President, Federal Reserve Bank of Richmond
Shadow Open Market Committee Symposium
New York, N.Y.
The Federal Open Market Committee in January formally announced a numerical objective for inflation, a step which has long been argued to be essential to anchoring longer-term expectations about the conduct of monetary policy.1 So it might seem a bit surprising, as this year draws to a close, to find a member of the Committee speaking at an event whose title is “The Fed’s Monetary Policy Adrift.” But on further reflection, I don’t think it should be surprising at all. Both the FOMC’s articulation of an inflation target and the sense that policy is adrift are related, I believe, to the extraordinary circumstances and resulting policy actions of the last few years. In my remarks this morning, I will discuss two dimensions of Federal Reserve policy that came in the wake of the financial crisis and Great Recession: first, the effort to provide stimulus and policy guidance at the zero bound; and second, the expansion of the scope of Fed policy beyond monetary policy to a broader engagement in credit policy. Before I begin, however, I need to recite a disclaimer that should be quite familiar to members of the Shadow Open Market Committee — my remarks reflect my own views and not necessarily those of any other members of the FOMC.2
Let me begin by noting that when the FOMC announced an explicit numerical objective for inflation this year, we had experienced an extended period of relative monetary stability. Specifically, since December of 1993, inflation (as measured by the price index for personal consumption expenditures) has averaged very close to 2 percent per year — a very good performance when compared against previous decades. To be sure, that performance has not been perfect; inflation averaged over 3 percent for the five years from mid-2003 to mid-2008, a subpar outcome for which we at the Fed should accept responsibility. Nevertheless, despite such swings, inflation has generally tended to return to around 2 percent, and this appears to have enhanced public confidence in the Fed’s willingness and ability to keep inflation low and stable. Critical to that process was the Fed’s demonstrated determination to act preemptively against inflationary pressures over the last three decades, particularly in 1994, an episode to which I will return.3
This period of relative success on our implicit inflation objective helped make the announcement of an explicit numerical inflation objective in the January statement possible. Actions speak louder than words, after all, and without having seen the Fed take action to preempt inflation, mere words might have done little, by themselves, to bolster credibility. The clear statement of the FOMC’s monetary policy objective was still important, though, to help dispel lingering doubts about the Committee’s intentions and to provide a clear benchmark for accountability.
Implicit in the Fed’s credibility is some measure of public understanding of how the Fed will typically respond to changes in economic circumstances. That understanding no doubt depends heavily on the Fed’s observed responses over the last 20 years or more, but the postwar historical record does not include any extended periods in which the Fed’s target interest rate was effectively at the zero lower bound. As a consequence, uncertainty about future Fed policy actions is bound to be greater now than in a more typical interest rate environment. This provides a compelling reason, in my view, for the FOMC to attempt to provide greater guidance about future policy conduct. The recent appearance of drift in policy may be attributable to the Committee’s search for more effective ways to communicate about future policy in a relatively unique setting.
Communicating Policy Actions
The most recent innovation in communication has been the use of a calendar date rather than a qualitative phrase to characterize the time period over which the Committee anticipates interest rates will be exceptionally low. Specifically, the Committee said in August 2011 that it “currently anticipates that economic conditions … are likely to warrant exceptionally low levels for the federal funds rate through at least mid-2013.” At subsequent meetings, “mid-2013” was changed to “late 2014,” and then “mid-2015.” The minutes of the August 2011 meeting said that the Committee viewed the change in language as “a shift toward more accommodative policy,” implying a desire to shift the yield curve downward. The language of the statement, however, was phrased as simply a forecast of future Committee behavior.
The current formulation of the forward guidance raises the question: How can a change in the forecast of future policy settings also be a shift to more accommodative policy? Indeed, market participants have seemed confused about the extent to which the forward guidance represents a commitment. Committee members have emphasized in public statements that the time frame should be viewed as contingent on incoming data. But describing the forward guidance language as “a shift to more accommodative policy” seems to imply that the Committee intends to choose policy settings in the future in a way they would not otherwise see fit at that time. The lack of clarity about forward guidance has contributed to a problem highlighted by Michael Woodford — namely, that observers may misinterpret a change in the forward guidance date as a pessimistic shift in the Committee’s assessment of the drivers of economic growth rather than as a clarification of its reaction function.4
Experimentation with more explicit forward guidance has been motivated by suggestions from some economists that the Fed can make current policy more stimulative by assuring the public that it will keep its interest rate target at the zero bound longer than it would if it were following its normal pattern of behavior.5 It’s not clear whether this mechanism can work, however, without raising expected inflation over some horizon.6 Adopting such a strategy without compromising longer-term credibility may be feasible in model environments, where absolute credibility can easily be assumed. In practice, however, a central bank’s credibility is often contingent and incomplete. My reading of recent history is that the Fed’s credibility is not so unassailable that inflation expectations can be dialed up for a time and then easily dialed back to price stability. At the very least, the precedent set by an opportunistic attempt to raise inflation temporarily is likely to cloud our credibility for decades to come.
It should be unobjectionable, however, to provide forward guidance that reduces unnecessary uncertainty about the central bank’s reaction function and thereby helps people make better predictions about future monetary policy. For example, the Committee could provide some sense of the economic conditions under which it’s likely to begin raising rates and reducing the size of its balance sheet. But it’s important to avoid spurious precision. Some of my colleagues have suggested that the Committee provide specific numerical thresholds to help characterize future policy. For example, they suggest that the Committee state that interest rates will be exceptionally low at least until the unemployment rate falls below some specific number, as long as inflation is projected to be close to the Committee’s 2 percent objective, and inflation expectations remain stable.
This approach would place great weight on a single indicator of labor market conditions, one that can easily lead you astray. This risk seems particularly germane now, given the difficulty of disentangling the trend and cyclical components of labor force participation. The January statement in which the Committee announced its 2 percent inflation objective also explained that “[the] Committee considers a wide range of indicators” in assessing labor market conditions. Crisp numerical thresholds may work well in the classroom models used to illustrate policy principles, but one or two economic statistics do not always capture the rich array of policy-relevant information about the state of the economy.
Proponents of numerical thresholds sometimes reply to this criticism by citing the inflation “safety valve” clause that says: “as long as inflation is projected to be close to the Committee’s 2 percent objective and inflation expectations remain stable.” They argue that if a poorly specified unemployment threshold caused us to hold interest rates low for too long, inflation expectations would rise and a rate increase would be indicated. This strikes me as an inadequate defense because it essentially requires that we lose a measure of credibility before it can be invoked. Our policy should strive to maintain the stability of inflation expectations. At times, this requires a preemptive tightening of monetary policy, before inflation expectations have deteriorated or inflation has surged. In February 1994, for example, the FOMC began tightening monetary policy, despite well-behaved inflation and an unemployment rate over 7 percent.7
In addition to forward guidance, the FOMC’s other main initiative at the zero bound has been asset purchases, with the most recent installment being the purchases of agency mortgage-backed securities that began after the September meeting. Back in 2009, I fully supported the first wave of purchases of U.S. Treasury securities because it was clear that heightened uncertainty had increased the demand for safe liquid assets, such as reserve account balances. Furnishing an elastic supply of central bank liabilities in that instance helped prevent deflation.
Since then, the FOMC’s asset purchase programs have increasingly focused on altering the composition of the Fed’s asset holdings in order to affect the net public supply of assets with particular characteristics and thereby affect their relative prices. The idea is that some type of market segmentation breaks the standard arbitrage relationships that would generally keep various asset prices aligned. Thus, purchases of longer-term Treasury securities are thought to reduce the slope of the yield curve, and purchases of mortgage-backed securities, or MBS, are thought to reduce their spreads over comparable Treasuries.
There is ample room for skepticism about the effect of the Fed’s asset purchases on asset returns. A broad array of investors seems to be capable of operating across multiple asset markets, and the markets in which the Fed has been active tend to be relatively broad and deep. Moreover, the empirical evidence on the effects of Fed asset purchases, which is based on yield movements around the announcements of asset purchases, is ambiguous, given the difficulty of parsing policy signals from pure supply effects.
When the Fed expands reserves by buying private assets, it extends public sector credit to private borrowers. To the extent that purchases of private claims have any effect, they do so by distorting the relative cost of credit among different borrowers. Such differential effects are unlikely to be beneficial, on net, unless borrowers in the favored sector would otherwise face artificially high rates. I think it’s difficult to make this case for agency MBS, a sector that historically has benefited from heavy subsidies, which arguably contributed to dangerously high homeowner leverage. So I do not see the rationale for reducing the interest rates paid by conforming home mortgage borrowers relative to those paid by, say, small-business borrowers. Moreover, purchasing agency MBS encourages the continuation of a housing finance model based heavily on government-sponsored enterprises, at a time when the housing sector would be better served by a new model that relies less on government credit subsidies.
Credit Market Intervention
The pattern of Federal Reserve credit market intervention has evolved over time. The most recent articulation of an explicit credit policy, as such, is the Joint Statement of the Department of Treasury and the Federal Reserve of March 23, 2009, which stated that “[government] decisions to influence the allocation of credit are the province of the fiscal authorities.” This expresses well the core idea of a “credit accord” that Professor Marvin Goodfriend first advocated many years ago while at the Richmond Fed, and that I and others have endorsed.8 The apparent contradiction between the March 2009 Treasury-Fed statement and the FOMC’s recent interventions to steer credit to the housing market also may be contributing to the perception that Federal Reserve credit policy is adrift.
Uncertainty regarding Fed credit policy has precedents, whether it involves direct lending or purchases of private sector assets. For several decades prior to the recent crisis, policy regarding lending to financially stressed firms was often characterized by “constructive ambiguity.”9 Financial firms and their creditors were encouraged to believe they would not be rescued in the event of distress, while officials preserved their ability to intervene should a crisis actually arise. Constructive ambiguity sought to obtain the ex-ante incentive benefits of commitment without giving up the discretion to act freely ex post. But taking their cues from central bank actions rather than its words, market participants’ expanded their reliance on implied commitments of central bank liquidity support. This created excruciating dilemmas in times of stress, as was vividly illustrated during 2008: Disappoint short-term creditors and massive investor realignment destabilizes markets; rescue short-term creditors and the additional precedent reinforces expectations of future rescues and further intensifies moral hazard. Constructive ambiguity became increasingly hopeless in the face of accumulating instances of intervention, and the toxicity of credit policy opacity is now quite clear. Financial stability is likely to remain elusive without constraints on ad hoc rescues of firms facing financial stress.
Unconstrained credit policy thus poses a thorny problem for the modern central bank, as Professor Goodfriend has forcefully argued.10 Independent management of their balance sheet is essential to a central bank’s ability to conduct monetary policy in a way that is relatively free of the short-term pressures associated with electoral politics. But an immediate consequence of a central bank’s independence is the capacity to use its balance sheet to direct the flow of credit toward particular market segments, circumventing the constitutional checks and balances that would otherwise apply to such fiscal initiatives. Marvin Goodfriend and my predecessor, Al Broaddus, writing in 1994, warned that central bank forays into fiscal policy would be perceived as redistributional and would risk entanglement in partisan politics. The political backlash following the Federal Reserve’s 2008 actions, I believe, validates their concerns.
The reactive evolution of Fed credit policy over recent decades parallels the way monetary policy drifted into instability during the 1960s and 1970s.11 The process of solving the inherent time consistency problem and restoring monetary stability was long and costly. Legislative and constitutional solutions were proposed, but success depended on the Fed itself making price stability a priority and culminated with the FOMC’s adoption of the self-imposed constraint of a numerical inflation target.
Limiting Central Bank Lending
The process of establishing credible limits on central bank lending could be even more difficult than the pursuit of price stability. Whether self-imposed lending constraints could be effective remains to be seen. One approach would be for the Fed to operationalize the principles articulated in the March 2009 Joint Statement of the Treasury and the Fed. The alternative to self-imposed restraint is legislative action. The Dodd-Frank Act pared back the Fed’s ability to lend beyond the banking system by limiting the Fed’s so-called “13(3)” powers to lend to nondepository firms in “unusual and exigent circumstances.” These restrictions are modest, however. One could imagine legislation that limits the Fed to a narrowly defined set of ordinary lending activities — very short-term lending to sound, solvent banks, against good collateral, at rates above interbank market rates. If the Federal Reserve cannot limit credit policy of its own accord, legislation may be the best option. And the restraint of credit policy would not be complete unless limits on reserve bank lending are complemented by limits on the Fed’s ability to buy private sector assets.
Expansive central bank lending has its supporters, and some are likely to argue that such restraints would inhibit performance of the “lender of last resort function” that is traditionally thought to be an essential central bank role. Professor Goodfriend is persuasive, I believe, in demonstrating that this is a misreading of the historical record. A century ago, central bank lending was thought of primarily as a means of rapidly increasing the supply of paper bank notes when the demand for those notes surged, either in connection with seasonal agricultural cycles or in connection with financial panics in which depositors sought to convert their deposits into currency. This is consistent with the purpose of the Federal Reserve Act, which, according to the preamble, is “to furnish an elastic currency.”12
I will conclude by noting a theme that runs through both the monetary and credit sections of my remarks: humility. Central banks are at times asked to do too much — and at times they ask themselves to do too much. Given their fiscal independence and their historically expansive authority, one can see why people look to central banks as public-sector benefactors. But central bank success arguably has been associated more with restraint than ambition. The Fed tamed inflation when it backed away from overly ambitious notions of the role monetary policy could play in labor market outcomes. I believe that future financial stability will depend similarly on central bank modesty about its ability to redirect credit flows constructively. The independence and effectiveness of the modern central bank will require limiting aspirations.
Board of Governors of the Federal Reserve System, “Longer-Run Goals and Policy Strategy.” News and Events, Monetary Policy Press Release, January 25, 2012.
I am grateful to John Weinberg for assistance in preparing these remarks.
Marvin Goodfriend, “Monetary Policy Comes of Age: A Twentieth Century Odyssey,” Federal Reserve Bank of Richmond Economic Quarterly, Winter 1997, vol. 83, no. 1, pp. 1-22.
Michael Woodford, “Methods of Policy Accommodation at the Interest Rate Lower Bound.” Delivered at the Federal Reserve Bank of Kansas City’s 2012 Jackson Hole Symposium, Jackson Hole, WY, August 31, 2012.
Woodford, 2012. Also, Gauti Eggertsson and Michael Woodford, “Optimal Monetary and Fiscal Policy in a Liquidity Trap.” Presented at the National Bureau of Economic Research International Seminar on Macroeconomics 2004, pp. 75-144, National Bureau of Economics, 2006.
Ivan Werning, “Managing a Liquidity Trap: Monetary and Fiscal Policy,” National Bureau of Economics Research Working Paper 17344, August 2011. Werning writes that forward commitment can raise current growth even if prices are so sticky that there is no effect on inflation. If prices are not perfectly fixed, however, commitment to greater future stimulus raises inflation.
Board of Governors of the Federal Reserve System, Federal Open Market Committee, Greenbook, February 1994; and Robert Hetzel, “The Monetary Policy of the United States: A History” (New York, NY, Cambridge University Press, 2008, chapter 15).
Marvin Goodfriend, “Why We Need an ‘Accord’ for Federal Reserve Credit Policy: A Note,” Journal of Money, Credit and Banking, August 1994, v. 26, no. 3, pp. 572-80. This was also printed in the Federal Reserve Bank of Richmond Economic Quarterly special issue on the 50th anniversary of the Treasury-Fed Accord, 2001. Also Jeffrey Lacker, “Government Lending and Monetary Policy.” Speech at the National Association for Business Economics 2009 Washington Economic Policy Conference, Alexandria, VA, March 2, 2009; and Charles Plosser, “Ensuring Sound Monetary Policy in the Aftermath of the Crisis.” Speech at the U.S. Monetary Policy Forum, New York, February 27, 2009.
Marvin Goodfriend and Jeffrey Lacker, “Limited Commitment and Central Bank Lending,” Federal Reserve Bank of Richmond Economic Quarterly, 1999, v. 85, n. 4. Also Lacker, “Reflections on Economics, Policy, and the Financial Crisis.” Speech to the Kentucky Economics Association, Lexington, KY, September 24, 2010.
Marvin Goodfriend, “The Ellusive Promise of Independent Central Banking,” Institute for Monetary and Economic Studies, Bank of Japan, discussion paper 12-E-09, September 2012.
Goodfriend and Lacker (1999).
Elmus Wicker, “The Great Debate on Banking Reform: Nelson Aldrich and the Origins of the Fed” (Columbus, OH, Ohio State University Press, 2005).