President, Federal Reserve Bank of Richmond
National Association for Business Economics
2009 Washington Economic Policy Conference
The U.S. economy is now in the second year of a recession that began at the end of 2007.1 The deterioration in economic activity has been particularly sharp since September of last year. But throughout this downturn, a singular feature has been the extent of the disruption to financial markets and losses suffered by financial institutions. The financial dimension of this contraction has brought an historic expansion in government lending to financial market participants, mostly through an expanding array of Federal Reserve initiatives. In contrast, the Fed's response to most recessions in recent decades has been limited to adjustments of the target federal funds rate. We've done that in this cycle, too, bringing the funds rate target from 5 ¼ percent in September of 2007 to between zero and 25 basis points now.
These two dimensions of the Fed's response are interconnected, since both involve the use of our balance sheet, but they have different impacts on the economy. So today, I'd like to speak about the economic effects of government lending and how that relates to our broader monetary policy goals. In doing so, I'll note what I see as important differences between monetary and credit policy and I’ll offer my thoughts on the Fed's role in the extension of government credit. And I'm sure you won't be surprised to hear that these are my own views and not necessarily those of any of my colleagues on the FOMC.
Views about the role of government credit in promoting financial and macroeconomic stability tend to be shaped by views about the role of credit in the business cycle. In one popular view, credit market disturbances, such as the recent rise in losses on mortgage-backed securities, cause banks and other credit intermediaries to pull back credit supply as they attempt to repair their balance sheets. The reduction in lending to households and firms forces them to reduce their spending on goods and services and creates an additional drag on growth. An alternative view is that shocks to the economy affect spending more directly, and that as growth declines, the creditworthiness of households and firms deteriorates, causing credit flows to fall and spreads to widen.
These two views represent opposite directions of causality between credit and aggregate spending. In reality, both of these directions may well be in operation at the same time, and determining the quantitative importance of each is very hard. But my reading of recent events emphasizes the second view, in which the effect of slowing growth on credit conditions predominates. This view has received much less attention than it deserves, I believe, so let me say a few words about the current cycle in light of this issue.
The antecedent of the contraction we are in was the boom in home sales, prices and construction coming out of the last recession. Untangling the causes of that boom poses research challenges that will launch a thousand dissertations, I expect. The most plausible suspects at this point include financial innovation, regulatory laxity, accommodative monetary policy and a global savings glut; but all worked through the expanding availability of mortgage credit. Even with favorable financing conditions though, the increasingly leveraged purchases of homes would not have made sense without confidence – in hindsight misplaced confidence – in a continued upward path for home prices.
Whatever the causes of the boom, the result was what turned out to be a glut of housing, which, as people’s beliefs about demand growth adjusted, led to historic declines in prices. The most immediate effect was a collapse in residential investment, and large consequent declines in employment in construction and related sectors. The reduction in home owners’ wealth as home prices declined, together with growing uncertainty about labor market prospects, caused household spending to slow beginning in mid-2007, and then decline outright in mid-2008. The dimming outlook for consumer spending also dampened business investment spending in turn, and spread the employment slowdown beyond the residential construction sector. This, in turn, further dampened consumer spending.
These trends reduced prospects for and increased uncertainty about household incomes and firm revenues. As a result, households and firms are riskier lending prospects than they were a couple of years ago, given the change in the overall macroeconomic environment. Note that surveys that ask lenders whether they have “tightened terms” in recent months do not really get at this question. Any given profile of borrower characteristics – income, balance sheet and credit score, for example – is likely to translate into a riskier loan now, so banks are likely to have tightened qualification cut-offs even without any reduction in their risk appetite.2
The downturn in home prices in many regions has resulted in increased losses on home mortgages, particularly subprime mortgages. Uncertainty about the ultimate depth of the decline in home values has meant ongoing uncertainty about the magnitude of aggregate losses that will be realized on mortgage-related assets. Financial market participants have also faced uncertainty about where these losses will turn up. Mortgage risks were split up and spread widely, both within the United States and abroad, through securitization and use of the insurance capabilities provided by credit derivative contracts, making it difficult to assess any individual institution’s share of the aggregate exposure. In addition, financial market participants have at times faced uncertainty about prospective public sector intervention. The disparate responses to potential failures at several high-profile organizations may have made it difficult for market participants to forecast whether official support would be forthcoming for a given counterparty, and where in the capital structure that support would land.
Most of what has been observed in financial markets since the summer of 2007 seems readily intelligible as a consequence of the increased uncertainty facing market participants resulting from the significant economic downturn. Apprehension about potential losses caused lenders to demand higher risk premia in interbank credit markets for institutions with at least some presumed mortgage-related exposure. Market participants became especially concerned about the heightened risk associated with lending at longer maturities, and so risk premia became especially elevated for term lending. Some borrowers were unwilling to pay higher premia for term loans, and shortened the tenor of their funding. Others sought to protect themselves against an erosion in counterparties’ perception of their creditworthiness by paying the unusually high premia in order to “lock in” funding or by hoarding liquid assets despite high opportunity costs. More broadly, the proliferation of intermediation channels in recent years has meant that for many borrowers, the next best financing option may not be much more costly. For example, many commercial paper issuers have back-up lines of credit with banks that they can draw on in the event they are unsatisfied with market pricing. Thus observing that a given intermediation channel is “frozen,” “clogged,” or “dried up” may not indicate dysfunction, per se, but may indicate instead just a portfolio reallocation in response to a shift in risk assessments.
In response to the credit market turmoil, the Federal Reserve, the Treasury and the FDIC have undertaken a sequence of interventions. Some of these have been fairly direct extensions of the Fed’s standard discount window lending – such the Term Auction Facility, or even the primary dealer credit facility, which extended credit to a specific set of non-bank financial institutions. Others involve the use of Fed credit to support specific classes of assets – for instance, commercial paper and now asset backed securities.3 Still others, of course, have provided direct assistance to specific institutions, such as Bear Stearns and AIG. Citibank and Bank of America obtained asset guarantees provided jointly by the Fed, the FDIC and the Treasury.
Each of these credit programs involves Federal Reserve lending or lending commitments. Before last October, the Fed was able to “sterilize” new lending through offsetting asset sales that soaked up the additional bank reserves, which otherwise would have increased the monetary base and pushed the federal funds rate below its target at the time. After October, the cumulative amount lent became too large to sterilize, and further lending added to the monetary base. Luckily, and perhaps not coincidentally, the implementation of these large credit programs has coincided with a time in which additional monetary stimulus is warranted.
Even though the conventional measure of the stance of monetary policy is the central bank’s interest rate target, monetary policy fundamentally is always about the amount of monetary liabilities issued by the central bank – also known as the “monetary base.” After all, hitting an interest rate target requires varying the quantity of central bank money, reducing the supply to raise rates and increasing the supply to reduce rates. Even when the policy rate has been driven down to zero, central banks can still dictate the supply of central bank money. And changes in the monetary base can still provide economic stimulus. Even if the funds rate does not change in response to an increase in the monetary base, some other rates of return must change to induce banks to voluntarily hold the additional supply of bank reserves.
These government lending programs, by targeting particular market sectors, alter the allocation of credit across markets. Consequently, while some market segments benefit from reduced funding costs, others may actually see their costs rise as credit is diverted to those markets that have been targeted for support. An alternative approach to expanding the monetary base is to do it in a way that is more neutral across market segments. Since risky financial assets are presumably priced in relation to U.S. Treasury securities, which are free of credit risk, purchasing Treasuries is likely to have little effect on the relative credit spreads on different financial instruments. This is one reason I expressed my preference, in my dissent at the last FOMC meeting, for managing the monetary base by purchasing U.S. Treasury securities rather than through targeted credit programs.
Another reason for that preference is that targeted credit programs, in addition to their immediate effect on the allocation of credit and resources across market segments, contribute to the moral hazard problem inherent in the provision of government-funded credit or guarantees. Safety net support for financial institutions encourages private market participants to view some institutions as “too big to fail,” and weakens those institutions’ incentive to monitor and manage the risks they face in their business strategies and financial market transactions.4 Intervention to support particular asset classes similarly weakens incentives by encouraging private market participants to discount the cost of credit losses that would depress asset prices. And this weakening of incentives, by inducing greater risk-taking, eventually increases the ultimate cost of providing safety net protection.
Some have questioned the empirical relevance of moral hazard in the current episode of financial market turmoil, citing the fact the equity holders in institutions receiving government support have suffered significant – and in a few cases, total – losses. But this observation misses an important point about the moral hazard effects of government lending. The most direct effect of government credit or guarantees is to lower the cost of private credit to protected borrowers. Limiting the circumstances in which the benefiting institution will have insufficient liquidity to survive, means that private debt holders bear less risk and have less incentive to constrain risk-taking by the borrower. Absent regulatory constraints, this encourages those institutions to take on more risks than they otherwise would, including by becoming more highly leveraged. The more leveraged a firm, the greater the incentive of management and equity to take on risk. Indeed, limited liability means that equity holders could find a negative net present value investment to be worthwhile, if it is risky enough. So the fact that equity has absorbed large losses is not strong evidence against moral hazard effects if the safety net protects – and reduces the cost of – debt. Indeed, the risk-shifting effect of moral hazard makes large losses to equity more likely to occur, because it makes large gambles more attractive.
It is also commonly argued that moral hazard, while a real and undesirable consequence of the safety net, is an unavoidable cost of the need to respond forcefully to prevent the disorderly failure of a large complex institution. This view seems to be based on the assumption that the systemic risks posed by large financial institutions – stemming from concentrated and correlated exposures among a complex web of counterparties – are an inherent and unavoidable trait of modern financial markets. But the pervasiveness of such characteristics can itself be an endogenous consequence of moral hazard. If systemic risks at large financial institutions are particularly protected by the safety net of government credit, then such institutions will have an extra incentive to acquire precisely those risks. This may be why the unexpectedly large exposures of large banking organizations to home-mortgage-related risks stemmed from their provision of backstop liquidity commitments to a wide array of off-balance-sheet securitization arrangements. Institutions that are viewed as too big to fail may have had a comparative advantage in supplying contingent liquidity that was most likely to be needed in the event of dire macroeconomic shocks because those are the circumstances most likely to elicit broad-scale government lending support.
The regulatory and supervisory regime surrounding institutions that benefit from access to government lending support plays a critical role in constraining and preventing the excessive risk-taking that would otherwise be induced by the moral hazard effects of that support. The dramatic recent expansion in government lending has extended safety net support beyond the set of institutions previously covered by that regime. If no corrective action is taken, the next economic expansion would likely see more excessive risk-taking that could again destabilize the financial system. It is critical that the scope of regulatory and supervisory oversight should match the extent of access to government credit support in order to contain moral hazard effectively. For that match to take place, the boundaries of government credit support need to be well-defined. In my view, it would be preferable for those boundaries to be rolled back.
The Federal Reserve System has been digesting and analyzing the lessons learned from the recent episode and is exploring steps to strengthen supervisory practices and processes accordingly. More broadly, many observers have urged a restructuring or reengineering of our approach to financial regulation.5 No doubt much will be said about financial regulatory reform in the weeks ahead, and a discussion of the relevant issues would be beyond my scope here. I will just offer the observation that restoring compatibility between the scope of government support and the scope of government supervision seems essential to a healthy and sustainable financial system. That vantage point suggests that when we do get around to considering concrete proposals for our financial regulatory structure, our choices about whom to regulate and how to regulate them ought to be driven by our decisions about who is eligible for public sector credit and under what conditions. In 1983, John Kareken of the University of Minnesota and the Minneapolis Fed described financial deregulation as “putting the cart before the horse,” suggesting that expanding the powers of banking and thrift institutions without appropriate attention to design of the financial safety net could be a risky move.6 His analysis was prescient, given the savings and loan debacle that followed later in that decade. Karaken’s emphasis was on deregulation in the presence of deposit insurance, but in the current episode, lending by the Fed and the Treasury has become just as important a part of the federal financial safety net. Nobody is talking about deregulation now, but the same principle applies: namely, redesigning our financial regulatory system before establishing the boundaries of the financial safety net would be like putting the cart before the horse.
I have spent some time discussing government lending, but the title of my talk is “Government Lending and Monetary Policy,” so I would like to say a few words now about the relationship between the two. Earlier, I described how the dramatic expansion in Federal Reserve Bank lending in the last few months has caused a dramatic increase in the size of our collective balance sheet and the monetary base.7 I noted that this is a time in which additional monetary stimulus is needed, and so the two strategies are not in conflict.
Nevertheless, monetary policy and credit policy are two different things. Monetary policy consists of changes in the monetary base – the sum of outstanding currency and bank reserves. Credit policy, in contrast, changes the Fed’s assets while holding the amount of the monetary base fixed – sterilized lending is an example. Monetary policy aims at keeping the price level stable and relatively predictable, and by doing so, contribute to maximum sustainable economic growth. Credit policy is also aimed at promoting growth, but it is a form of fiscal policy in that it uses the public sector’s balance sheet to alter the allocation of resources.
As I said, Federal Reserve lending has been financed to a large degree recently by increases in the monetary base, but that lending could just as well be performed by the U.S. Treasury and financed by the issue of Treasury securities.8 No immediate change in the assets and liabilities of the public would be required, since the additional amount of debt the Treasury issued would exactly match the additional need for assets by the Federal Reserve Banks if the monetary base were to remain unchanged.
There is one significant difference between lending performed by the U.S. Treasury and lending performed by the Federal Reserve Banks, however. The Treasury can lend only under explicit authorization from Congress. The Federal Reserve, in contrast, has independent control of its balance sheet and funds itself outside of the normal appropriations process. That independence was affirmed in the 1951 Fed-Treasury Accord, which freed the Fed from an obligation to suppress interest rates for the purpose of limiting the cost of public debt.9 Central bank independence is now widely recognized as an important mechanism for insulating monetary policymaking from inflationary political pressures, and allowing it to respond quickly to short-run macroeconomic developments.
A New Accord?
This observation led my former colleague, Marvin Goodfriend, to argue 15 years ago for transferring much of the Fed’s lending activities to the Treasury. He wrote:
“Congress bestows such independence only because it is necessary for the central bank to do its job effectively. Hence, the presumption ought to be that the Fed should perform only those functions that must be carried out by an independent central bank.”10
While both the Fed and the Treasury can extend credit, only the Fed issues money. Thus, the Fed’s primary focus should be the management of its monetary liabilities.
Goodfriend advocated an understanding or agreement between Fed and Treasury on credit policy, analogous to the 1951 Accord.11 A new “credit accord” that assigns to the Treasury the responsibility for all but very short-term lending to solvent institutions would have a number of advantages, I believe. On a practical level, at some point in the future, the Fed will need to withdraw monetary stimulus to prevent a resurgence of inflation when the economy begins to recover. That time could arrive before credit markets are deemed to be fully enough “healed” to warrant winding down particular credit programs. If monetary policy and credit programs remain tied together, as they currently are, we risk having to terminate credit programs abruptly, or else compromise on our inflation objective. Separating credit programs from monetary policy would make it easier to devise a successful “exit strategy,” and would reduce market uncertainty about how any potential tension between monetary and credit policy will be resolved.
Government lending, whether by the Fed or by the Treasury, fundamentally represents fiscal policy in the sense that it channels taxpayer funds to private sector entities. The presumption ought to be that such lending is subject to the checks and balances of the appropriations process laid out in the Constitution. Using the Fed’s balance sheet is at times the path of least resistance, because it allows government lending to circumvent the Congressional approval process. This risks entangling the Fed in attempts to influence credit allocation, thereby exposing monetary policy to political pressures.
Granted, there are circumstances in which timeliness precludes explicit Congressional authorization. But the understanding could stipulate that the emergency lending is transferred to the books of the Treasury after a brief period of time has elapsed. Moreover, the Treasury could be given explicit line of credit authority, as they now have for Fannie Mae, Freddie Mac and other entities. For longer-term credit programs that are meant to support specific market segments, and which are designed and implemented over a period of weeks, there would not seem to be any practical impediment to seeking explicit Congressional authorization.
I spoke earlier of my sense that the scope of the financial safety net will need to be scaled back. This could be difficult. When a financial crisis threatens one or more institutions that appear to pose “systemic” risks to a broad array of counterparties, it can be hard to contemplate not intervening. But it may also be the case that the systemic risks are partly the result of expectations about the likelihood of government intervention. This is a classic example of a so-called time consistency problem.12 One would like market participants to believe you are committed to resist lending, though following through later will be difficult. A credit accord could help limit the financial safety net by placing a hurdle in the way of intervention beyond a well-defined set of circumstances. Transferring authority for most government lending to the U.S. Treasury and subjecting that authority to a legislated framework can help commit authorities to a bounded government safety net.
Raising questions about the efficacy of government lending, as I have done, does not imply a view that financial markets are working perfectly. Indeed, financial markets are undergoing tremendous strains as they adjust to large and hard-to-predict losses. I think the fundamental problem with our financial system involves how our large institutions accumulated such large, concentrated exposures. Regulatory shortcomings seemed to contribute, and much needed attention will be given to this problem in the coming months. But part of the story of how and why institutions exposed themselves to the losses they are now experiencing has to do with the incentive effects of the financial safety net. And I hope that this problem also gets the attention it deserves.
I am grateful to John Weinberg for help in preparing this speech.
For a model in which it is optimal for credit standards vary over the business cycle, see John Weinberg, “Cycles in Lending Standards?” Federal Reserve Bank of Richmond Economic Quarterly, Summer 1995, vol. 81, no. 3, pp. 1-18.
For details on the Fed's credit market interventions see www.federalreserve.gov/monetarypolicy/bst.htm.
Gary H. Stern and Ron J. Feldman, "Too Big to Fail: The Hazards of Big Bank Bailouts" (Washington, DC: The Brookings Institution Press, 2004).
Working Group on Financial Reform, "Financial Reform: A Framework for Financial Stability" (Washington DC: Group of Thirty, January 15, 2009).
John Kareken, “Deposit Insurance Reform or Deregulation is the Cart, Not the Horse,” Federal Reserve Bank of Minneapolis Quarterly Review, Spring 1983, vol. 7, no. 2, pp. 1-9.
The two do not necessarily move together one-for-one. Since last September the U.S. Treasury has issued notes through the Supplementary Financing Program and deposited the proceeds with the Federal Reserve, effectively draining reserves from the banking system.
Although Federal Reserve staff may have unique skills and expertise, there is ample precedent for sharing skills with the Treasury; for example, the Treasury contracts with the Federal Reserve Banks to perform fiscal agency functions.
For narrative accounts of the events leading to the 1951 Treasury-Federal Reserve Accord, see the articles by Robert L. Hetzel and Ralph F. Leach in the Federal Reserve Bank of Richmond Economic Quarterly Special Issue commemorating the 50th anniversary of the Accord (Winter 2001). That issue also contains the article by Marvin Goodfriend cited below, as well as a related article by J. Alfred Broaddus and Goodfriend, “What Assets Should the Federal Reserve Buy?” The associated Web site contains historical documents, biographies, and additional material. Go to Research and click on “Treasury-Fed Accord Special Report.”
Marvin Goodfriend, “Why We Need an ‘Accord’ for Federal Reserve Credit Policy: A Note,” Federal Reserve Bank of Richmond Economic Quarterly, Winter 2001, vol. 87, no. 1, see p. 24. Reprint of article in Journal of Money, Credit, and Banking, vol. 26 (August, 1994).
See also Charles Plosser, “Ensuring Sound Monetary Policy in the Aftermath of Crisis.” Speech to U.S. Monetary Policy Forum, New York, N.Y., February 27, 2009.