention in cases where the capital level of a nonbank financial intermediary fell precipitously low. And second, it lacked a procedure for safely resolving such an institution. While the Dodd-Frank Act of 2010 made progress on both fronts, significant problems remain (see, for example, our comments on the shortcomings of Dodd-Frank, on “living wills,” and on the difficulties of designating non-bank SIFIs). Only when these challenges are fully addressed will the key lessons of Lehman have been absorbed.