Donald L. Kohn, Vice Chairman
At the Kellogg Distinguished Lecture Series, Kellogg School of Management, Northwestern University, Evanston, Illinois
November 16, 2009
Policy Challenges for the Federal Reserve
I titled my talk "Policy Challenges for the Federal Reserve." I did that before I knew what I was going to discuss, because just about any topic involving the Federal Reserve would entail policy challenges. So let me begin by narrowing the topic just a bit: I would like to talk about the challenges that lie at the intersection of monetary policy and financial stability.
Our economy is just beginning to recover from a horrendous episode in which mispricing of risk--especially in home mortgage lending, but more broadly as well--led to financial instability that in turn led to a severe recession. Much attention is, appropriately, being focused on the implications of this episode for the reform of financial regulation. The Congress, financial regulators, and analysts are debating those critical issues.
I will focus on some possible implications of the recent episode for monetary policy. The questions I want to address are, first, how should we take account of financial stability in the conduct of monetary policy--for example, should financial stability be another specific responsibility of monetary policy, in addition to its responsibilities for promoting maximum employment and stable prices? And second, what do the crisis and our response imply for the monetary policy tools of the Federal Reserve? I don't have the answers to these questions, but I thought it would be useful to discuss them.1 Importantly, I speak only for myself, not for my colleagues on the Federal Open Market Committee.2
Monetary Policy and Financial Stability
Traditionally, most analysts thought that when monetary policy successfully promoted economic stability, it would also promote financial stability. When business cycle swings are moderate and inflation rates are low and predictable, households and firms can make and follow through on long-term plans for spending, saving, and investment. Lending institutions can better evaluate the likely profitability of capital projects, thus reducing their risk of credit loss. And the moderate swings in long-term interest rates and other asset prices that were thought to be fostered by such an environment should tend to limit the exposure of balance sheets.
However, although monetary policy over the past several decades did help foster economic stability, some have questioned whether it also contributed substantially to the lax practices that led to the buildup of financial vulnerabilities and the resulting severe recession. Two aspects of the Federal Reserve's conduct of monetary policy are cited by these critics. One, some assert that we responded asymmetrically to movements in asset prices, easing aggressively in response to declines and not tightening correspondingly when asset prices rose. This perceived asymmetry is alleged to have given market participants the sense that they could engage in a one-way bet--that the Federal Reserve would cushion asset-price declines with cheap money but would not increase interest rates to make it more expensive for them to finance bets that asset prices would rise. Second, some believe that the Federal Reserve kept policy too loose around 2003 and then tightened too slowly and predictably in 2004 and 2005, in effect encouraging if not underwriting the bubble in house prices that lay behind so many of our troubles over the past two years.
My perspective is that policymakers have kept their eyes firmly on medium-term macroeconomic stability--that is, on the legislated objectives of maximum employment and stable prices. We responded to developments in asset and credit markets to the extent that they affected the macroeconomic outlook, but we did not assign them extra weight in our deliberations. In particular we did not react asymmetrically to asset-price developments; it only looks that way on the surface because asset prices tend to rise slowly and fall rapidly. Had we been more ready to ease than to tighten monetary policy, the economy would have tended to run above its productive potential on average and inflation would have risen. In fact, inflation fell over the 1980s and 1990s.
Inflation did rise from 2004 through the middle of 2008, but largely because the prices of petroleum and other commodities increased. Those increases reflected rising demand from rapidly industrializing emerging market economies, not easy monetary policy in the United States. As we went through this period, a variety of measures of resource utilization, underlying inflation rates, and market indicators of future commodities prices gave us good reason to believe that our policy was consistent with a return of headline inflation to a low and stable underlying trend. We did raise interest rates gradually, in accord with our announcements that the federal funds rate was likely to rise at a "measured pace," but following this path was dependent on economic conditions evolving about as we anticipated. Similarly, in the current circumstances, the Federal Open Market Committee has emphasized that its expectation that the federal funds rate is likely to remain at an exceptionally low level "for an extended period" depends on the outlook for resource utilization, inflation, and inflation expectations following the trajectories we expect.
A policy focused on price and economic stability over the medium run and that responds symmetrically to inflationary and deflationary shocks can nonetheless contribute to excessive leverage, risk-taking, and maturity transformation, and indeed that appears to have happened; to some extent, we were the victims of our own success. Improved monetary policy after the 1970s, along with a host of other factors, including technological advances and greater global integration of capital and product markets, helped reduce the size and frequency of fluctuations in output and lower inflation. This economic performance evidently led investors to become complacent about economic risks--for example, the risk that house prices could fall, and the small probability of a major downturn in the economy.
Moreover, persistently modest inflation gradually reduced inflation expectations, which brought down nominal interest rates. The tendency toward lower interest rates was accentuated earlier this decade by the sluggish recovery of investment from the recession of 2001, which combined with high saving propensities in the growing economies of Asia to lower real interest rates globally. The expectations of many investors about the returns they should be receiving apparently adjusted slowly to the new reality of low nominal interest rates, exacerbating the so-called search for yield that contributed to these investors venturing into riskier and more complex assets. And the benign economic environment encouraged intermediaries to lower credit standards and to try to pick up a little extra return by borrowing for the short term to fund long-term lending, leaving them increasingly leveraged and vulnerable.
But while monetary policy may have helped establish a stable macroeconomic environment in which investors overreached for yield, many other factors contributed to the financial crisis. Inadequate risk assessment and management by many financial market participants and the lagging adjustment of public oversight to the evolving structure of financial markets and rising risk levels were, in my view, the critical causes of the most severe financial crisis since the 1930s. And addressing these problems is the first priority and most effective safeguard against a repetition of the severe dislocations we experienced.
But does avoiding such a repetition also require monetary policy to include financial stability along with price stability and high employment in the objectives it tries to accomplish with its policy adjustments? When the monetary authorities judge that important asset prices or rates of credit expansion are deviating from sustainable long-run trends, should they adjust their policy setting to damp those price and credit movements--beyond whatever actions might be called for to preserve macroeconomic stability over the usual two- to three-year planning horizon for monetary policy?
To preview, I don't think we know enough to answer those questions with any confidence--to judge whether the benefits of such extra action would outweigh the costs. Given our state of knowledge, we should keep monetary policy focused on our mandates for maximum employment and stable prices. I can't rule out circumstances in which additional monetary policy actions specifically targeted at perceived asset price or credit imbalances and vulnerabilities would serve the pursuit of our mandate over the medium term. But given the bluntness of monetary policy as a tool for addressing developments that could lead to financial instability, the side effects of using policy for this purpose, and other difficulties, such circumstances are likely to be very rare. I continue to believe that the best approach is to promoting financial stability is to strengthen supervision and regulation.
Conducting monetary policy with additional emphasis on financial stability would imply that inflation and economic activity would generally be somewhat more variable over the medium run. Households and businesses might have greater difficulty planning saving and investment because they would be less certain of the medium-term trajectory of the economy and prices. Moreover, greater inflation variability might reduce the credibility of the central bank's pursuit of price stability, reducing its ability to lean against deviations in output as well. Another analytical point that should be taken into account when contemplating extra policy action is that the cost of using monetary policy to damp perceived asset-price or credit-driven imbalances would depend on the prevailing conditions in the economy. In general, the further the economy is from price stability or full employment, the greater the costs of further deviations. For example, raising interest rates and restraining output, employment, and inflation to lean against an incipient asset bubble will hurt more when inflation is already well below our objective or unemployment is high than when we are contemplating small deviations around long-run steady-state values.
From this perspective, the key analytical question is whether, by accepting somewhat larger deviations of employment and inflation from the levels otherwise attainable, we are able to sufficiently enhance financial stability in a way that significantly reduces the probability of a very severe financial crisis and the attendant serious economic costs. In my view, the theoretical and empirical models that would underpin this approach and its implementation are not adequately developed. Particularly in the absence of reliable models to guide us, policymakers attempting to conduct such a policy would confront serious challenges in identifying threatening imbalances and in calibrating and timing effective and appropriate policy actions.
Moreover, a cursory review of the evidence is not encouraging regarding the direct use of monetary policy to promote financial stability. For example, would a small adjustment in policy rates damp the optimism or complacency that was feeding the speculative excesses producing an asset or credit bubble? We don't know. But the impression from our experience in 1999 is that tightening policy had no effect on the intensity of dot-com speculation. And the housing bubble and the spread of lax mortgage lending practices built up for several years after we began raising interest rates in 2004. If policy adjustments intended to deflate asset prices or credit expansion only reduced inflation and output without affecting the bubble, they will have ultimately moved the economy further from its sustainable path when the bubble breaks. And if large adjustments are necessary to damp speculation, medium-term variations in the economy will be commensurately greater.
In addition, timing policy action accurately is critical to realizing greater benefits than costs in leaning against potential speculative excesses. Adjusting policy shortly before a bubble breaks, when its existence will be most evident, is not likely to be optimal, as it will do little to damp the excesses and will add to the downward pressure on the economy. Central banks would need to spot these threats very early, when they may not be generally recognized as divergences from fundamentals. Efforts to spot emerging imbalances early would raise the odds on identifying developments as threats to stability, when in fact they are natural and sustainable adaptations to changing relative prices, for example, from shifts in technology. Even with the access to the wide range of data, sophisticated models, and market analysis enjoyed by central banks, our ability to discern deviations of asset prices from fundamental values is very limited.
The difficulties I've just outlined lead me to a strong preference for using prudential regulation to deal with potential problems in credit and asset markets. Historically, bank supervision has been focused on individual institutions rather than on the system as a whole. Unfortunately, such microprudential regulation in practice was not as effective as it should have been, and an important challenge for policymakers will be to strengthen the supervision of individual institutions. However, the experience of this crisis also strongly suggests that policymakers need to pay close attention to developments across the financial system--that is, to engage in macroprudential supervision and regulation. Both microprudential and macroprudential oversight are essential to making the financial system more resilient to the inevitable cycles in asset prices, and less prone to large cycles. In this regard, many improvements are possible: better alignment of regulatory capital requirements with risk; making capital requirements less procyclical; identification of systemically important intermediaries and holding them to standards commensurate with their importance in the financial system; improved techniques for supervision of all intermediaries; spotting emerging increases in risk and adjusting regulation to address those problems; and addressing the elements in the financial markets that tend to amplify cycles, such as accounting and margining practices. The list of potential regulatory changes is long; without more experience, the efficacy of some of these adjustments is hard to assess; and putting them into practice will be difficult. But proceeding with careful analysis and implementation of these approaches would seem to have a better chance of yielding net benefits to our economy than using monetary policy to damp asset and credit cycles beyond what is required to keep our economy on an even keel.
Current circumstances well illustrate the challenges of using monetary policy as a tool to promote financial stability. The prices of riskier assets have risen rapidly in the past few months, with, for example, equity prices recovering a considerable portion of their earlier decline and risk spreads on corporate bonds dropping substantially. In addition, commodities prices have risen and the dollar has retraced a good portion of its run-up last fall and winter. These developments have led some observers to question whether another round of potentially destabilizing asset-price movements is in train. The movements in asset prices reflect several factors. One is a reversal of the extreme panicky conditions of late last year and earlier this year when extraordinary uncertainty about the economy and the soundness of many counterparties led to a widespread flight to liquidity and safety. As the economy stabilized and then began to improve, as financial markets became less volatile, and as additional information about the health of individual borrowers emerged, investors have become more willing to shift back into longer-term and riskier assets. The rise in commodities prices probably reflects the turnaround in economic prospects, especially with many resource-importing emerging market economies leading the rebound.
Another important factor has been the very low level of policy interest rates in the United States and in most other industrialized economies. Indeed, one of the purposes of these policies is to induce investors to shift into riskier and longer-term assets in order to lower the cost and increase the availability of capital to households and businesses. The more accommodative financial conditions, in turn, are intended to induce an increase in spending at a time when the level of output is expected to remain depressed for some time relative to the capacity of the economy to produce.
As I've already noted, our abilities to discern the "correct" values of assets is quite limited. At present, however, the prices of assets in U.S. financial markets do not appear to be clearly out of line with the outlook for the economy and business prospects as well as the level of risk-free interest rates. Most bond spreads and equity premiums are still appreciably higher than a few years ago and comparable to their levels in past recessions. Moreover, money and credit have been quite weak, suggesting that asset price movements have not been fueled by increased leverage that would leave financial intermediaries vulnerable to a reversal of recent gains. The improvement in the spreads and functioning of securities markets has not been accompanied by any loosening of very tight credit conditions for bank credit; indeed, banks have continued to tighten terms and standards in recent months, albeit at a slowing rate.
Still, my current assessments could be wrong--asset prices may in fact be in the process of rising excessively. However, consider the complications of using monetary policy to lean against a presumed bubble at this time. A decision to reduce monetary accommodation now would mean taking immediate steps to raise short-term interest rates or to reduce the support for private credit markets by selling longer-term securities that the Federal Reserve has acquired over the past year. Tightening financial conditions at a time when an economic recovery has just begun, when labor markets are continuing to weaken, when inflation is below its optimal level for the longer run, and when significant strains persist in the financial system would incur a considerable short-run cost in order to achieve possible long-run benefits whose extent is, at best, quite uncertain. Still, we will need to be alert to any tendencies for movements in prices for commodities and assets to result in a sustained increase in inflation and inflation expectations; unanchored inflation expectations would ultimately destabilize output and undermine our ability to foster higher levels of employment. And we need to continue to work with banks to improve their risk-management and credit-evaluation systems so that the banks are not vulnerable to unexpected movements in interest rates or asset prices in the future.
In sum, it seems to me that under most circumstances monetary policy is not the appropriate tool to use to address asset-price developments or growing vulnerabilities in financial markets. As I argued earlier, microprudential and macroprudential policies seem likely to me to be more effective and targeted at the problem than monetary policy adjustments, and in my view these tools should be the first that policymakers deploy. From my perspective, central bankers would need convincing evidence that such other tools would be inadequate and that significant asset-price misalignments were developing that would have serious economic costs before attempting to use monetary policy to address them.
Federal Reserve Tools to Provide Liquidity in Stress Situations
In addition to highlighting shortcomings in the U.S. regulatory regime, the financial crisis also raised questions about the adequacy of the Federal Reserve's tools for monetary policy implementation, particularly in periods of economic stress, and for crisis management more generally. In routine circumstances, the Federal Reserve implements monetary policy primarily through open market transactions, with the discount window playing an important but supporting role. Customarily, the bulk of the Federal Reserve's assets are U.S. Treasury securities, and the Federal Reserve uses short-term repurchase (repo) transactions, generally in relatively small amounts, to make the adjustments to the quantity of reserves in the banking system that are necessary to achieve the Federal Open Market Committee's target for the federal funds rate. Notably, the Federal Reserve's counterparties for these repo transactions are a small number (currently 18) of broker-dealers, known as primary dealers. Even though these dealers do not have reserve requirements, and indeed are not permitted to hold reserves at the Federal Reserve, the dissemination of reserves from the Federal Reserve's open market transactions conducted through the primary dealers, who pass the funds on through the highly liquid money markets, is largely sufficient in routine circumstances to implement monetary policy smoothly. Ordinarily, little credit is extended through the discount window; banks are able to obtain their funding and reserves in the open market and generally turn to the window only to cover very short-term liquidity shortfalls arising from operational glitches as opposed to more fundamental funding problems. This structure allows the Federal Reserve to implement policy quite efficiently in routine circumstances with minimal interference in private credit markets.
However, during the financial crisis, some aspects of the Federal Reserve's monetary policy implementation framework worked less well than in routine circumstances. A hallmark of the financial crisis was the impaired functioning--and in some cases, even a complete breakdown--of critical financial markets. Market participants became highly uncertain about several matters, such as the financial strength of their counterparties; the future value of assets, including any collateral they might be lending against; and how their own needs for capital and liquidity might evolve. They fled to the safest and most liquid assets and would not engage in the trading and arbitrage that usually transmit monetary policy impulses throughout the financial system. For example, from the early days of the crisis, trading in the interbank funding markets for terms longer than overnight dried up considerably. The greatly reduced access of banks to term funding likely was an important factor that contributed to a sharp pullback in their willingness to lend to households and firms.
The fact that primary dealers rather than commercial banks were the regular counterparties of the Federal Reserve in its open market operations, together with the fact that the Federal Reserve ordinarily extended only modest amounts of funding through repo agreements, meant that open market operations were not particularly useful during the crisis for directing funding to where it was most critically needed in the financial system. The Federal Reserve attempted to address the reduced availability of term funding to banks partly by lowering the cost of discount window credit and by increasing the maximum maturity of such credit from 1 day to 30 days and then later to 90 days. However, banks' willingness to use the discount window was greatly undermined by their concerns about the "stigma" that they would bear if their use of the window somehow was detected by counterparties or market analysts. To address banks' need for term funding, the Federal Reserve created the Term Auction Facility (TAF), under which it auctions large blocks of term funds to banks. The TAF combines aspects of open market operations and the discount window. The legal form of the TAF is the same as that of regular discount window loans. But by providing funds through an auction mechanism rather than through a standing facility, the TAF resembles open market operations rather than the standard discount window and, partly as a result, it appears to have largely avoided the stigma problem that limited the effectiveness of the discount window. Important questions for the Federal Reserve going forward are whether the benefits of the TAF warrant its maintenance on an ongoing basis or whether, now that the TAF has been developed, it can be brought off the shelf sufficiently quickly if warranted by circumstances.
The dysfunction in money markets was by no means limited to the United States. In globally integrated financial markets, many banks overseas had dollar obligations that they were financing with short-term dollar borrowing. When funding markets were disrupted, the intense bidding of these banks for dollar financing put upward pressure on interest rates in the United States. To relieve this pressure, the Federal Reserve entered into liquidity swaps with foreign central banks to enable those central banks to lend dollars to commercial banks in their jurisdictions. We also need to decide whether and in what form we should keep this type of liquidity provision readily available for possible future use.
Although the Federal Reserve routinely conducts open market operations through primary dealers, as noted above, it also does so through an auction mechanism; that mechanism implies both that no individual firm has any assurance of access to Federal Reserve liquidity and that the aggregate amount of funds provided through open market operations does not bear any direct relationship to dealers' funding needs. The Federal Reserve does not have legal authority to lend to firms other than depository institutions in routine circumstances. Moreover, given that dealers are not subject to the type of regulation and supervision applied to banks, routine lending to such firms could lead to moral hazard, with dealers' access to central bank liquidity increasing their incentives to take excessive risks. Thus, no regular standing facility for primary dealers is maintained, even though, like large commercial banks, primary dealers are highly leveraged, engage in significant maturity transformation, and are closely interconnected with the rest of the financial system. Dealers ordinarily fund themselves primarily in the triparty repo market, borrowing heavily to finance their securities positions. However, during the financial crisis, dealers' access to triparty funding declined sharply. Concerned with declining collateral values, repo counterparties increased required interest rates and ratcheted down the lendable values of collateral. As a consequence, dealers were forced to dump assets on the market in fire sales, further driving down collateral values. To break this vicious cycle and head off the collapse of the financial system that could have ensued, the Federal Reserve determined that it was necessary to lend to primary dealers on a fully collateralized basis. Later in the crisis, the Federal Reserve took similar actions to lend in support of money market mutual funds, which, like primary dealers, resemble banks in their economic function but lack regular access to the discount window.
Both of these actions were based on clear authority found in section 13(3) of the Federal Reserve Act, but they required, consistent with the statutory provision, that a five-member majority of the Federal Reserve Board determine that "unusual and exigent circumstances" prevailed--a determination that had not been made in decades. An important question for the future is what kind of authority we should have to lend to certain key groups of nonbank financial intermediaries or in support of markets. The Federal Reserve Act was designed when most credit flowed through banks, but over time securities markets have assumed a much more prominent role in the distribution of credit. Our ability to preserve financial stability may be enhanced by making sure the Federal Reserve has authority to lend against good collateral to other classes of sound, regulated financial institutions that are central to our financial markets--not on a routine basis, but in some circumstances when the Board of Governors finds that the absence of such lending would threaten market functioning and economic stability. The collateral would have to be of good quality and the institutions sound to minimize any credit risk the Federal Reserve might take. And the institutions would need to be tightly regulated and closely supervised to limit the moral hazard of permitting access to the discount window, even when such access is not routinely granted. I want to be quite clear that I am not referring to lending under section 13(3) to individual troubled institutions, like American International Group. That sort of lending is more appropriately done by the fiscal authorities and conducted only in association with the exercise of new authority to resolve systemically important financial institutions.
There are clear disadvantages to expanding the safety net to cover a wider range of firms as we did in the recent crisis. But we have seen that when market functioning is or could be substantially impaired, the effectiveness of our regular monetary policy tools is considerably diminished, with potentially severe economic consequences. In these circumstances, lending in support of markets and classes of important intermediaries becomes an essential extension of monetary policy to foster our legislated mandate of promoting maximum employment and stable prices.
After a serious setback to economic stability such as we have recently experienced, monetary policymakers must ask whether the strategies and tools they have been using need to be adapted to fulfill their responsibilities for price and economic stability in modern financial markets. As with most interesting policy questions, the answers aren't entirely clear. I've tried to outline some of the issues here in the hope of fostering further discussion of these critical questions.
1. I have addressed the first of these questions before--most recently a year ago in a speech given at the Cato Institute--but in light of the experience over the past several years, it seems appropriate to reexamine the issue. (See Donald L. Kohn (2008), "Monetary Policy and Asset Prices Revisited," speech delivered at the Cato Institute's 26th Annual Monetary Policy Conference, Washington, November 19.) Return to text