Last week I participated in a conference hosted by the Federal Reserve Bank of Boston, at which I discussed the new lending programs and asset acquisitions pursued by the Federal Reserve over the last two years. Previously I shared with Econbrowser readers empirical evidence on the effects these targeted liquidity operations seem to have had. Below I reproduce my remarks from the conference on the underlying motivation for using such measures, in which I suggested that the critical question is what was the underlying cause of the financial stress to which the Fed was responding. I distinguished between two possible interpretations of how the financial crisis arose.
Perspective 1: Everybody just panicked
The first interpretation of what went wrong is that financial markets were pricing risk correctly in 2006 but began to overprice risk in 2007. Keister and McAndrews analyzed a situation in which banks out-of-the-blue stop lending to each other, while Gorton interpreted events in terms of a classic bank run, in which the liquidation value of entities is feared to have fallen below their short-run liabilities, creating an incentive for lenders to refuse to renew short-term credit. In the benign version of this theory, the troubled entities would in fact be solvent if it were not for the “fire-sale” prices at which distressed assets must be sold in such an environment. If allowed to proceed unchecked, these fears could prove self-fulfilling and result in a rapid collapse of credit.
In terms of appropriate policy responses to this problem, I would distinguish between actions that might have helped if implemented earlier in the decade and options that were available if we begin the analysis in the fall of 2007. If we are looking at what might have been done years earlier that could have helped, the obvious answer is to consider regulatory reforms that might have prevented financial markets from reaching a point at which the liquidation spiral could be set off in the first place. Bank panics are not an inevitable result of private financial intermediation. The key principle for avoiding them is to ensure that the liabilities of financial institutions consist not just of short-term borrowing, but also of equity contributed by the owners. As long as this equity cushion exceeds potential liquidation losses, there is no incentive for short-run creditors to rush to get their cash back, and no insolvency for the bank in the event that the bank does experience a run. It was a regulatory failure to allow an explosion of off-balance sheet entities that borrowed short and lent long but were immune from bank capital requirements.
On the other hand, if we ask what policy options were available after we had entered the fall of 2007, this particular policy prescription is of no help, as the horses were already out and the barn had no capital. Since there are profound negative externalities from simply watching asset prices and lending collapse, there would seem to be a clear case for the Fed to fulfill the function of lender of last resort, lending and buying assets where others won’t until the panic subsides and rational valuations return, and trying to do so in such a way that otherwise solvent enterprises were shielded from a panic bankruptcy.
Perspective 2: The core problem in credit markets preceded the crisis
An alternative perspective is that risk was incorrectly priced in the years leading up to the crisis with rationality only returning in 2007-2008. During 2004-2006 there was $2.7 trillion in new subprime and alt-A mortgage debt generated; (Ashcraft and Schuermann). Much of this was extended without documentation of the borrowers’ income, little or no money down, negative amortization, and called for huge increases in the borrowers’ monthly payments a few years into the loan. Yet somehow through the magic of securitization, this debt was repackaged into tranches that overwhelmingly received AAA credit ratings.
Such massive capital flows only made sense if one believed that house prices would continue to expand rapidly. Because this process was funneling such huge sums into the U.S. housing market, for a while house prices did just that, more than doubling between 2000 and 2005 according to the Case-Shiller 20-city house price index. U.S. household mortgage debt tripled in a little over a decade. According to this second interpretation, when house prices inevitably came crashing down, they brought with them defaults not just on the hybrid subprime and alt-A mortgages, but also put many otherwise sound borrowers underwater.
If it is claimed that the run-up in house prices and mortgage debt were a horrible miscalculation, what were the market failures that produced it? There is a long list of contributing factors. The originate-to-distribute model left the loan originators and securitizers with profits and lesser-informed buyers with the losses, creating agency problems; (Ashcraft and Schuermann). Intra-firm compensation schemes left decision-makers personally with the upside and stockholders with the downside, inducing excessive risk-taking; (Diamond and Rajan; Bebchuk and Spamann). The public-private GSEs Fannie Mae and Freddie Mac were woefully undercapitalized, giving private players the upside and the taxpayers the downside, and perhaps emboldening private securitizers to take even bigger risks (Hamilton). Both the compensation and procedures of the ratings agencies may have contributed to inaccurate perception of the safety of MBS (Ashcraft and Schermann), as did the mistaken perception that entities like AIG had the ability to insure against aggregate default risk. Moral hazard problems induced from the (ex post correct) belief that the U.S. government would absorb the downside on such gambles may have been another factor inducing excessive risk-taking.
If this perspective is the correct one, we can again distinguish between policies that would have made sense earlier in the decade and policies that were realistic options once we entered the crisis phase in 2008. If the above list of contributing market failures is correct, obviously addressing these with regulatory reforms before we reached the crisis point would have been the first-best option. On the other hand, if we condition on previous policy mistakes and ask what could have been done with options available in the fall of 2008, I disagree with those who reason that the way to correct the moral hazard problem is to hang tough in this situation and simply watch the losers go down. There are huge macroeconomic externalities from the resulting collapse of credit, which is why the government claiming it will not bail out the gamblers is not a credible strategy. Instead, this perspective suggests that the key policy question once we find ourselves in the fall of 2008 is how to allocate the necessary capital losses among lenders, stockholders, and the taxpayers in a way that minimizes the disruptive externalities of a credit collapse. If this is the correct perspective, the primary effect of targeted liquidity measures is simply to allocate these potential losses to the Federal Reserve. It is far from clear that this is the appropriate way for a democratic society to answer the question of who should bear the losses.
Finding the middle ground
I laid out the two perspectives above as diametrically opposed views. I nevertheless believe that the correct interpretation of events would acknowledge that each account contains some truth. It is hard to deny that there was some degree of misallocation of capital in the explosion of house prices and mortgage debt or that the resulting real estate price collapse was a key cause of the devaluation of securities and loss of bank equity that precipitated the banking panic phase. The remarks I presented at the Jackson Hole conference in August 2007 laid out precisely this scenario.
We might disagree on how much of that $2.7 trillion in new subprime and alt-A debt represented a malfunctioning capital market, and characterize the middle ground between the two views in terms of choice of a number between 0 and 2.7. If that number is big enough, it may be that no realistically feasible level of bank equity would have been sufficient to assure solvency in the face of a deterioration of confidence, and there is certainly the potential for fire-sale asset price deterioration and a necessary role for the Federal Reserve to fulfill its role of lender of last resort. But obviously from this hybrid perspective, the Fed is performing a combination of liquidity provision and residual loss absorption through these operations, and would want to undertake the latter only with extreme care and thoughtfulness.
Participants in this session were asked to address two basic questions. The first is whether the Fed’s targeted liquidity operations were necessary and effective. My answer is probably yes, though I would have a hard time persuading someone if they were not already convinced of that. The second question is whether such operations should be considered an important part of central banks’ arsenal of tools in the future. To that my answer is categorically no. From virtually any perspective of our current problems, it would have made far more sense to address these problems with proper regulatory supervision prior to the crisis instead of targeted liquidity operations after the crisis unfolds.
You can view my complete set of comments prepared for the conference here.