JOURNAL OF MONETARY ECONOMICS Aims and Scope: The Journal of Monetary Economics publishes important research contributions to a wide range of modern macroeconomic topics including work along empirical, methodological and theoretical lines. In recent years, these topics have been: asset pricing; banking, credit and financial markets; behavioral macroeconomics; business cycle analysis; consumption, labor supply, and saving; dynamic equilibria (theory and computational methods); economic growth and development; expectation formation, information and aggregate economic activity; fiscal shocks and fiscal policies; expectation formation; forecasting, macroeconometrics, and time series analysis; information and aggregate economic activity; international trade, exchange rates, and open economy macroeconomics; labor markets; macroeconomic data and history; monetary policy; monetary theory; money demand and money supply behavior; optimal contracting and economic activity; productivity measurement and theory; pricing in product markets and labor markets; and real investment (inventories, fixed, human capital). The Journal has eight regular issues per year, with the Carnegie-Rochester Conference Series on Public Policy as the January and July issues. Founding Editors: KARL BRUNNER and CHARLES I. PLOSSER Editor: ROBERT G. KING, Department of Economics, Boston University, 270 Bay State Road, Boston, MA 02215, USA Co-Editors: Urban Jermann, University of Pennsylvania; Ricardo Reis, Columbia University Senior Associate Editors: MARIANNE BAXTER, Boston University; JANICE EBERLY, Northwestern University; MARTIN EICHENBAUM. Northwestern University; SERGIO REBELO, Northwestern University; STEPHEN WILLIAMSON, University of Washington Associate Editors: KLAUS ADAM, University of Mannheim; NICHOLAS BLOOM, Stanford University; YONGSUNG CHANG, University of Rochester; MARIO CRUCINI, Vanderbilt University; HUBERTO ENNIS, Federal Reserve Bank of Richmond; KRISTOPHER GERARDI, Federal Reserve Bank of Atlanta; MIKHAIL GOLOSOV, Yale University; FRANCOIS GOURIO, Boston University; JONATHAN HEATHCOTE, Federal Reserve Bank of Minneapolis; RICARDO LAGOS, New York University; EDWARD NELSON, Federal Reserve Board; GIORGIO PRIMICERI, Northwestern University; ESTEBAN ROSSI-HANSBERG, Princeton University; PIERRE-DANIEL SARTE, Federal Reserve Bank of Richmond; FRANK SCHORFHEIDE, University of Pennsylvania; CHRISTOPHER SLEET, Carnegie-Mellon University; JULIA THOMAS, Ohio State University; ANTONELLA TRIGARI, Università Bocconi; LAURA VELDKAMP, New York University; ADRIEN VERDELHAN, Massachusetts Institute of Technology; ALEXANDER WOLMAN, Federal Reserve Bank of Richmond CRC Editors: THOMAS F. COOLEY, New York University; MARVIN GOODFRIEND, Carnegie Mellon University CRC Advisory Board: ANDREW ABEL, University of Pennsylvania; MARK AGUIAR, University of Rochester; MARK BILS, University of Rochester; YONGSUNG CHANG, University of Rochester; HAROLD COLE, University of Pennsylvania; JANICE EBERLY, Northwestern University; BURTON HOLLIFIELD, Carnegie Mellon University; BENNETT T. McCALLUM, Carnegie Mellon University; THOMAS PHILIPPON, New York University; CHARLES I. PLOSSER, Federal Reserve Bank of Philadelphia; CHRISTOPHER SLEET, Carnegie Mellon University; GIANLUCA VIOLANTE, New York University; TONI WHITED, University of Rochester; STANLEY E. ZIN, New York University Submission fee: There is a submission fee of US$250 for all unsolicited manuscripts submitted for publication. There is a reduced fee for full-time students (US$150). To encourage quicker response referees will be paid a nominal fee and the submission fee will be used to cover these refereeing expenses. There are no page charges. Cheques should be made payable to the Journal of Monetary Economics. When a paper is accepted the fee will be reimbursed. Publication information: Journal of Monetary Economics (ISSN 0304-3932). For 2011, volume 58 (8issues) is scheduled for publication by Elsevier (Radarweg 29, 1043 NX Amsterdam, the Netherlands). Further information on this journal is available from the Publisher or from the Elsevier Customer Service Department nearest you or from this journal’s website (http://www.elsevier.com/locate/jme). Information on other Elsevier products is available through Elsevier’s website (http://www.elsevier.com). Periodicals Postage Paid at Rahway, NJ, and at additional mailing offices. USA mailing notice: Journal of Monetary Economics (ISSN 0304-3932) is published 8 times per year by Elsevier (Radarweg 29, 1043 NX Amsterdam, The Netherlands). Periodical postage rate paid at Rahway NJ and additional mailing offices. USA Postmaster: Send change of address to Journal of Monetary Economics, Elsevier Customer Service Department, 3251 Riverport Lane, Maryland Heights, MO 63043, USA. Airfreight and mailing in USA by Mercury International Limited, 365 Blair Road, Avenel, NJ 07001.
The paper used in this publication meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper).
Printed by Henry Ling Ltd., The Dorset Press, Dorchester, UK
Journal of Monetary Economics 58 (2011) 1–12
Contents lists available at ScienceDirect
Journal of Monetary Economics journal homepage: www.elsevier.com/locate/jme
Central banking in the credit turmoil: An assessment of Federal Reserve practice$ Marvin Goodfriend n Carnegie Mellon University and NBER, Tepper School 257, Pittsburgh, PA 15213, USA
a r t i c l e in fo
abstract
Article history: Received 10 May 2010 Received in revised form 14 September 2010 Accepted 16 September 2010 Available online 7 October 2010
Central banking is understood in terms of the fiscal features of monetary, credit, and interest on reserves policies. Monetary policy expanding reserves by buying Treasuries transfers all revenue from money creation directly to the fiscal authorities. Credit policy selling Treasuries to fund loans or acquire non Treasury securities is debt financed fiscal policy. Interest on reserves frees monetary policy to fund credit policy independently of interest rate policy. An ambiguous boundary of responsibilities between the Fed and the fiscal authorities contributed to economic collapse in fall 2008. ‘‘Accord’’ principles are proposed to clarify Fed credit policy powers and secure its independence on monetary and interest rate policy. The Fed needs more surplus capital from the fiscal authorities to be fully flexible against both inflation and deflation at the zero interest bound. & 2010 Elsevier B.V. All rights reserved.
1. Introduction The credit market turmoil that began in August 2007 and precipitated the Great Recession challenged central banks around the world as never before. Central banks increased aggregate bank reserves enormously, and lowered targeted short term interest rates to zero in many countries. For instance, the Federal Reserve grew bank reserve balances from around 10 billion dollars in early September 2008 to over 1 trillion dollars as it drove the federal funds rate nearly to zero. Central bank lending expanded to facilitate private credit flows. For instance, Federal Reserve loans to depository institutions stood at over 400 billion dollars at the end of April 2009. Previously, the most expansive, prolonged Fed lending was a loan of roughly 5 billion dollars to Continental Illinois Bank from May 1984 until February 1985.1 The Fed extended its credit reach well beyond depository institutions. By April 2009, the Fed had purchased around 350 billion dollars of mortgage backed securities guaranteed by Fannie Mae, Freddie Mac, and Ginnie Mae; and the Fed had extended around 200 billion dollars of loans to a special purpose vehicle created to purchase commercial paper.2
$ Presented at the 75th Anniversary Carnegie Rochester Conference on Public Policy, April 16–17, 2010. The paper benefited from earlier conference presentations at the Bank of Japan, Princeton University, De Nederlandsche Bank, and the Federal Reserve Bank of Atlanta, as well as seminars at the Bank of England, the Board of Governors of the Federal Reserve System, the Federal Reserve Banks of Cleveland, Kansas City, New York, and Richmond, the London School of Economics, and Saint Vincent College. Comments by the conference discussants D. Altig, V.V. Chari, L. Ohanian, F. Smets, and H. Ugai were particularly valuable. n Corresponding author. Tel.: + 1 412 268 8459. E-mail address:
[email protected] 1 For a brief period following 9/11 Fed lending to banks rose above 30 billion dollars. Fed lending discussed throughout the text refers to overnight loans. 2 Federal Reserve Statistical Release H.4.1 ‘‘Factors Affecting Reserve Balances.’’
0304-3932/$ - see front matter & 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jmoneco.2010.09.008
2
M. Goodfriend / Journal of Monetary Economics 58 (2011) 1–12
Still farther afield, the Fed extended credit to three limited liability companies in conjunction with efforts to stabilize institutions that it deemed to be critically important. In mid March 2008 the Fed agreed to extend roughly 29 billion dollars to Maiden Lane I so that it could acquire a variety of mortgage obligations, derivatives, and hedging products to facilitate the acquisition of Bear Stearns by JP Morgan Chase. Maiden Lane II and III were both created to restructure the Fed’s lending to AIG in the aftermath of its financial support for AIG in September 2008. Together, the Fed lent Maiden Lane II and III roughly 50 billion dollars to purchase, respectively, residential mortgage backed securities from AIG, and multi sector collateralized debt obligations on which AIG wrote credit default swap contracts.3 All together, the Fed grew its balance sheet from around 900 billion dollars in mid 2007 to over 2 trillion dollars as of April 2009. The Fed did so while reducing its purchases of US Treasury securities from around 800 to 550 billion dollars. The Fed funded its enormous increase in lending with around 250 billion dollars from the sale of Treasury securities, around 300 billion dollars of additional deposits provided by the Treasury, and the creation of around 800 billion dollars of bank reserves for a grand total of around 1.3 trillion dollars of Fed lending as of April 2009. Since then, the Fed mainly shifted the composition of its assets shrinking its lending to depositories and through various special facilities, rebuilding its holdings of Treasuries to around 800 billion dollars, increasing its holdings of mortgage backed securities to around 1 trillion dollars, and acquiring about 170 billion dollars of federal agency debt securities.4 The Fed and other central banks around the world have undergone a ‘‘stress test’’ that is still very much in progress. Yet enough time has passed to take stock, not so much to evaluate the timing, magnitude, and effectiveness of particular actions, but to observe how central banks put their various powers to work, and to use the observations to rethink central banking more generally. This essay presents a framework for thinking about central banking in light of these extraordinary developments. The reconsideration begins by classifying core central banking initiatives as monetary policy, credit policy, or interest on reserves policy. Briefly, monetary policy refers to open market operations that expand or contract high powered money (bank reserves and currency) by buying or selling Treasury securities. Credit policy shifts the composition of central bank assets, holding their total fixed. Interest on reserves policy involves adjusting interest paid on bank reserves. The three fold classification did not matter much for the Fed in the past. Until the recent credit turmoil Fed credit policy played a relatively minor role. The Fed could not pay interest on reserves. And monetary policy was utilized to target the federal funds rate. However, the classification is essential to understand the extraordinary central banking initiatives in the current context. The heart of the paper is the idea that monetary, credit, and interest on reserves initiatives all involve fiscal policy in important but different ways, and that it is essential to understand each initiative in terms of its fiscal features. Fiscal policy involves the use of public funds acquired with current taxes or by borrowing against future taxes. For the purpose of this paper, fiscal policy should be understood to include the lending of public funds to particular borrowers financed by selling Treasury securities against future taxes. From a fiscal policy perspective, monetary policy saves the government interest that it would pay otherwise on outstanding Treasuries. Mechanically, the interest saving is achieved because the central bank returns to the Treasury after expenses all the interest it receives on the Treasuries it acquires in the conduct of monetary policy. Thus, expansionary monetary policy provides the fiscal authorities with a flow of additional revenue. In particular, all the revenue from monetary policy is transferred via the acquisition of Treasuries to the fiscal authorities to allocate as they see fit. Credit policy is not monetary policy because it does not alter the stock of bank reserves or currency outstanding. Credit policy involves lending to particular borrowers or acquiring non Treasury securities with proceeds from the sale of Treasuries. The fiscal authorities then receive interest on the credit assets instead of interest on the Treasuries held by the central bank. In effect, credit policy commits future taxes to back loans or non Treasury security purchases via the sale of Treasuries. Thus, credit policy involves the fiscal allocation of public funds in a way that monetary policy does not. Interest on bank reserves is not monetary or credit policy since it involves neither the size nor the composition of the central bank balance sheet. Interest on reserves uses public funds to pay interest on bank reserve balances at the central bank; therefore interest on reserves also involves the allocation of public funds in a way that monetary policy does not. These and other fiscal features of monetary, credit, and interest on reserves policies are employed below to identify and evaluate the role that each type of initiative plays in central bank stabilization policy. The paper also considers questions of central bank independence. Flexibility and decisiveness are essential for effective central banking. Independence is essential to enable a central bank to react promptly to macroeconomic or financial shocks without the approval of the Treasury or the legislature. Central bank initiatives must be regarded as legitimate by the fiscal authorities and the public. The problem is to identify the limits of independence on monetary, credit, and interest on reserves policies to preserve a workable, sustainable division of responsibilities between the central bank and the fiscal authorities. Monetary policy can be conducted independently by a central bank because the objectives of monetary policy price stability and full employment are reasonably clear and coherent, and confining purchases to Treasuries transfers all the revenue from money creation directly to the fiscal authorities.
3 4
Board of Governors of the Federal Reserve System (2009b), February, appendix. Board of Governors of the Federal Reserve System (2009a), starting in June 2009.
M. Goodfriend / Journal of Monetary Economics 58 (2011) 1–12
3
Neither condition is satisfied for central bank credit policy. Objectives that guide and circumscribe central bank credit policy have not been clear, and credit policy inherently involves the fiscal allocation of public funds. Prior to the passage of the Dodd Frank financial reform legislation of 2010, Alan Greenspan wrote that in 1991 ‘‘at the urging of the Federal Reserve Board of Governors, Section 13 3 of the Federal Reserve Act was considered, and amended, by Congress. The section grant[ed] virtually unlimited authority to the Board to lend in ‘‘unusual and exigent circumstances’’ ’’.5 Given the inherent fiscal nature of credit policy, it is not surprising that expansive Fed credit policy in the recent turmoil created conflict with the Congress. In light of that experience, the Dodd Frank law limits the Fed’s power to lend, requiring Fed lending extended beyond depository institutions to be approved by the Treasury Secretary and to be part of a broad program not directed to any particular borrower. This paper develops a set of principles as the basis for a Treasury Fed ‘‘accord’’ to clarify and limit the Fed’s credit policy powers and preserve its independence on monetary and interest on reserves policy. Sections 4.2 and 4.3 below suggest that ambiguity in the responsibility for the provision of fiscal support for the financial system contributed to panic and economic collapse in fall 2008. Among other things, the ‘‘Accord’’ principles are motivated by the realization that an independent central bank cannot be relied upon to deliver or decide upon the delivery of fiscal support for the financial system. The essay proceeds as follows. Section 2 classifies central banking initiatives into monetary policy, credit policy, and interest on reserves policy. Section 3 explains how monetary, credit, and interest on reserves policies work in terms of their fiscal features. Section 4 assesses five actual Federal Reserve initiatives in the credit turmoil the Term Auction Facility, Fed lending to facilitate the acquisition of Bear Stearns by JP Morgan Chase, Fed support for AIG, emergency authority to pay interest on reserves, and the joint statement by the Treasury and the Fed on the role of the Fed in preserving financial and monetary stability. Section 5 develops and presents the ‘‘accord’’ principles for clarifying the boundary of central bank credit policy. Section 6 considers monetary and fiscal proposals to strengthen policy flexibility at the zero interest bound and in the exit strategy use by the central bank of non monetary managed liabilities, and enlarged surplus capital on the central bank balance sheet. Section 7 is a brief conclusion. 2. Monetary policy, credit policy, and interest on reserves policy Monetary policy consists of open market operations that expand or contract high powered money (bank reserves plus currency) by buying or selling Treasury securities. Until the recent credit turmoil, the Fed satisfied virtually all of its asset acquisition needs in support of monetary policy by purchasing Treasury securities, an acquisition policy known as ‘‘Treasuries only’’.6 This was done to avoid carrying credit risk on the Fed’s balance sheet. Pure monetary policy works by varying the aggregate quantity of bank reserves to influence the spread between the federal funds rate and interest paid on reserves. For example, an open market purchase of Treasury securities that adds reserves to the banking system lowers the federal funds rate relative to whatever rate the Fed pays on reserves which until October 2008 was always zero. The Fed utilized monetary policy exclusively in the past to manage the federal funds rate in the pursuit of interest rate policy directed by the Federal Open Market Committee. At the start of the credit turmoil in the summer of 2007, the Fed had accumulated on its balance sheet roughly 850 billion dollars of Treasury securities obtained in the course of supplying the economy with currency and bank reserves. Credit policy involves changing the composition of the central bank’s asset portfolio between Treasury securities, on one hand, and credit to the private sector or to non Treasury government entities on the other hand, holding high powered money fixed. For instance, the hundreds of billions of dollars of TAF discount window credit auctioned by the Fed to depositories from the fall of 2007 through the summer of 2008 was credit policy financed by selling Treasuries from the Fed’s portfolio. A combination credit and monetary policy initiative could involve the funding of central bank lending to depositories or the purchase of non Treasury securities with newly created bank reserves. The trillion dollars of bank reserves that currently finances a like volume of mortgage backed securities and federal agency debt on the Fed balance sheet reflects a combination credit and monetary policy. Importantly, even though most MBS and agency securities held by the Fed are liabilities of Fannie Mae and Freddie Mac, and the two agencies were placed into government conservatorship in September 2008, these securities are not the equivalent of US Treasuries. They do carry a very strong guarantee by the US Treasury. But the Treasury’s ‘‘credit enhancement’’ is not the same as the legally binding ‘‘full faith and credit’’ obligation of the US government that backs Treasury securities.7 Only Congress can confer ‘‘full faith and credit’’ backing. Ginnie Mae MBS have such backing, but Fannie and Freddie liabilities do not. Interest on reserves policy consists of varying interest that a central bank pays on bank reserves, holding monetary policy and credit policy fixed. As explained in Section 3 below, the spread (possibly zero) between the federal funds rate and interest paid on reserves is pinned down by the aggregate quantity of bank reserves provided by the central bank. Hence, holding monetary policy fixed, an adjustment in interest paid on reserves tends to be passed directly to the federal 5 6 7
Greenspan (2010), p. 17). Most Treasuries have been purchased outright with a fraction held under repurchase agreements for liquidity purposes. Goldfarb (2010).
4
M. Goodfriend / Journal of Monetary Economics 58 (2011) 1–12
funds rate. Thus, a central bank can pursue interest rate policy without employing monetary policy by varying interest on reserves alone. Soon after the Fed began to pay interest on reserves in October 2008 it cut interest rates nearly to zero and interest on reserves has not mattered much since. As discussed in Section 4.4 below, however, interest on reserves will help the Fed exit the zero interest bound, if need be, without first shrinking its balance sheet. Even if an abundance of aggregate bank reserves continues to put downward pressure on the federal funds rate, depository institutions will not lend in the interbank market at interest below the rate they can earn on reserve balances held at the Fed. Hence, the Fed will be able to elevate market interest rates over time by paying ever higher interest on reserves, regardless of the size of its balance sheet.
3. Fiscal aspects of monetary, credit, and interest on reserves policies Monetary policy involves fiscal policy in two ways. First, monetary policy governs the tax rate on bank reserves. The tax rate on reserves is the wedge between the overnight interbank interest rate, i.e., the federal funds rate, and interest paid on reserve balances held overnight at the central bank. Monetary policy varies the scarcity of reserves in the banking system in order to influence the marginal liquidity services yield on reserves. For instance, draining reserves to increase their scarcity raises the marginal liquidity services yield. A higher liquidity services yield, in turn, requires a higher interest opportunity cost of holding reserves overnight at the central bank in equilibrium, and hence a higher spread between the federal funds rate and interest paid on reserves. To sum up, monetary policy may be understood to manage the federal funds rate (the market interest rate) given interest paid on reserves at zero or otherwise, by varying the scarcity and thereby the tax rate on reserves. Second, as pointed out above, a ‘‘Treasuries only’’ asset acquisition policy leaves the decisions regarding the use of the revenue from the creation of high powered money to the fiscal authorities. For example, in 2006 the Fed transferred around 30 billion dollars to the Treasury when the Fed portfolio was nearly all Treasuries, the Fed did not pay interest on reserves, and the federal funds rate averaged around 6 percent. Given the huge volume of Treasury debt outstanding and likely to remain outstanding, the Fed could satisfy all its asset acquisition needs independently with ‘‘Treasuries only’’ for the foreseeable future if it wished to do so. Credit policy has no effect on the federal funds rate because it does not change aggregate bank reserves or interest paid on reserves. As pointed out above, the correct way to think of central bank credit policy is as debt financed fiscal policy. At the margin, the central bank returns to the Treasury the interest earned on Treasury securities that it holds; so when the central bank sells Treasuries to finance loans or to purchase mortgage backed securities, the result is just as if the Treasury financed the loans or purchases by borrowing from the public. Central bank credit policy works by interposing the government between private borrowers and lenders and exploiting the government’s creditworthiness the power to borrow credibly against future taxes to facilitate flows to distressed or favored borrowers. Doing so involves a fiscal policy decision to put taxpayer funds at risk. In contrast to holding Treasuries, or those securities with ‘‘full faith and credit’’ backing, all central bank lending carries some credit risk and exposes the central bank, and ultimately taxpayers, to losses and controversial disputes regarding credit allocation. Even central bank lending that is collateralized fully exposes taxpayers to losses if the borrower fails subsequently. For instance, emergency central bank lending that finances the withdrawal of uninsured claimants of a financial institution that fails subsequently strips that institution of collateral that would be available otherwise to cover the cost of insured deposits or other government guarantees. Even if the central bank lends only against good collateral so as not to take appreciable credit risk itself, lending to depositories and emergency credit extended to other financial institutions that have federal guarantees has the capacity to impose significant losses on taxpayers. Interest on reserves policy enables a central bank to employ a fiscal policy instrument interest on reserves to implement interest rate policy without imposing a tax in the form of below market interest on bank reserves.8 To do so, interest on reserves is set at the intended interest rate target and the central bank creates an abundance of bank reserves sufficient to drive the interbank interest rate down to the interest on reserves floor. In effect, this operating procedure attains Milton Friedman’s ‘‘optimum quantity of money’’ with respect to bank reserves, although not with respect to currency unless interest rates are zero.9 Implementing interest rate policy by employing interest on reserves together with satiation of the reserves market could yield a number of practical benefits. An abundance of costless, safe reserves could displace costly and risky private credit in the payments system and enable the central bank to limit its own credit in support of the payments system. Eliminating the tax on reserves could raise deposit rates, lower loan rates, and thereby reduce banking costs for the public. Eliminating the reserve tax secures the central bank’s control of short term interest rates, since banks would no longer have an incentive to substitute away from central bank reserves in the provision of transactions services. Finally, implementing interest rate policy at the interest on reserves floor frees monetary policy to fund credit policy independently of interest rate policy. As discussed in Section 4.4 below, it was for this reason that the Fed asked Congress in May 2008 to expedite its authority to pay interest on reserves. 8 9
Goodfriend (2002) and Keister et al. (2008) discuss the use of interest on reserves to implement interest rate policy. Friedman (1969).
M. Goodfriend / Journal of Monetary Economics 58 (2011) 1–12
5
The revenue from monetary policy would likely increase on average over time by paying interest on reserves at the market interbank rate. There would be a small loss of revenue to the fiscal authorities associated with interest paid to banks on preexisting reserve balances. However, the expansion of reserves could finance the acquisition of long term Treasuries by the central bank whose yields exhibit a positive term premium relative to short term interest rates.10 The ‘‘interest rate risk’’ incurred would be manageable if central bank liabilities continued to consist overwhelmingly of non interest bearing currency. 4. Fiscal aspects of five Federal Reserve initiatives This section describes five Fed initiatives in the credit turmoil: the Term Auction Facility, Fed lending to facilitate the acquisition of Bear Stearns by JP Morgan Chase, Fed support for AIG, emergency authority to pay interest on reserves, and the joint statement by the Treasury and the Fed on the role of the Fed in preserving financial and monetary stability. The descriptions highlight the role that fiscal policy plays in each of these initiatives, and how at times the fiscal aspects of these initiatives created problems for the Fed and for the effectiveness of its interventions to stabilize the economy. 4.1. The term auction facility In December 2007, the Fed approved the establishment of the Term Auction Facility (TAF) under which it auctioned term loans against a wide variety of collateral to depository institutions judged to be in sound condition.11 After January 2009 the minimum bid rate was interest paid on reserves. TAF loans were provided for 28 and 84 day terms. Roughly 400 billion dollars of TAF loans were outstanding in April 2009. The TAF program was established as a pure credit policy in as much as the Fed financed TAF loans with funds acquired by selling Treasury securities from its portfolio, with no effect on aggregate bank reserves. The TAF worked as follows. The credit turmoil was marked by an unprecedented elevation in rates at which banks could borrow in the interbank market. For example, the elevation was especially pronounced in the spread between 3 month LIBOR and the expected 3 month path for federal funds rate target. Banks recognized a substantial credit risk in lending to each other given that interbank lending is generally unsecured. Even if collateral were taken, the ability to liquidate it at a reasonable price could be impaired severely in a widespread default. Banks reacted by shortening the maturity at which they were willing to lend, and charging a substantial term premium for interbank lending at longer horizons such as one and three months. Bank positions in the interbank market can be highly persistent. For instance, big banks tend to be borrowers and smaller banks lenders. When the credit turmoil hit, those banks that were persistent borrowers endured a sharp persistent jump in their funding costs. Persistent borrowers at term LIBOR, for instance, would bid most aggressively for TAF term credit. By substituting TAF credit for more expensive term funding they could lower their borrowing costs. Persistent lenders of funds in the interbank market could sell their excess reserves to the Fed in exchange for Treasury securities sold by the Fed to fund its TAF loans. Since the TAF program had no effect on total bank reserves, and little if any effect on the balance of supply and demand in the federal funds market, and little effect on the creditworthiness of borrowing banks, it should not have been expected to have much sustained effect on the marginal rate paid by persistent interbank borrowers. The Fed says that the TAF program was designed to increase the access of depository institutions to funding in order to support the ability of such institutions to meet the credit needs of their customers.12 Whether or not the TAF program had much effect on the marginal interbank rate, the TAF program can be understood to have reduced funding costs of those banks caught with a persistent funding shortfall. Understood this way, the TAF program provided infra marginal relief on funding costs for persistent interbank borrowers. The TAF provided credit relief at little cost to the Fed the rate earned on TAF credit exceeded interest on the Treasury securities sold to fund it, and TAF credit was fully collateralized. However, it cannot be said that the TAF provided interest savings to banks at little risk to the taxpayer. As discussed in Section 2 above, even Fed lending that is collateralized fully exposes the deposit insurance fund, and ultimately taxpayers to losses if the borrower fails subsequently. If TAF credit financed the exit of uninsured or unsecured lenders to a bank that failed with TAF loans is outstanding, then the TAF would have stripped the bank of collateral that would have been available otherwise to cover the cost of insured deposits or other government guarantees. 4.2. Fed lending to facilitate the acquisition of Bear Stearns by JP Morgan Chase In mid March 2008 Bear Stearns was pushed to the brink of failure after losing the confidence of investors and its access to short term funding. The Fed judged that a disorderly failure of Bear Stearns would have threatened overall financial stability. After talking with the Treasury and SEC, the Fed determined that it would invoke emergency authority to provide special financing to facilitate the acquisition of Bear Stearns by JP Morgan Chase.13 In June, when the acquisition was 10 11 12 13
Campbell et al. (1997), p. 415, report an average 10-year term premium of around 1.4 percent per annum relative to the short rate. Armantier et al. (2008). Board of Governors of the Federal Reserve System (2009b), February, p. 47. Geithner (2008).
6
M. Goodfriend / Journal of Monetary Economics 58 (2011) 1–12
completed, the Fed extended roughly 29 billion dollars to the limited liability company Maiden Lane I, which was formed to facilitate the transaction by acquiring a variety of mortgage obligations, derivatives, and hedging products from Bear Stearns. This is not to question the Fed’s decision to provide financial support for the acquisition of Bear Stearns by JP Morgan Chase. The point is that the Fed’s financial support went well beyond ordinary lending to depository institutions. Institutions ordinarily eligible to borrow at the Fed discount window are depositories that hold balances at the Fed. Investment banks were not in this group. Hence, the Fed had to invoke emergency powers to lend in support of the acquisition. The Fed usually provides loans against good collateral to depositories deemed to be in sound financial condition. The Fed went beyond these three conditions in this case. First, it lent to a limited liability company Maiden Lane I. Second, it lent against assets of questionable value. Third, its loan amounted to a purchase. Maiden Lane I was funded by a 29 billion dollar loan from the Fed and a 1 billion dollar loan from JP Morgan Chase. The first 1 billion dollar loss was to be borne by JPMC any further loss up to 29 billion was to be borne by the Fed. And any realized gains beyond the 30 billion initial financing, which could occur because of revaluing the underlying assets, would accrue to the Fed. This arrangement meant that by lending to Maiden Lane I, the Fed had all of the upside of the asset valuations and all but a small fraction of the downside. In effect, the Fed ‘‘purchased’’ the assets, again, a variety of risky mortgage obligations, derivatives, and hedging products acquired from Bear Stearns.14 The Fed financed its loan to Maiden Lane I with funds from the sale of Treasury securities. Hence, the loan to Maiden Lane I was pure credit policy, which amounted to a debt financed fiscal policy purchase of a pool of risky private financial assets. The Fed effectively acknowledged this in two ways. The Fed brought Maiden Lane onto its balance sheet and recognized implicitly that its loan to Maiden Lane amounted to a purchase of the assets in Maiden Lane.15 And the Fed received a letter from the Treasury saying ‘‘if any loss arises out of the special facility extended by the FRBNY to JPMC, the loss will be treated by the FRBNY as an expense that may reduce the net earnings transferred by the FRBNY to the Treasury general fund’’.16 In April 2008, Paul Volcker described the Fed’s lending to facilitate the acquisition of Bear Stearns by JP Morgan Chase as follows: ‘‘Simply stated, the bright new financial system for all its talented participants, for all its rich rewards has failed the test of the market place. To meet the challenge, the Federal Reserve judged it necessary to take actions that extend to the very edge of its lawful and implied powers, transcending certain long embedded central banking principles and practices. The extension of lending directly to non banking financial institutions while under the authority of nominally ‘‘temporary’’ emergency powers will surely be interpreted as an implied promise of similar actions in times of future turmoil. What appears to be in substance a direct transfer of mortgage and mortgage backed securities of questionable pedigree from an investment bank to the Federal Reserve seems to test the time honored central bank mantra in time of crisis ‘‘lend freely at high rates against good collateral’’ to the point of no return.’’17 In retrospect, Volcker’s remarks can be seen as a kind of ‘‘life preserver’’ thrown to the Fed.18 Without judging whether the Fed’s actions were called for under the circumstances, but describing the Fed as having acted at the ‘‘very edge of its lawful and implied powers,’’ Volcker’s public comments could have prompted the Fed and the Treasury during the period of relative calm in April 2008 to urge Congress to appropriate resources to stabilize the financial system, should those resources be needed as the credit turmoil ran its course. Instead, Congress was not then so involved, and the Fed remained exposed to having its balance sheet utilized as an ‘‘off budget’’ arm of fiscal policy. 4.3. Federal Reserve support for AIG Allowing the Fed to be the front line of fiscal support for the financial system proved to be a problem in the fall of 2008. On September 7 the Treasury and the Federal Housing Finance Agency announced they would place Fannie Mae and Freddie Mac into conservatorship.19 Shortly thereafter, Lehman Brothers came under pressure as short term secured funding was withdrawn from the investment bank, and Lehman filed for bankruptcy on Monday, September 15th. The financial condition of American International Group (AIG), a large, complex insurance conglomerate, had also deteriorated rapidly and on Tuesday, September 16th with the full support of the Treasury, the Fed announced an 85 billion dollar loan to AIG to support the firm whose failure it judged would have significant adverse effects on the economy.20 A full scale financial panic developed on Wednesday, September 17th after a major money market mutual fund ‘‘broke the buck’’ prompting widespread withdrawals from prime money funds and forcing the liquidation of their commercial paper holdings. The ‘‘flight to safety’’ pushed the 3 month Treasury bill yield to zero on September 17th. 14
The Fed publishes the current valuations of these assets regularly. Federal Reserve Statistical Release H.4.1 ‘‘Factors Affecting Reserve Balances,’’ July 3 and September 10, 2008, and 1A Memorandum Items, September 10, 2008. 16 Paulson (2008). 17 Volcker (2008), p. 2. 18 Wessel (2009), pp. 173–174 makes a related point. 19 The summary of these events comes from Board of Governors of the Federal Reserve System (2009b), February, pp. 6–8. 20 The fiscal details of the Fed’s support for AIG are reported in Board of Governors of the Federal Reserve System (2009b), February, p. 51. 15
M. Goodfriend / Journal of Monetary Economics 58 (2011) 1–12
7
The Fed’s financial support for AIG was criticized immediately by some prominent members of Congress as a questionable commitment of taxpayer funds.21 At that point, and in light of the ongoing panic in financial markets, Fed Chairman Bernanke had little choice but to call Treasury Secretary Paulson to say that the Fed had been stretched to its limits and could not do anymore. Although Paulson apparently had been resisting such a move for months, Bernanke said it was time for the Treasury Secretary to go to Congress to seek funds and authority for a broader rescue of the financial system.22 On Thursday eve, September 18th, Paulson and Bernanke made their case to the congressional leadership that the Congress should authorize public funds to help stabilize the financial system. By that weekend, Congress and Paulson had agreed on the outlines of the 700 billion dollar Troubled Asset Relief Program (TARP).23 To convince Congress to appropriate the funds, Bernanke argued that otherwise the US economy was at risk of a severe contraction, if not another Great Depression. When the House of Representatives rejected the initial TARP bill on Monday, September 29th, stocks plunged.24 To overcome resistance to funding the TARP program, Bernanke continued to argue that the legislation was needed to prevent a severe contraction. By the time Congress was sufficiently worried to pass TARP and it was signed into law on Friday, October 3rd, the public was frightened as well. Equity markets in the United States fell by over 30 percent in the four weeks to October 10th. Risk spreads rose dramatically throughout the credit markets as never before in the credit turmoil. High yield corporate bond spreads over comparable off the run Treasuries spiked briefly to 16 percentage points and remained above 10 percentage points, well above their previous peak in the credit turmoil of 6 percentage points. A relatively modest contraction of economic activity due to financial distress associated with the deflation of house prices became the Great Recession. This is not to argue against the Fed’s support for AIG. The Fed faced a no win situation in deciding whether or not to support AIG. A decision to commit substantial taxpayer resources in support of the financial system or one that denies taxpayer resources is inherently a highly charged, political, fiscal policy matter. Whatever the Fed decided would have lacked sufficient political legitimacy and undermined its independence for further fiscal action. The Congress would have become involved chaotically either way. The point is that an independent central bank cannot be responsible for delivering or deciding upon the delivery of fiscal support for the financial system. 4.4. Authority to pay interest on reserves The Financial Services Regulatory Relief Act of 2006 gave the Fed the authority starting in 2011 to pay interest on reserves. In May 2008 the Fed asked Congress to expedite that authority to assist in its emergency credit policy initiatives. Following the passage of the Emergency Economic Stabilization Act of 2008, the Fed announced on October 6 that it would begin paying interest on required and excess reserve balances. Initially, the rate paid on excess reserves was set at a spread below the targeted federal funds rate. Shortly thereafter, with the federal funds rate trading consistently below the target rate, the spreads were eliminated. Interest on reserves helped set a floor under the federal funds rate as the Fed created nearly a trillion dollars of reserves to help finance its credit initiatives in the fourth quarter of 2008. Chairman Bernanke in his written testimony for the July 2009 Monetary Policy Report to Congress expressed the view that the authority to pay interest on reserves is perhaps the most important tool enabling the Fed to raise interest rates without first shrinking its balance sheet. However, in his July 2009 Wall Street Journal op ed, Bernanke noted that the federal funds rate slipped below interest paid on reserves in the fall of 2008 because some large lenders in the federal funds market, such as government sponsored enterprises (GSEs) Fannie Mae and Freddie Mac, and the Federal Home Loan Banks (FHLBs), are legally ineligible to receive interest on balances that they hold at the Fed.25 Thus, lending by the GSEs, the FHLBs, and others in the federal funds market could impair the power of interest on reserves to put a floor under the federal funds rate again when the Fed tries to exit the near zero federal funds rate setting. Depository institutions eligible to receive interest on reserves have an incentive to attract federal funds from the GSEs and the FHLBs, and to deposit those funds at the Fed. Such arbitrage would tend to keep the federal funds rate from falling far below interest on reserves.26 Nevertheless, such arbitrage cannot be counted upon absolutely to stabilize the federal funds rate close to interest on reserves, especially in periods of financial distress. Allowing the federal funds rate to fluctuate below interest on reserves would complicate interest rate policy needlessly by creating doubt about whether short term interest rates that matter for borrowing and lending will follow interest on reserves or the federal funds rate. The Fed has since been working to build the technical capacity to immobilize reserves to help raise the federal funds rate, if need be, without first shrinking its balance sheet. These techniques have problems of their own discussed in Section 6.1. Fortunately, a direct solution to the interest on reserves problem is available. The Fed’s July 2009 Monetary Policy Report to Congress points out on p. 37 that interest paid on bank reserves worked successfully for other central banks to put 21 22 23 24 25 26
Blackstone and Yoest (2008) and Andrews et al. (2008). Hilsenrath et al. (2008). The Economist (2008), p. 1. See The Wall Street Journal, September 30, 2008. Bernanke (2009). Bech and Klee (2009).
8
M. Goodfriend / Journal of Monetary Economics 58 (2011) 1–12
a floor under interbank rates in their economies even as bank reserves expanded aggressively.27 The Fed could ask the relevant regulatory agencies, the Treasury, and Congress to help secure the interest on reserves floor in the United States by modifying regulations for the federal funds market to exclude all but depository institutions from lending in that market, or alternatively by allowing those institutions eligible to lend in the federal funds market to earn interest on balances at the Fed. So strengthened, interest on reserves policy would provide the Fed with a precise, flexible, and reliable means of raising interest rates as the economy recovers, regardless of the size of the Fed’s balance sheet. 4.5. Joint statement by the Treasury and the Federal Reserve On March 23, 2009 the Treasury and the Fed issued a joint statement that recalled the 1951 Treasury Fed Accord on monetary policy discussed in Section 5 below.28 Released in a period of great financial distress reflected in a DOW below 7000, and one year after the acquisition of Bear Stearns by JPMC, the March 23rd statement sought to clarify the boundary of responsibilities on monetary and credit policy between the Treasury and the Fed.29 Entitled ‘‘The Role of the Federal Reserve in Preserving Financial and Monetary Stability,’’ the points of agreement were listed under the following headings:
Treasury Federal Reserve cooperation in improving the functioning of credit markets and fostering financial stability. The Federal Reserve to avoid credit risk and credit allocation. Need to preserve monetary stability. Need for a comprehensive resolution regime for systemically critical financial institutions.
Notably, the statement agreed that (i) Fed credit policy should aim to improve financial conditions broadly, and not allocate credit to narrowly defined sectors or classes of borrowers, (ii) government decisions to influence the allocation of credit are the province of the fiscal authorities, (iii) Fed credit policy should not constrain monetary policy needed to foster maximum sustainable employment and price stability, (iv) the Treasury will remove or liquidate the Maiden Lane facilities on the Fed’s balance sheet, and (v) the Fed’s independence with regard to monetary policy is critical for ensuring that monetary policy decisions are made with regard only to the long term economic welfare of the nation. The joint statement was welcome and had much to recommend it; nevertheless, it did not provide a set of principles that could serve comprehensively to clarify the boundary of responsibilities between the Treasury and the Fed. 5. Clarifying the boundary of central bank credit policy The 1951 Accord between the Treasury and the Fed was one of the most dramatic events in US financial history. The Accord ended an arrangement dating from World War II in which the Fed agreed to use its monetary policy powers to keep interest rates low to help finance the war effort. The Truman administration urged an extension of the agreement to keep interest rates low in order to hold down the cost of the huge Federal government debt accumulated during the war. Fed officials argued that keeping interest rates low would require inflationary money growth that would destabilize the economy and ultimately fail.30 The Accord famously reasserted the principle of Fed independence so that monetary policy might serve exclusively to stabilize inflation and macroeconomic activity. Congress early on recognized that the Fed needed financial independence in order to conduct monetary policy effectively. The Fed is exempted from the congressional appropriations process in order to keep the political system from abusing its money creating powers. The Fed finances its operations from interest earnings on its portfolio of securities. The Fed was given wide latitude regarding the size and composition of its balance sheet so it could react promptly, decisively, and independently to economic and financial conditions. In the early 1980s under the strong, independent leadership of Paul Volcker the Fed succeeded in establishing low inflation as the nominal anchor for monetary policy. Thus, Fed independence is today the institutional foundation for effective monetary policy. The Fed has long executed credit policy in addition to monetary policy as ‘‘lender of last resort’’ to depository institutions. Credit policy is also subject to misuse for fiscal policy purposes. However, as long as Fed lending was relatively modest, temporary, and confined to depository institutions deemed solvent, and the Fed took good collateral against its loans, the potential for fiscal misuse was limited. Although the Fed has long needed an accord for credit policy, the lack of one was not a pressing matter.31 The enormous expansion of Fed credit in the turmoil lending beyond depository institutions and acquiring non Treasury securities demands an accord for Fed credit policy to supplement the accord on monetary policy. A credit accord should set guidelines for Fed credit policy so that pressure to misuse Fed credit policy for fiscal purposes does not 27 28 29 30 31
Bowman et al. (2010). Department of the Treasury and the Federal Reserve (March 23, 2009). Lacker (2009) and Plosser (2009). Hetzel (2001) and Stein (1969). Goodfriend (1994) and Schwartz (1992).
M. Goodfriend / Journal of Monetary Economics 58 (2011) 1–12
9
undermine the Fed’s independence and impair the central bank’s power to stabilize financial markets, inflation, and macroeconomic activity. Congress bestowed independence on the Fed only because it is essential for the Fed to do its job effectively.32 A healthy democracy requires full public disclosure and discussion of the expenditure of public funds. The congressional appropriations process enables Congress to evaluate competing budgetary programs and to establish priorities for the allocation of public resources. Hence, the Fed precisely because it is exempted from the appropriations process should avoid, to the fullest extent possible, taking actions that can properly be regarded as within the province of fiscal policy and the fiscal authorities. As emphasized repeatedly above, when the Fed purchases Treasury securities it transfers all the revenue from monetary policy to the fiscal authorities and hence does not infringe on their fiscal policy prerogatives. Monetary policy, perhaps with the help of interest on reserves, respects the integrity of fiscal policy fully. Fed credit policy is another matter entirely, because all financial securities other than Treasuries or their equivalent carry some credit risk and all lending involves the Fed in potentially controversial disputes regarding credit allocation. When the Fed extends credit to private or other public entities lacking the ‘‘full faith and credit’’ backing of the US government, the Fed is allocating credit to particular borrowers, and therefore taking a fiscal action and invading the territory of the fiscal authorities. As emphasized in Sections 3 and 4.1 above, even fully collateralized lending that is riskless for the Fed exposes taxpayers to losses if the borrower fails subsequently. Fed credit that finances the exit of uninsured or unsecured lenders to a financial institution that fails while the loan is outstanding will have stripped the bank of collateral that could otherwise be available to cover the cost of insured deposits or other government guarantees. It is important to appreciate the difficulties to which the Fed exposes itself in the pursuit of credit policy initiatives that go beyond ordinary last resort lending to depository institutions. The Fed must decide how widely to expand its lending reach. Lending farther afield creates ‘‘an implied promise of similar actions in times of future turmoil,’’ as Volcker put it, which the Fed may then be inclined to accommodate.33 Fed presence in one credit market can drain lending from nearby credit channels and prompt calls for support in neighboring credit classes. The Fed must determine the relative pricing of its loans based on risk and collateral. The Fed must be accountable for its credit allocations and the returns or losses on its loans or security purchases. The public deserves transparency on Fed credit extensions beyond ordinary lending to depository institutions. Yet, congressional oversight opens the door to political interference in the Fed’s lending or non Treasury acquisitions. The Fed is exposed to pressure to exploit the central bank’s off budget status to circumvent the appropriations process. Moreover, the Fed and the fiscal authorities must cooperate on banking, financial, and payments system policy matters. This interdependence exposes the Fed to political pressure to make undesirable concessions with respect to its credit policy initiatives in return for support on other matters. Worse, the Fed could be pressured to make concessions on monetary policy to deflect pressure regarding credit policy.
5.1. ‘‘Accord’’ principles for central bank credit policy By its very nature then, credit policy has the potential to create friction between the independent central bank and the fiscal authorities. That friction is evident in the tense relationship between the Fed and Congress in the aftermath of the credit turmoil. The problem is that credit policy undoes ‘‘Treasuries only’’ so to speak, and uses some of the revenue from monetary policy to acquire non Treasury assets without the authorization of the fiscal authorities. Credit policy must direct public funds to specific borrowers, at a minimum favoring one class of creditors or one sector of the economy over another. As discussed in Section 2 above, even the central bank acquisition of government agency debt or securities packaged by government agencies is problematic. Except in the rare cases when Congress has granted ‘‘full faith and credit’’ backing to government agency debt or securities packaged by government agencies, acquisition of such securities by the central bank has allocative consequences because it steers credit in a particular direction and confers an implied preferential status enhancing that agency’s creditworthiness. Central bank credit policy must be circumscribed with clear, coherent boundaries.34 One could deny credit policy powers to the central bank altogether by requiring the central bank to pursue a ‘‘Treasuries only’’ asset acquisition policy. But credit policy has been useful in the recent turmoil and last resort lending to temporarily illiquid but solvent depositories has long been a valued part of independent central banking. Moreover, conventional last resort lending is reasonably compatible with central bank independence. Last resort lending to supervised, solvent depositories, on a short term basis, against good collateral provides multiple layers of protection against ex post losses and ex ante distortions. So the fiscal policy consequences of conventional last resort lending are likely to be minimal, and the scope for conflict with the fiscal authorities small. 32
The following paragraphs are from Broaddus and Goodfriend (2001). See the excerpt from Volcker’s April 2008 speech quoted in Section 4.2 above, and the discussion of the ‘‘limited commitment’’ problem in Goodfriend and Lacker (1999). 34 Friedman (1962), pp. 232–234. 33
10
M. Goodfriend / Journal of Monetary Economics 58 (2011) 1–12
On the other hand, expansive credit initiatives those that extend a central bank’s credit reach in scale, maturity, and collateral to unsupervised non depository institutions and the purchase of non Treasury securities inevitably carry substantial credit risk and have significant allocative consequences. Expansive credit initiatives infringe significantly on the fiscal policy prerogatives of the Treasury and Congress and properly draw the scrutiny of the fiscal authorities. Hence, expansive credit initiatives jeopardize central bank independence and should be circumscribed by agreement between the fiscal authorities and the central bank. Such reasoning suggests the following three principles as the basis for a Treasury Fed ‘‘accord’’ for central bank credit policy. To reiterate, Congress bestows Fed independence only because it is necessary for the Fed to do its job effectively. Hence, the Fed should perform only those functions that must be carried out by an independent central bank. The main idea is to preserve the Fed’s independence to react flexibly and decisively to stabilize economic and financial conditions while maintaining a credible commitment to low inflation. Principle 1: As a long run matter, a significant, sustained departure from a ‘‘Treasuries only’’ asset acquisition policy is incompatible with Fed independence. Principle 2: The Fed should adhere to ‘‘Treasuries only’’ except for occasional, temporary, well collateralized ordinary last resort lending to solvent, supervised depository institutions. Principle 3: Fed credit initiatives beyond ordinary last resort lending should be undertaken only with prior agreement of the fiscal authorities, and only as bridge loans accompanied by take outs arranged and guaranteed in advance by the fiscal authorities. 5.2. Pinnacle financial oversight authority The Dodd Frank financial reform law of 2010 establishes a ‘‘Financial Stability Oversight Council’’ chaired by the Treasury Secretary with nine other voting members including the Fed Chairman and five non voting members. The Stability Council is authorized to identify risks to the financial stability of the United States, promote market discipline, and respond to emerging threats to the stability of the United States financial system. The Stability Council was established as a pinnacle authority with responsibility for the entire financial system. The idea was that the pinnacle authority should have jurisdiction to address threats to financial stability originating anywhere in the economy with all the regulatory tools available in government. In other words, the guiding principle was to create a ‘‘one stop shop’’ systemic regulator. Among other things, to be the ‘‘one stop shop’’ systemic regulator that reformers had in mind, the pinnacle authority would have the responsibility to grant or deny fiscal support for particular firms or sectors in financial distress. During the debate on the Dodd Frank legislation, consideration was given to making the Fed the pinnacle financial oversight authority. The discussions in Sections 4.2 and 4.3 above make clear, however, that the pinnacle financial oversight authority cannot be lodged in an independent central bank. To grant or deny taxpayer support for the financial system is fiscal policy. To force a central bank to make fiscal policy, especially such contentious fiscal policy decisions, would politicize the central bank and destroy its independence. The Dodd Frank law chose correctly to lodge the pinnacle authority in a Stability Council outside the Fed. 6. Central banking at the zero bound and in the exit strategy The Fed’s 2 trillion dollar balance sheet and near zero interest rate policy stance appear to have achieved some stability in economic and financial conditions.35 However, the Fed must remain poised to tighten financial conditions on short notice if conditions warrant or to expand its balance sheet further if economic conditions weaken again. To be fully flexible at the zero interest bound the Fed must position itself to raise interest rates promptly against inflation if that becomes necessary, even after expanding its balance sheet well beyond two trillion dollars against deflation if that proves to be necessary. Credibility against deflation is tied to credibility against inflation. This section considers two proposals to strengthen policy flexibility at the zero interest bound and in the exit strategy. First, it considers monetary policy options proposed by the Fed to immobilize reserves and help raise the federal funds rate, if need be, without first shrinking the Fed’s balance sheet. Second, it proposes to employ fiscal policy to enlarge surplus capital on the Fed’s balance sheet to secure its financial independence to pay interest on reserves in the exit strategy even after acting aggressively against deflation. 6.1. Monetary policy options to raise interest rates The Fed has proposed a number of options to raise the federal funds rate in addition to interest on reserves.36 The Fed’s suggestions involve monetary policy since they work by reducing or immobilizing aggregate bank reserves. The Fed 35 36
Gagnon et al. (2010). Board of Governors of the Federal Reserve System (2009b), July, pp. 34–37.
M. Goodfriend / Journal of Monetary Economics 58 (2011) 1–12
11
contemplates publicly four options for draining reserves. The Fed itself acknowledges that two have serious drawbacks. First, the Fed could reduce reserves by selling some of its holdings of Treasury securities. The Fed recognizes that this option is limited by the stock of Treasuries available in its portfolio. Second, the Treasury could sell securities and deposit the proceeds with the Fed. But the Fed rightly does not want to rely on the Treasury to achieve its policy objectives. The Fed is more favorably disposed to the third option. The Fed could drain bank reserves and absorb federal funds otherwise lent by GSEs, FHLBs, and other institutions by arranging large scale reverse repurchase agreements. Such reverses would involve the sale by the Fed of securities from its portfolio with an agreement to buy the securities back at a slightly higher price. There are problems with this approach, too. Large scale reverses would expose the Fed to substantial counterparty risk. This could complicate the Fed’s management of financial markets, especially in time of financial turmoil. Simply put, the Fed should not put itself in the position of having to depend heavily on contractual arrangements with the private sector. Fourth, the Fed could drain bank reserves by offering interest earning term deposits to banks, analogous to certificates of deposit that banks offer their customers. The Fed is favorably disposed to this option, too. But, again, this option is not without problems. Fed term deposits would compete with Treasury bills and potentially create friction with the Treasury. Term deposits would be close substitutes for bank reserves. The introduction and management of interest on term deposits could destabilize the demand for reserves and complicate federal funds rate targeting with monetary policy as contemplated. Another problem is that to raise the federal funds rate significantly with monetary policy, the Fed would have to return aggregate reserves to a level near to those prior to the credit turmoil. Large scale reverse repurchase agreements or term deposit operations would have to be undertaken in advance over a span of time to pre position monetary policy to take the modest actions needed to adjust the federal funds rate precisely and flexibly when the time comes. Finally, draining bank reserves with large scale non monetary managed liabilities would turn the Fed into a financial intermediary and jeopardize its independence by facilitating the perpetual funding of credit policy independently of monetary policy or interest rate policy. Moreover, there is no reason for the Fed to issue managed liabilities if the regulation of the federal funds market is modified as suggested in Section 4.4 above to secure the interest on reserves floor for the federal funds rate.
6.2. Enlarging surplus capital to secure financial independence In order to employ interest on reserves independently to exit the zero interest bound without first shrinking its balance sheet, the Fed must secure its financial independence to pay interest on reserves. Confining its assets to short term Treasury bills would hedge much if not all of the Fed’s interest on reserves risk, since T bill rates would follow interest on reserves closely as interest rates moved higher. However, short term T bills are perfect substitutes for reserves at the zero interest bound, and confining the Fed’s portfolio to T bills would preclude further monetary stimulus at zero interest. To act decisively against deflation at the zero interest bound, the Fed must acquire long term securities instead of T bills and be prepared to expand its balance sheet well beyond 2 trillion dollars, if need be.37 In other words, the Fed must be willing to assume a large maturity mismatch between its assets and liabilities and expose itself to substantial interest rate risk. The Fed must also be prepared to reverse field should inflation become the problem, and raise interest on reserves with trillions of dollars of reserves and long term securities on its balance sheet. Three broad outcomes are possible. There would not be a problem if the Fed managed stabilization policy so that long term interest rates remain stable near a sustainable 3 percent real yield plus 1 to 2 percent inflation expectations consistent with the Fed’s implicit inflation target. If the yield curve also remained upward sloping as interest on reserves exited the zero bound, then the Fed could finance interest on reserves with earnings on its long term securities. A negative cash flow problem could arise if the Fed were either insufficiently preemptive against deflation or insufficiently preemptive against inflation. If the Fed were too slow against deflation, buying long term securities at high prices and very low interest, then interest earnings could be insufficient to pay interest on reserves subsequently if the Fed had to raise interest on reserves against inflation before it could shrink its balance sheet. Alternatively, if the Fed were insufficiently preemptive against inflation, a negative cash flow problem could arise if subsequently the Fed had to raise interest on reserves above long term interest rates before it could shrink its balance sheet. To be fully flexible against both deflation and inflation at the zero bound the Fed should enlarge surplus capital enough to self insure the payment of interest on reserves in any scenario. The Fed could build up its capital account in one of two ways with the cooperation of the fiscal authorities. The fiscal authorities could transfer new Treasury securities directly to the Fed. Or the fiscal authorities could let the Fed retain interest earnings to build up surplus capital gradually. The first alternative is preferable because it is immediate. Either way, the enlargement of the Fed’s capital account would have no fiscal cost as long as the Fed did not draw on the interest or the principal of its surplus capital. The Fed would simply return 37 Goodfriend (2000), Section 3, pp. 1018–1028 discusses the mechanics of monetary policy stimulus at the zero interest bound, Section 2, pp. 1013– 1018 discusses negative interest on reserves policy. See Curdia and Woodford (this issue) and Gertler and Karadi (this issue) for an analysis of credit policy stimulus at the zero bound.
12
M. Goodfriend / Journal of Monetary Economics 58 (2011) 1–12
to the Treasury all the interest on the Treasuries in its enlarged capital account.38 Nevertheless, enlarging the Fed’s surplus capital would secure its financial independence and greatly improve its macroeconomic stabilization powers at the zero interest bound. 7. Conclusion The proposed classification of central bank policies into monetary policy, credit policy, and interest on reserves policy could be utilized productively in the Fed’s internal deliberations and in its external communications to (1) improve the transparency of the Fed’s operations for purposes of accountability and credibility, (2) distinguish the fiscal aspects of Fed policies for the purpose of clarifying the boundary of its independent responsibilities, (3) help secure the Fed’s operational and financial independence to raise interest rates precisely and flexibly to sustain a non inflationary recovery from the Great Recession, and (4) reinforce the sense that the Fed has the political independence and the determination to unwind its emergency liquidity measures while limiting their inflationary potential. References Andrews, A.L., de la Merced, M.J., Walsh, M.W., 2008. Fed’s $85 Billion Loan Rescues Insurer. The New York Times, September 17, p. 1. Armantier, O., Krieger, S., McAndrews, J., 2008. The Term Auction Facility, Current Issues in Economics and Finance. Federal Reserve Bank of New York July, pp. 1–9. Bech, M.L., Klee, E., 2009. The Mechanics of a Graceful Exit: Interest on Reserves and Segmentation in the Federal Funds Market. Federal Reserve Bank of New York Staff Report No. 416, December. Bernanke, B., 2009. The Fed’s Exit Strategy (Op-ed). Wall Street Journal, July 21. Blackstone, B., Yoest, P., 2008. Bailouts Turn Up Heat on Fed Chief. Wall Street Journal.com September 18. Board of Governors of the Federal Reserve System, 2009a. Federal Reserve System Monthly Report on Credit and Liquidity Programs and the Balance Sheet, June 2009 to the present. Board of Governors of the Federal Reserve System, 2009b. Monetary Policy Report to Congress, February 24 and July 21. Board of Governors of the Federal Reserve System and Other Financial Regulators, 2010. Advisory on Interest Rate Risk Management, January 6. Bowman, D., Gagnon, E., Leahy, M., 2010. Interest on Excess Reserves as a Monetary Policy Instrument: The Experience of Foreign Central Banks. International Finance Discussion Papers, Board of Governors of the Federal Reserve System, No. 996, March. Broaddus, J.A., Goodfriend, M., 2001. What assets should the Federal Reserve buy? Federal Reserve Bank of Richmond. Economic Quarterly (Winter), 7–22. Campbell, J.Y., Lo, A.W., MacKinlay, A.C., 1997. The Econometrics of Financial Markets. Princeton University Press, Princeton, NJ. Curdia, V., Woodford, M., this issue. The central bank balance sheet as an instrument of monetary policy. Journal of Monetary Economics, January, doi:10.1016/j.jmoneco.2010.09.011. Department of the Treasury and the Federal Reserve, 2009. The role of the Federal Reserve in preserving financial and monetary stability, Joint Press Release, Board of Governors of the Federal Reserve System, March 23. Friedman, M., 1962. Should there be an independent monetary authority?. In: Yeager, L.B. (Ed.), In Search of a Monetary Constitution. Harvard University Press, Cambridge, MA. Friedman, M., 1969. The Optimum Quantity of Money and Other Essays. Aldine Publishing Company, Chicago, IL. Gagnon, J., Raskin M., Remache, J., Sack, B., 2010. Large-scale asset purchases by the Federal Reserve: did they work? Federal Reserve Bank of New York Staff Report No. 441, March. Geithner, T., 2008. Actions by the New York Fed in response to liquidity pressures in financial markets (Testimony), Federal Reserve Bank of New York, April 3. Gertler, M., Karadi, P., this issue. A model of unconventional monetary policy. Journal of Monetary Economics, January, doi:10.1016/j.jmoneco.2010.10. 004. Goldfarb, Z.A., 2010. Rep. Barney Frank Warns of Fannie, Freddie Risks. The Washington Post, A10 March 6. Goodfriend, M., 2002. Interest on reserves and monetary policy. Federal Reserve Bank of New York. Policy Review (May), 77–84. Goodfriend, M., 2000. Overcoming the zero bound on interest rate policy. Journal of Money, Credit, and Banking (November), 1007–1035. Goodfriend, M., 1994. Why we need an ‘‘Accord’’ for Federal Reserve Credit Policy. Journal of Money, Credit, and Banking (August), 572–580. Goodfriend, M., Lacker, J., 1999. Limited commitment and central bank lending. Federal Reserve Bank of Richmond. Economic Quarterly (Fall), 1–27. Greenspan, A., 2010. The Crisis. Mimeo, Second Draft, March 9. Hetzel, R., 2001. The Treasury-Fed Accord: a new narrative account. Federal Reserve Bank of Richmond. Economic Quarterly (Winter), 33–64. Hilsenrath, J., Solomon, D., Paletta, D., 2008. Crisis mode: Paulson, Bernanke strained for consensus in bailout. Wall Street Journal.com November 10. Keister, T., Martin, A., McAndrews, J., 2008. Divorcing money from monetary policy. Federal Reserve Bank of New York. Policy Review (September), 41–56. Lacker, J., 2009. Government lending and monetary policy. Federal Reserve Bank of Richmond, March 2. Paulson, H., 2008. Letter to Timothy Geithner, March 17. Plosser, C., 2009. Ensuring sound monetary policy in the aftermath of the crisis. Federal Reserve Bank of Philadelphia, February 27. Schwartz, A., 1992. The misuse of the Fed’s discount window. Federal Reserve Bank of St. Louis Economic Review (September/October), 58–69. Stein, H., 1969. The liberation of monetary policy. In: The Fiscal Revolution in America. University of Chicago Press, Chicago, pp. 241–80. The Economist, 2008. America’s Bail-out Plan: The Doctors’ Bill. September 25. The Wall Street Journal, September 30, 2008, p.1. Volcker, P., 2008. Remarks by Paul Volcker at a Luncheon of the Economic Club of New York. April 8. Wessel, D., 2009. In: In Fed We Trust: Ben Bernanke’s War on the Great Panic. Crown Business, New York.
38 In early 2010, financial regulators in the United States issued collectively an advisory to remind institutions of supervisory expectations regarding sound practices for managing interest rate risk. See Board of Governors of the Federal Reserve System and Other Financial Regulators (2010). To build up the Fed’s surplus capital is to follow that advice.
Journal of Monetary Economics 58 (2011) 13–16
Contents lists available at ScienceDirect
Journal of Monetary Economics journal homepage: www.elsevier.com/locate/jme
Discussion
Discussion of goodfriend Lee E. Ohanian UCLA, United States
a r t i c l e in f o Article history: Received 19 October 2010 Accepted 19 October 2010 Available online 27 October 2010
Marvin Goodfriend’s paper addresses a set of the classic questions in monetary policy that extend at least back to Walter Bagehot’s 1873 book Lombard Street: What should a central bank do, how should they do it, and what institutional arrangements must exist for the Central Bank to effectively carry out its duties? My discussion reviews the contributions of Marvin’s analysis, then focuses on three themes. These themes are (1) the fundamental importance of Treasury backing of the Fed when the Fed pays interest on reserves, (2) related issues associated with the exiting from a policy that pays interest on reserves, particularly if inflation or inflation expectations are rising, and (3) the broad theme of what institutions promote central bank independence. Marvin’s paper provides an excellent summary of Fed programs adopted during the financial crisis, and presents an outstanding discussion of these programs with a focus on their implications for Fed independence. He develops this discussion around a three part taxonomy of Fed procedures: traditional monetary policy (targeting overnight rates), credit policy (the acquisition of risky securities) and interest rate policy associated with paying interest on reserves. This taxonomy shows just how much monetary policy has changed. Fed procedure formerly was largely confined to traditional monetary policy (a Fed funds target), with no credit policy and no interest on reserves. But recently, traditional monetary policy has taken a backseat to non standard credit policies, in which the Fed has purchased a variety of non Treasury (risky) securities and in which the Fed has paid interest on reserves. These differences are highlighted in the following table: Characterizing federal reserve policy Policy action
Pre-2008
2008 and After
Funds rate target Credit policy Interest on reserves
Yes No No
Irrelevant Enormous Yes
My interpretation of Marvin’s essay is as follows: ‘‘Central bank independence is incompatible with anything other than narrow central banking: occasional and limited lender of last resort functions to only depository institutions, and a central bank balance sheet with treasuries only.’’ This theme differs from today’s Fed. Specifically, it argues against actions such as acquiring mortgage backed securities, the rescue of various financial institutions, and also argues against extending funding to non depository taking institutions, even though this latter action has been supported by some of the strongest Fed critics, such as John Taylor. E-mail address:
[email protected] 0304-3932/$ - see front matter & 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jmoneco.2010.10.007
14
L.E. Ohanian / Journal of Monetary Economics 58 (2011) 13–16
Marvin develops a set of six principles to help the Fed become a politically independent central bank with narrowly defined policy procedures and to help the Fed get there while trying to minimize the risk of high inflation. Several of these principles are associated with paying interest on reserves and issues about exiting from such a policy. Marvin is correct to focus attention in this area, particularly the principle that the Treasury must commit to support the Fed under such a policy. To see this within a model context, I will show that paying interest on reserves may increase the risks of inflation, and increase economic volatility upon the exit of such a policy. To see how reserves on interest may exacerbate inflation risk, consider two equations that appear in several classes of models. The first equation is a standard period government budget constraint. The second equation is present value budget balance. Period government budget constraint: Bt Rt1 Bt1 ¼ Gt þ Pt Pt
Tt
Mt Mt1 Pt
Present value government budget constraint, where Qs is discounting, 1 X Ms Ms1 ¼0 Qs Gs Ts Ps s t Combining these two equations yields a third equation which shows that the real value of debt is equal to the present value of future surpluses, including seignorage: 1 X Bt Ms Ms1 ¼ Qs Gs Ts Pt Ps t tþ1 Given these equations, suppose the Fed is paying interest on reserves, which gives rise to the following modified government budget constraint: 1 X Bt Ms Ms1 , Rm ¼ Qs Gs þðRm 1ÞMs1 Ts s s Z1 Pt Ps t tþ1 Now suppose that inflation expectations increase. This could lead to possible cash flow problems for the Fed, which in turn could lead to a self fulfilling inflation. To see this, note that in the absence of Treasury support, this cash flow problem may force the Fed to increase money creation to make it feasible for the Fed to pay interest on reserves. Moreover, the government budget constraint shows that failure of the Treasury to support the Fed which I interpret as no change in spending or federal tax collections puts further upward pressure on the government budget constraint, and thus ultimately implies higher seignorage is required to satisfy this constraint. Note that this is not a run on the Fed, which can print money. Instead, it is a run on the dollar based purely on expectations. This example suggests the possibility that inflation could become self fulfilling in the absence of Treasury commitment. A simple way to address the potential problem of Fed cash flow problems would be to let the Fed pile up a buffer stock. This could be accomplished by the Fed keeping profits from trading activities, rather than the long standing practice of turning those profits over to the Treasury. Next, I consider some implications of an exit from an interest on reserves policy. To do this, I use a Lucas Stokey cash credit economy, which delivers an interest elastic demand for money. Variables are cash consumption goods, credit consumption goods, hours worked, nominal interest on government bonds, and nominal interest on reserves. In the budget constraint below, the variable X refers to lump sum taxes/transfers. Preferences are given by X t max b fuðcmt ,crt Þ,vð1 ht Þg t
The budget constraint, with multiplier l, is given by Rb bt1 ð1 tt Þwt ht Rm mt1 bt mt þ þ cmt þ crt ¼ t1 þ þ t1 þXt pt pt pt pt pt The cash in advance (CIA) constraint, with multiplier m, is given by: Now, consider pegging interest on reserves to the government bond rate: b Rm t ¼ aRt
The household first order conditions of this model for labor, money holdings, and bond holdings, are given by vuð1 ht Þ ¼ lt ð1 tt Þwt
lt pt
lt pt
¼b
Rm t ðlt þ 1 þ mt þ 1 Þ pt þ 1
¼b
Rbt lt þ 1 pt þ 1
L.E. Ohanian / Journal of Monetary Economics 58 (2011) 13–16
15
First note that in the absence of interest rate pegging, we obtain ucmt lt þ mt ¼ ¼ Rbt ucrt lt I assume a monetary policy that pegs the interest rate on reserves to be proportional to the market bond interest rate. One issue to first note is that this policy may lead to the type of indeterminacy stressed by Sargent and Wallace in their classic paper about interest rate pegging, which is associated with indeterminacy of the initial price level in the economy. While this issue is interesting and merits discussion, I will leave it aside for the time being to focus on what can happen in such an economy when policy shifts from paying interest on reserves to no longer paying interest on reserves. With the interest rate pegging policy, however, the relative price between cash and credit goods is constant, as is velocity: Rm t ¼a Rbt The first order conditions for this problem shows that the relative price of cash credit goods, which ordinarily is the market nominal rate, is now constant as a result of interest on reserves. Consumption velocity is also constant. Thus, the inflation process is a pure quantity theoretic process, just as in a standard CIA model, as opposed to a Cagan type process with variable velocity.
6
Primary surplus and debt (federal government)
90 80
4
70 2
60
0 19 6 19 0 62 19 6 19 4 6 19 6 6 19 8 7 19 0 7 19 2 7 19 4 76 19 7 19 8 8 19 0 8 19 2 8 19 4 8 19 6 8 19 8 9 19 0 9 19 2 9 19 4 9 19 6 9 20 8 00 20 0 20 2 0 20 4 0 20 6 08
50
2
40 30
4
20 6
10 0
8 Surplus as % of GDP
24
Debt as % of GDP, right scale
Spending plus transfers (% of GDP)
22 20 18 16 14 12
19
6 19 0 6 19 2 6 19 4 6 19 6 6 19 8 7 19 0 7 19 2 7 19 4 7 19 6 7 19 8 8 19 0 8 19 2 8 19 4 8 19 6 8 19 8 9 19 0 9 19 2 9 19 4 9 19 6 9 20 8 0 20 0 0 20 2 0 20 4 0 20 6 08
10
16
L.E. Ohanian / Journal of Monetary Economics 58 (2011) 13–16
But unlike a pure CIA model, interest on reserves turns money into a quasi credit good, and thus the first order condition for the allocation of time no longer fluctuates as a result of variations in inflation. This is useful, as it can improve the efficiency of allocations which otherwise are distorted by inflation. But there are challenging issues upon leaving such a policy, particularly if the exit is associated with rising inflation and/or inflation expectations. Exiting the policy means that the relative price of cash and credit goods is no longer fixed, and that the allocation of time, as well as resources across sectors will be distorted by variations in expected inflation. Perhaps most important, an exit from interest on reserves changes the inflation process from a pure quantity equation to a Cagan type process in which variable velocity, as well as money growth, impact inflation. Exiting an interest on reserves policy, particularly with rising inflation, means an additional source of real volatility in the economy, welfare reducing volatility that is driven by inflation. Moreover, it also means a change in inflation dynamics that could possibly increase the inflation rate through rising velocity. Thus, I firmly agree with Marvin that exiting from such a policy be accompanied by an explicit and full commitment from the Treasury to be prepared to either raise taxes or reduce spending during such a transition. Should we worry about Treasury commitment? Probably. The figures show graphs of federal debt as a percentage of GDP, the primary surplus as a percentage of GDP, and federal spending as a percentage of GDP. The graphs paint a picture of expanding expenditures and debt, with little sign of fiscal discipline in sight. Note that there is no tendency for the primary surplus to rise when debt rises. In fact, there is a negative correlation between the surplus and the debt, raising questions about Treasury commitment to supporting the fed. In terms of government spending, we see that government purchases and transfers are rising as a percentage of GDP. This will likely become even higher in the future, reflecting unfunded government pension liabilities, the health care expansion, and demographics associated with social security. These patterns also raise questions about Treasury commitment. The final issue I will discuss is related to promoting central bank independence. Marvin argues that only narrow central banking that is, a central bank that holds only treasuries, and one that only intervenes as an occasional and limited lender of last resort to depository institutions is consistent with a sufficiently independent central bank. I am sympathetic, but how do we commit the Fed to this course of action? Eliminating systemic problems would clearly help, but this is of course hard to do, particularly if there continues to be large maturity mismatches between assets and liabilities. The huge political demands for intervention make limited actions in financial markets almost impossible. Marvin’s excellent paper advances an important question. Maintaining a narrowly confined and independent Fed requires other intervention organizations for possible future crises. It is interesting that the Resolution Trust Corporation was reasonably successful in dealing with the Savings and Loan crisis about 20 years ago. But the Dodd Frank Reform Act puts the Fed front and center again in this process as it specifies the Fed as the main organization that must deal with institutions in financial exigency, and in my view, codifies bailouts. In a word of huge possible financial moral hazard, Fed independence will continue to be strained, and the solid principles that form the basis of Marvin’s excellent essay may be lost to the failure of effectively managing too big to fail. Reference Bagehot, 1986. Lombard street: a description of the money market. In: John-Stevas (Ed.), The Collected Works of Walter Bagehot. Oxford University Press, New York.
Journal of Monetary Economics 58 (2011) 17–34
Contents lists available at ScienceDirect
Journal of Monetary Economics journal homepage: www.elsevier.com/locate/jme
A model of unconventional monetary policy Mark Gertler a,, Peter Karadi a,b,1 a b
New York University, United States National Bank of Hungary, Hungary
a r t i c l e in fo
abstract
Article history: Received 4 May 2010 Received in revised form 7 October 2010 Accepted 8 October 2010 Available online 17 October 2010
We develop a quantitative monetary DSGE model with financial intermediaries that face endogenously determined balance sheet constraints. We then use the model to evaluate the effects of the central bank using unconventional monetary policy to combat a simulated financial crisis. We interpret unconventional monetary policy as expanding central bank credit intermediation to offset a disruption of private financial intermediation. Within our framework the central bank is less efficient than private intermediaries at making loans but it has the advantage of being able to elastically obtain funds by issuing riskless government debt. Unlike private intermediaries, it is not balance sheet constrained. During a crisis, the balance sheet constraints on private intermediaries tighten, raising the net benefits from central bank intermediation. These benefits may be substantial even if the zero lower bound constraint on the nominal interest rate is not binding. In the event this constraint is binding, though, these net benefits may be significantly enhanced. & 2010 Elsevier B.V. All rights reserved.
1. Introduction Over most of the post war period the Federal Reserve conducted monetary policy by manipulating the Federal Funds rate in order to affect market interest rates. It largely avoided lending directly in private credit markets. After the onset of the sub prime crisis in August 2007, the situation changed dramatically. To address the deterioration in both financial and real activity, the Fed directly injected credit into private markets. It began in the fall of 2007 by expanding the ease at which financial institutions could obtain discount window credit and by exchanging government debt for high grade private debt. The most dramatic interventions came following the collapse of the shadow banking system that followed the Lehman Brothers failure. At this time the Fed began directly lending in high grade credit markets. It provided backstop funding to help revive the commercial paper market. It also intervened heavily in mortgage markets by directly purchasing agency debt and mortgage backed securities. There is some evidence to suggest that these policies have been effective in reducing credit costs. Commercial paper rates relative to similar maturity Treasury Bills fell dramatically after the introduction of backstop facilities in this market. Credit spreads for agency debt and mortgage backed securities also fell in conjunction with the introduction of the direct lending facilities. The Fed’s balance sheet provides the most concrete measure of its credit market intervention: since August 2007 the quantity of assets it has held has increased from about 800 billion to over two trillion, with most of the increase coming after the Lehman collapse. Most of the increase in assets the central absorbed were financial instruments previously held by the shadow banks. Further, it financed its balance sheet expansion largely with interest bearing reserves, which are in effect overnight government debt. Thus, over this period the Fed has attempted to offset the disruption of a considerable Corresponding author.
E-mail address:
[email protected] (M. Gertler). Much thanks to Bob Hall and Hal Cole for comments on an earlier draft and to Luca Guerrieri for computational help.
1
0304-3932/$ - see front matter & 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jmoneco.2010.10.004
18
M. Gertler, P. Karadi / Journal of Monetary Economics 58 (2011) 17–34
fraction of private financial intermediation by expanding central bank intermediation. To do so, it has exploited its ability to raise funds quickly and cheaply by issuing (in effect) riskless government debt. Overall, the Fed’s unconventional balance sheet operations appeared to provide a way for it to stimulate the economy even after the Federal Funds reached the zero lower bound. At the same time, operational models of monetary policy have not kept pace with the dramatic changes in actual practice. There is of course a lengthy contemporary literature on quantitative modeling of conventional monetary policy, beginning with Christiano et al. (2005) and Smets and Wouters (2007). The baseline versions of these models, however, assume frictionless financial markets. They are thus unable to capture financial market disruptions that could motivate the kind of central bank interventions in loan markets that are currently in play. Similarly, models which do incorporate financial market frictions, such as Bernanke et al. (1999) have not yet explicitly considered direct central bank intermediation as a tool of monetary policy. Work that has tried to capture this phenomenon has been mainly qualitative as opposed to quantitative (e.g. Kiyotaki and Moore, 2008; Adrian and Shin, 2009). Accordingly, the objective of this paper is to try to fill in this gap in the literature: the specific goal is to develop a quantitative macroeconomic model where it is possible to analyze the effects of unconventional monetary policy in the same general manner that existing frameworks are able to study conventional monetary policy. To be clear, we do not attempt to explicitly model the sub prime crisis. However, we do try to capture the key elements relevant to analyzing the Fed’s credit market interventions. In particular, the current crisis has featured a sharp deterioration in the balance sheets of many key financial intermediaries. As many observers argue, the deterioration in the financial positions of these institutions has had the effect of disrupting the flow of funds between lenders and borrowers. Symptomatic of this disruption has been a sharp rise in various key credit spreads as well as a significant tightening of lending standards. This tightening of credit, in turn, has raised the cost of borrowing and thus enhanced the downturn. The story does not end here: the contraction of the real economy has reduced asset values throughout, further weakening intermediary balance sheets, and so on. It is in this kind of climate, that the central bank has embarked on its direct lending programs. To capture this kind of scenario, accordingly we incorporate financial intermediaries within an otherwise standard macroeconomic framework. To motivate why the condition of intermediary balance sheets influences the overall flow of credit, we introduce a simple agency problem between intermediaries and their respective depositors. The agency problem introduces endogenous constraints on intermediary leverage ratios, which have the effect of tying overall credit flows to the equity capital in the intermediary sector. As in the current crisis, a deterioration of intermediary capital will disrupt lending and borrowing in a way that raises credit costs. To capture unconventional monetary policy in this environment, we allow the central bank to act as intermediary by borrowing funds from savers and then lending them to investors. Unlike private intermediaries, the central bank does not face constraints on its leverage ratio. There is no agency problem between the central bank and its creditors because it can commit to always honoring its debt (which as we noted earlier is effectively government debt.) Thus, in a period of financial distress that has disrupted private intermediation, the central bank can intervene to support credit flows. On the other hand, we allow for the fact that, everything else equal, public intermediation is likely to be less efficient than the private intermediation. When we use the model to evaluate these credit interventions, we take into account this trade off. Section 2 presents the baseline model. The framework is closely related to the financial accelerator model developed by Bernanke et al. (BGG, 1999).2 That approach emphasized how balance sheet constraints could limit the ability of non financial firms to obtain investment funds. Firms effectively borrowed directly from households and financial intermediaries were simply a veil. Here, as we discussed, financial intermediaries may be subject to endogenously determined balance sheet constraints. In addition, we allow for the central bank to lend directly to private credit markets. Another difference from BGG is that, we use as a baseline framework the conventional monetary business cycle framework developed by Christiano et al. (CEE, 2005a), Smets and Wouters (SW, 2007) and others. We adopt this approach because this framework has proven to have reasonable empirical properties. Here we use it to study not only conventional interest policy but also unconventional credit market interventions by the central bank. Section 3 presents a quantitative analysis of the model. We illustrate how financial factors may amplify and propagate some conventional disturbances. We also consider a disturbance to the underlying quality of intermediary assets (a ‘‘valuation shock’’) and then show how this kind of disturbance could create a contraction in real activity that mirrors some of the basic features of the current crisis. As we show, either an actual decline in asset quality or the expectation (e.g. ‘‘news’’) of a future decline can trigger a crisis. We then illustrate the extent to which central bank credit interventions could moderate the downturn. Finally, we show the stabilization benefits from credit policy are magnified if the zero lower bound on nominal interest rates is binding. In Section 4, we undertake a normative analysis of credit policy. We first solve for the optimal central bank credit intervention in crisis scenario considered in Section 3. We do so under different assumptions about the efficiency costs of central bank intermediation. We then compute for each case the net welfare gains from the optimal credit market
2 The theory underlying the financial accelerator was developed in Bernanke and Gertler (1989). For quantitative frameworks, in addition to BGG, see Carlstrom and Fuerst (1997), Iacovello (2005), Goodfriend and McCallum (2007), Gilchrist et al. (2009), Jermann and Quadrini (2010), Mendoza (2010) and Christiano et al. (2010). As an example of recent theory, see Brunnermeier and Sannikov (2010).
M. Gertler, P. Karadi / Journal of Monetary Economics 58 (2011) 17–34
19
intervention. We find that as long as the efficiency costs are quite modest, the gains may be quite significant. As we discuss, this finding suggests a formal way to think about the central bank’s choice between direct credit interventions versus alternatives such as equity injections to financial intermediaries. Within our baseline model the two policies are equivalent if we abstract from the issue of efficiency costs. For certain types of lending, e.g. securitized high grade assets such as mortgage backed securities, the costs of central bank intermediation might be relatively low. In this case, direct central bank intermediation may be justified. In other cases, e.g. C&I loans that requires constant monitoring of borrowers, central bank intermediation may be highly inefficient. In this instance, capital injections may be the preferred route. Concluding remarks are in Section 5.
2. The baseline model The core framework is the monetary DSGE model with nominal rigidities developed by CEE and SW. To this, we add financial intermediaries that transfer funds between households and non financial firms. An agency problem constrains the ability of financial intermediaries to obtain funds from households. We also include a disturbance to the quality of capital. Absent financial frictions, this shock introduces only a modest decline in output, as the economy works to replenish the effective capital stock. With frictions in the intermediation process, however, the shock creates a significant capital loss in the financial sector, which in turn induces tightening of credit and a significant downturn. As we show, it is in this kind of environment that there is a potential role for central bank credit interventions. There are five types of agents in the model: households, financial intermediaries, non financial goods producers, capital producers, and monopolistically competitive retailers. The latter are in the model only to introduce nominal price rigidities. In addition, there is a central bank that conducts both conventional and unconventional monetary policy. Without financial intermediaries the model is isomorphic to CEE and SW. As we show, though, the addition of financial intermediaries adds only a modest degree of complexity. It has, however, a substantial effect on model dynamics and associated policy implications. We now proceed to characterize the basic ingredients of the model.
2.1. Households There is a continuum of identical households of measure unity. Each household consumes, saves and supplies labor. Households save by lending funds to competitive financial intermediaries and possibly also by lending funds to the government. Within each household there are two types of members: workers and bankers. Workers supply labor and return the wages they earn to the household. Each banker manages a financial intermediary and similarly transfers any earnings back to the household. The household thus effectively owns the intermediaries that its bankers manage. The deposits it holds, however, are in intermediaries that is does not own. Finally, within the family there is perfect consumption insurance. As we make clear in the next section, this simple form of heterogeneity within the family allows us to introduce financial intermediation in a meaningful way within an otherwise representative agent framework. At any moment in time the fraction 1 f of the household members are workers and the fraction f are bankers. Over time an individual can switch between the two occupations. In particular, a banker this period stays banker next period with probability y, which is independent of history (i.e., of how long the person has been a banker.) The average survival time for a banker in any given period is thus 1=ð1 yÞ. As will become clear, we introduce a finite horizon for bankers to insure that over time they do not reach the point where they can fund all investments from their own capital. Thus every period ð1 yÞf bankers exit and become workers. A similar number of workers randomly become bankers, keeping the relative proportion of each type fixed. Bankers who exit give their retained earnings to their respective household. The household, though, provides its new bankers with some start up funds, as we describe in the next sub section. Let Ct be consumption and Lt family labor supply. Then households preferences are given by 1 X w 1þj ð1Þ bi lnðCt þ i hC t þ i1 Þ Lt þ i maxEt 1þj i 0 with 0 o b o 1, 0 o h o1 and w, j 40. As in CEE and SW we allow for habit formation to capture consumption dynamics. As in Woodford (2003), we consider the limit of the economy as it become cashless, and thus ignore the convenience yield to the household from real money balances. Both intermediary deposits and government debt are one period real bonds that pay the gross real return Rt from t 1 to t. In the equilibrium we consider, the instruments are both riskless and are thus perfect substitutes. Thus, we impose this condition from the outset. Thus let Bt + 1 be the total quantity of short term debt the household acquires, Wt, be the real wage, Pt net payouts to the household from ownership of both non financial and financial firms and, Tt lump sum taxes. Then the household budget constraint is given by Ct ¼ Wt Lt þ Pt þTt þ Rt Bt Bt þ 1
ð2Þ
20
M. Gertler, P. Karadi / Journal of Monetary Economics 58 (2011) 17–34
Note that Pt is net the transfer the household gives to its members that enter banking at t. Finally, as will be clear later, it will not matter in our model whether households hold government debt directly or do so indirectly via financial intermediaries (who in turn issue deposits to households.) Let Rt denote the marginal utility of consumption. Then the household’s first order conditions for labor supply and consumption/saving are standard:
Rt Wt ¼ wLjt
ð3Þ
with
Rt ¼ ðCt hC t1 Þ1 bhEt ðCt þ 1 hC t Þ1 and Et bLt,t þ 1 Rt þ 1 ¼ 1
ð4Þ
with
Lt,t þ 1
Rt þ 1 Rt
2.2. Financial intermediaries Financial intermediaries lend funds obtained from households to non financial firms. In addition to acting as specialists that assist in channeling funds from savers to investors, they engage in maturity transformation. They hold long term assets and fund these assets with short term liabilities (beyond their own equity capital).3 In addition, financial intermediaries in this model are meant to capture the entire banking sector, i.e., investment banks as well as commercial banks. Let Njt be the amount of wealth or net worth that a banker/intermediary j has at the end of period t; Bjt + 1 the deposits the intermediary obtains from households, Sjt the quantity of financial claims on non financial firms that the intermediary holds and Qt the relative price of each claim. The intermediary balance sheet is then given by Qt Sjt ¼ Njt þBjt þ 1
ð5Þ
For the time being, we ignore the possibility of the central bank supplying funds to the intermediary. As we noted earlier, household deposits with the intermediary at time t, pay the non contingent real gross return Rt + 1 at t + 1. Thus Bjt + 1 may be thought of as the intermediary’s debt and Njt as its equity capital. Intermediary assets earn the stochastic return Rkt + 1 over this period. Both Rkt + 1 and Rt + 1 will be determined endogenously. Over time, the banker’s equity capital evolves as the difference between earnings on assets and interest payments on liabilities: Njt þ 1 ¼ Rkt þ 1 Qt Sjt Rt þ 1 Bjt þ 1
ð6Þ
¼ ðRkt þ 1 Rt þ 1 ÞQt Sjt þ Rt þ 1 Njt
ð7Þ
Any growth in equity above the riskless return depends on the premium Rkt + 1 Rt + 1 the banker earns on his assets, as well as his total quantity of assets, QtSjt. i Let b Lt,t þ i be the stochastic discount the banker at t applies to earnings at t + i. Since the banker will not fund assets with a discounted return less than the discounted cost of borrowing, for the intermediary to operate in period i the following inequality must apply: i
Et b Lt,t þ 1 þ i ðRkt þ 1 þ i Rt þ 1 þ i Þ Z 0,
i Z0
With perfect capital markets, the relation always holds with equality: the risk adjusted premium is zero. With imperfect capital markets, however, the premium may be positive due to limits on the intermediary’s ability to obtain funds.4 So long as the intermediary can earn a risk adjusted return that is greater than or equal to the return the household can earn on its deposits, it pays for the banker to keep building assets until exiting the industry. Accordingly, the banker’s objective is to maximize expected terminal wealth, given by Vjt ¼ maxEt
1 X i
0
i iþ1
ð1 yÞy b
Lt,t þ 1 þ i ðNjt þ 1 þ i Þ ¼ maxEt
1 X i
i iþ1
ð1 yÞy b
Lt,t þ 1 þ i ½ðRkt þ 1 þ i Rt þ 1 þ i ÞQt þ i Sjt þ i þ Rt þ 1 þ i Njt þ i
0
ð8Þ
3 In Gertler and Kiyotaki (2010), we consider a generalization of this framework that has banks manage liquidity risks (stemming from idiosyncratic shocks to firm investment opportunities) via an interbank market. In this setup, financial frictions may also affect the functioning of the interbank market. 4 See Justiniano et al. (2010a, 2010b) for evidence that this premium is highly countercyclical and in fact opened up widely during the 2007–2009 recession.
M. Gertler, P. Karadi / Journal of Monetary Economics 58 (2011) 17–34
21
i
To the extent the discounted risk adjusted premium in any period, b Lt,t þ i ðRkt þ 1 þ i Rt þ 1 þ i Þ, is positive, the intermediary will want to expand its assets indefinitely by borrowing additional funds from households. To motivate a limit on its ability to do so, we introduce the following moral hazard/costly enforcement problem: at the beginning of the period the banker can choose to divert the fraction l of available funds from the project and instead transfer them back to the household of which he or she is a member.5 The cost to the banker is that the depositors can force the intermediary into bankruptcy and recover the remaining fraction 1 l of assets. However, it is too costly for the depositors to recover the fraction l of funds that the banker diverted. Accordingly for lenders to be willing to supply funds to the banker, the following incentive constraint must be satisfied: Vjt Z lQt Sjt
ð9Þ
The left side is what the banker would lose by diverting a fraction of assets. The right side is the gain from doing so. We can express Vjt as follows: Vjt ¼ vt Qt Sjt þ Zt Njt
ð10Þ
vt ¼ Et fð1 yÞbLt,t þ 1 ðRkt þ 1 Rt þ 1 Þ þ bLt,t þ 1 yxt,t þ 1 vt þ 1 g Zt ¼ Et fð1 yÞ þ bLt,t þ 1 yzt,t þ 1 Zt þ 1 g
ð11Þ
with
where xt,t þ i Qt þ i Sjt þ i =Qt Sjt , is the gross growth rate in assets between t and t + i, and zt,t þ i Njt þ i =Njt is the gross growth rate of net worth. The variable vt has the interpretation of the expected discounted marginal gain to the banker of expanding assets QtSjt by a unit, holding net worth Njt constant, and while Zt is the expected discounted value of having another unit of Njt, holding Sjt constant. With frictionless competitive capital markets, intermediaries will expand borrowing to the point where rates of return will adjust to ensure vt is zero. The agency problem we have introduced, however, may place limits on this arbitrage. In particular, as we next show, when the incentive constraints is binding, the intermediary’s assets are constrained by its equity capital. Note first that we can express the incentive constraints as
Zt Njt þvt Qt Sjt Z lQt Sjt
ð12Þ
If this constraint binds, then the assets the banker can acquire will depend positively on his/her equity capital: Qt Sjt ¼
Zt l vt
Njt ¼ ft Njt
ð13Þ
where ft is the ratio of privately intermediated assets to equity, which we will refer to as the (private) leverage ratio. Holding constant Njt, expanding Sjt raises the bankers’ incentive to divert funds. The constraint (13) limits the intermediaries leverage ratio to the point where the banker’s incentive to cheat is exactly balanced by the cost. In this respect the agency problem leads to an endogenous capital constraint on the intermediary’s ability to acquire assets. Given Njt 4 0, the constraint binds only if 0 o vt o l. In this instance, it is profitable for the banker to expand assets (since vt 40). Note that in this circumstance the leverage ratio that depositors will tolerate is increasing in vt. The larger is vt, the greater is the opportunity cost to the banker from being forced into bankruptcy. If vt increases above l, the incentive constraint does not bind: the franchise value of the intermediary always exceed the gain from diverting funds. In the equilibrium we construct below, under reasonable parameter values the constraint always binds within a local region of the steady state. We can now express the evolution of the banker’s net worth as Njt þ 1 ¼ ½ðRkt þ 1 Rt þ 1 Þft þ Rt þ 1 Njt Note that the sensitivity of Njt + 1 to the ex post realization of the excess return Rkt + 1 ft . In addition, it follows that
ð14Þ Rt + 1 is increasing in the leverage ratio
zt,t þ 1 ¼ Njt þ 1 =Njt ¼ ðRkt þ 1 Rt þ 1 Þft þ Rt þ 1 xt,t þ 1 ¼ Qt þ 1 Sjt þ 2 =Qt St þ 1 ¼ ðft þ 1 =ft ÞðNjt þ 1 =Nt Þ ¼ ðft þ 1 =ft Þzt,t þ 1 Importantly, all the components of ft do not depend on firm specific factors. Thus to determine total intermediary demand for assets we can sum across individual demands to obtain Qt St ¼ ft Nt
ð15Þ
where St reflects the aggregate quantity of intermediary assets and Nt denotes aggregate intermediary capital. In the general equilibrium of our model, variation in Nt, will induce fluctuations in overall asset demand by intermediaries. Indeed, a crisis will feature a sharp contraction in Nt. 5
One way the banker may divert assets is to pay out large bonuses and dividends to the household.
22
M. Gertler, P. Karadi / Journal of Monetary Economics 58 (2011) 17–34
We can derive an equation of motion for Nt, by first recognizing that it is the sum of the net worth of existing banker/ intermediaries, Net, and the net worth of entering (or ‘‘new’’) bankers, Nnt. Nt ¼ Net þ Nnt Since the fraction y of bankers at t
ð16Þ 1 survive until t, Net is given by
Net ¼ y½ðRkt Rt Þft1 þ Rt Nt1
ð17Þ
Observe that the main source of variation in Net will be fluctuations in the ex post return on assets Rkt. Further, the impact on Net is increasing in the leverage ratio ft . As we noted earlier, newly entering bankers receive ‘‘start up’’ funds from their respective households. We suppose that the startup money the household gives its new banker a transfer equal to a small fraction of the value of assets that exiting bankers had intermediated in their final operating period. The rough idea is that how much the household feels that its new bankers need to start, depends on the scale of the assets that the exiting bankers have been intermediating. Given that the exit probability is i.i.d., the total final period assets of exiting bankers at t is ð1 yÞQt St1 . Accordingly we assume that each period the household transfers the fraction o=ð1 yÞ of this value to its entering bankers. Accordingly, in the aggregate, Nnt ¼ oQt St1
ð18Þ
Combining (17) and (18) yields the following equation of motion for Nt. Nt ¼ y½ðRkt Rt Þft1 þ Rt Nt1 þ oQt St1 Observe that o helps pin down the steady state leverage ratio QS/N. Indeed, in the next section we calibrate o to match this evidence. The resulting value, as we show, is quite small. 2.3. Credit policy In the previous section we characterized how the total value of privately intermediated assets, QtSpt, is determined. We now suppose that the central bank is willing to facilitate lending. Let QtSgt be the value of assets intermediated via government assistance and let QtSt be the total value of intermediated assets: i.e., Qt St ¼ Qt Spt þ Qt Sgt
ð19Þ
To conduct credit policy, the central bank issues government debt to households that pays the riskless rate Rt + 1 and then lends the funds to non financial firms at the market lending rate Rkt + 1. We suppose that government intermediation involves efficiency costs: in particular, the central bank credit involves an efficiency cost of t per unit supplied. This deadweight loss could reflect the costs of raising funds via government debt. It might also reflect costs to the central bank of identifying preferred private sector investments. On the other hand, the government always honors its debt: thus, unlike the case with private financial institutions there is no agency conflict than inhibits the government from obtaining funds from households. Put differently, unlike private financial intermediation, government intermediation is not balance sheet constrained.6 An equivalent formulation of credit policy has the central bank issue government debt to financial intermediaries. Intermediaries in turn fund their government debt holdings by issuing deposits to households that are perfect substitutes. Assuming the agency problem applies only to the private assets it holds, a financial intermediary is not constrained in financing its government debt holdings. Thus, only the funding of private assets by financial institutions is balance sheet constrained. As in the baseline scenario, the central bank is able to elastically issue government debt to fund private assets. It is straightforward to show that the equilibrium conditions in the scenario are identical to those in the baseline case. (The identical intermediary balance sheet constraint on private assets holds). One virtue of this scenario is that the intermediary holdings of government debt are interpretable as interest bearing reserves.7 Accordingly, suppose the central bank is willing to fund the fraction ct of intermediated assets: i.e., Qt Sgt ¼ ct Qt St
ð20Þ
It issues government bonds Bgt equal to ct Qt St to fund this activity. Its net earnings from intermediation in any period t thus equals (Rkt + 1 Rt + 1)Bgt. These net earnings provide a source of government revenue and must be accounted for in the budget constraint, as we discuss later. Since privately intermediated funds are constrained by intermediary net worth, we can rewrite Eq. (19) to obtain Qt St ¼ ft Nt þ ct Qt St ¼ fct Nt 6 As Wallace (1981) originally noted, for government financial policy to matter it is important to identify what is special about government intermediation. Sargent and Wallace (1981) provide an early example of how credit policy could matter, based on a setting of limited participation in credit markets. For related analyses see Holmstrom and Tirole (1998), Curdia and Woodford (2010) and Gertler and Kiyotaki (2010). 7 This analysis concentrates on the central bank’s direct lending programs which we think were the most important dimension of their balance sheet activities. See Gertler and Kiyotaki (2010) for a formal characterization of the different types of credit market interventions that the Federal Reserve and Treasury pursued in the current crisis.
M. Gertler, P. Karadi / Journal of Monetary Economics 58 (2011) 17–34
23
where ft is the leverage ratio for privately intermediated funds (see Eqs. (13) and (15)), and where fct is the leverage ratio for total intermediated funds, public as well as well private:
fct ¼
1 f 1 ct t
Observe that fct depends positively on the intensity of credit policy, as measured by ct . Later we describe how the central might choose ct to combat a financial crisis.
2.4. Intermediate goods firms We next turn to the production and investment side of the model economy. Competitive non financial firms produce intermediate goods that are eventually sold to retail firms. The timing is as follows: at the end of period t, an intermediate goods producer acquires capital Kt + 1 for use in production in the subsequent period. After production in period t+ 1, the firm has the option of selling the capital on the open market. There are no adjustment costs at the firm level. Thus, the firm’s capital choice problem is always static, as we discuss below. The firm finances its capital acquisition each period by obtaining funds from intermediaries. To acquire the funds to buy capital, the firm issues St claims equal to the number of units of capital acquired Kt + 1 and prices each claim at the price of a unit of capital Qt. That is, QtKt + 1 is the value of capital acquired and QtSt is the value of claims against this capital. Then by arbitrage: Qt Kt þ 1 ¼ Qt St
ð21Þ
We assume that there are no frictions in the process of non financial firms obtaining funding from intermediaries. The intermediary has perfect information about the firm and has no problem enforcing payoffs. This contrasts with the process of the intermediary obtaining funding from households. Thus, within our model, only intermediaries face capital constraints on obtaining funds. These constraints, however, affect the supply of funds available to non financial firms and hence the required rate of return on capital these firms must pay.8 Conditional on this required return, however, the financing process is frictionless for non financial firms. The firm is thus able to offer the intermediary a perfectly state contingent security, which is best though of as equity (or perfectly state contingent debt.) At each time t, the firm produces output Yt, using capital and labor Lt, and by varying the utilization rate of capital, Ut + 1. Let At denote total factor productivity and let xt denote the quality of capital (so that xt Kt is the effective quantity of capital at time t). Then production is given by a Yt ¼ At ðUt xt Kt Þa L1 t
ð22Þ
Following Merton (1973) and others, the shock xt is meant to provide a simple source of exogenous variation in the value of capital. In the context of the model, it corresponds to economic depreciation (or obsolescence) of capital. We emphasize though, that the market value of an effective unit of capital Qt is determined endogenously as we show shortly. Let Pmt be the price of intermediate goods output. Assume further that the replacement price of used capital is fixed at unity. Then at time t, the firm chooses the utilization rate and labor demand as follows: Pmt a
Yt ¼ duðUt Þxt Kt Ut
Pmt ð1 aÞ
Yt ¼ Wt Lt
ð23Þ
ð24Þ
Given that the firm earns zero profits state by state, it simply pays out the ex post return to capital to the intermediary. Accordingly Rkt + 1 is given by h i Pmt þ 1 axt þY1t þKt1þ 1 þ Qt þ 1 dðUt þ 1 Þ xt þ 1 ð25Þ Rkt þ 1 ¼ Qt Given that the replacement price of capital that has depreciated is unity, then the value of the capital stock that is left over is given by ðQt þ 1 dðUt þ 1 ÞÞxt þ 1 Kt þ 1 .9 Observe that the valuation shock xt þ 1 provides a source of variation in the return to capital. Note also that the current asset price will in general depend on beliefs about the expected future path of xt þ i . 8 Many non-financial corporations hold significant cash reserves which raises the issue of whether they can simply self-finance investment. However, the evidence suggests that these cash holdings are typically precautionary balances held by firms to meet unanticipated expenditure needs and not cash that is free to use for investment expenditures. In particular, the firms that hold cash balances as a buffer are typically those with imperfect access to credit markets, for which lines of credit are prohibitively expensive. See Acharya et al. (2010) and the references therein. 9 As we make clear in the next sub-section, we assume that adjustment costs are on net rather than gross investment, so that the replacing worn out equipment does not involve adjustment costs.
24
M. Gertler, P. Karadi / Journal of Monetary Economics 58 (2011) 17–34
2.5. Capital producing firms At the end of period t, competitive capital producing firms buy capital from intermediate goods producing firms and then repair depreciated capital and build new capital. They then sell both the new and re furbished capital. As we noted earlier, the cost of replacing worn out capital is unity. The value of a unit of new capital is Qt. While there are no adjustment costs associated with refurbishing capital, we suppose that there are flow adjustment costs associated with producing new capital. We assume households own capital producers and are the recipients of any profits. Let It be gross capital created and Int It dðUt Þxt Kt be net capital created, and Iss the steady state investment. Then discounted profits for a capital producer are given by 1 X I þI bTt Lt, t ðQt 1ÞInt f nt ss ðInt þ Iss Þ ð26Þ maxEt Int 1 þ Iss t t with Int It dðUt Þxt Kt where f ð1Þ ¼ f uð1Þ ¼ 0 and f 00 ð1Þ 4 0, and where dðUt Þxt Kt is the quantity of capital refurbished. As in CEE, we allow for flow adjustment costs of investment, but restrict these costs to depend on the net investment flow.10 Note that because of the flow adjustment costs, the capital producer may earn profits outside of steady state. We assume that they rebate these profits lump sum back to households. Note also that all capital producers choose the same net investment rate. (For this reason, we do not index Int by producer type.) The first order condition for investment gives the following ‘‘Q’’ relation for net investment: Int þIss Int þ 1 þ Iss 2 f uðÞ Et bLt,t þ 1 f uðÞ ð27Þ Qt ¼ 1þ f ðÞ þ Int1 þ Iss Int þ Iss 2.6. Retail firms Final output Yt is a CES composite of a continuum of mass unity of differentiated retail firms, that use intermediate output as the sole input. The final output composite is given by "Z #e=ðe1Þ 1
ðe1Þ=e
Yft
Yt ¼
0
df
ð28Þ
where Yft is output by retailer f. From cost minimization by users of final output: e Pft Yt Yft ¼ Pt Pt ¼
"Z
1 0
ð29Þ
#1=ð1eÞ Pft1e df
ð30Þ
Retailers simply re package intermediate output. It takes one unit of intermediate output to make a unit of retail output. The marginal cost is thus the relative intermediate output price Pmt. We introduce nominal rigidities following CEE. In particular, each period a firm is able to freely adjust its price with probability 1 g. In between these periods, the firm is able to index its price to the lagged rate of inflation. The retailers pricing problem then is to choose the optimal reset price Pt* to solve " # 1 i X P Y max Et gi bi Lt,t þ i t ð1 þ pt þ k1 ÞgP Pmt þ i Yft þ i ð31Þ Pt þ i k 1 i 0 where pt is the rate of inflation from t i to t. The first order necessary conditions are given by Et " # 1 i X Pt Y gP i i g b Lt,t þ i ð1 þ pt þ k1 Þ mPmt þ i Yft þ i ¼ 0 Pt þ i k 1 i 0
ð32Þ
with
m¼
1 1 1=e
From the law of large numbers, the following relation for the evolution of the price level emerges. g
P Pt ¼ ½ð1 gÞðPt Þ1e þ gðPt1 Pt1 Þ1e 1=ð1eÞ
10
Adjustment costs are on net rather than gross investment to make the capital decision independent of the market price of capital.
ð33Þ
M. Gertler, P. Karadi / Journal of Monetary Economics 58 (2011) 17–34
25
2.7. Resource constraint and government policy Output is divided between consumption, investment, government consumption, Gt and expenditures on government intermediation, tct Qt Kt þ 1 . We suppose further that government expenditures are exogenously fixed at the level G. The economy wide resource constraint is thus given by Int þIss ðInt þ Iss Þ þ Gþ tct Qt Kt þ 1 ð34Þ Yt ¼ Ct þIt þ f Int 1 þ Iss where capital evolves according to Kt þ 1 ¼ xt Kt þ Int
ð35Þ
Government expenditures, further, are financed by lump sum taxes and government intermediation: Gþ tct Qt Kt þ 1 ¼ Tt þ ðRkt Rt ÞBgt1
ð36Þ
where government bonds, Bgt 1, finance total government intermediated assets, Qt ct1 St1 . We suppose monetary policy is characterized by a simple Taylor rule with interest rate smoothing. Let it be the net nominal interest rate, i the steady state nominal rate, and Y*t the natural (flexible price equilibrium) level of output. Then it ¼ ð1 rÞ½iþ kp pt þ ky ðlogYt logYt Þ þ rit1 þ et
ð37Þ
where the smoothing parameter r lies between zero and unity, and where et is an exogenous shock to monetary policy, and where the link between nominal and real interest rates is given by the following Fisher relation 1 þit ¼ Rt þ 1
Et Pt þ 1 Pt
ð38Þ
We suppose that the interest rate rule is sufficient to characterize monetary policy in normal times. In a crisis, however, we allow for credit policy. In particular, we suppose that at the onset of a crisis, which we define loosely to mean a period where credit spreads rise sharply, the central bank injects credit in response to movements in credit spreads, according to the following feedback rule:
ct ¼ c þ nEt ½ðlogRkt þ 1 logRt þ 1 Þ ðlogRk logRÞ
ð39Þ
where c is the steady state fraction of publicly intermediated assets and logRk logR is the steady state premium. In addition, the feedback parameter is positive. According to this rule, the central bank expands credit as the spread increase relative to its steady state value. In addition, we suppose that in a crisis the central bank abandons its proclivity to smooth interest rates. In this case it sets the smoothing parameter r equal to zero. By proceeding this way we believe we are capturing how the central bank behaved in practice as the crisis unfolded. Further, under smoothing, most of the effect of monetary policy works through the effect on the expected path of future short rates. It is reasonable to suppose that during the crisis the central bank perceived that its ability to manage expectations of the future had diminished, leading it to adjust the current interest rate at a faster pace. This completes the description of the model. 3. Model analysis 3.1. Calibration Table 1 lists the choice of parameter values for our baseline model. Overall there are 18 parameters. Fifteen are conventional. Three (l, x, y) are specific to our model. We begin with the conventional parameters. For the discount factor b, the depreciation rate d, the capital share a, the elasticity of substitution between goods, e, and the government expenditure share, we choose conventional values. Also, we normalize the steady state utilization rate U at unity. We use estimates from Primiceri et al. (2006) to obtain values for most of the other conventional parameters, which include: the habit parameter h, the elasticity of marginal depreciation with respect to the utilization rate, z, the inverse elasticity of net investment to the price of capital Zi , the relative utility weight on labor w, the Frisch elasticity of labor supply j1 , the price rigidity parameter, g, and the price indexing parameter gp . Since the policy rule the authors estimate is somewhat non standard, we instead use the conventional Taylor rule parameters of 1.5 for the feedback coefficient on inflation, kp , and 0.5 for the output gap coefficient, ky , along with a value of 0.8 for the smoothing parameter. For simplicity, we use minus the price markup as a proxy for the output gap. Our choice of the financial sector parameters the fraction of capital that can be diverted l, the proportional transfer to is meant to be suggestive. We pick these parameters to hit the entering bankers x, and the survival probability y following three targets: a steady state interest rate spread of one hundred basis points; a steady state leverage ratio of four; and an average horizon of bankers of a decade. We base the steady state target for the spread on the pre 2007 spreads between mortgage rates and government bonds and between BAA corporate versus government bonds. The steady state
26
M. Gertler, P. Karadi / Journal of Monetary Economics 58 (2011) 17–34
Table 1 Parameters. Households
b h
w j
0.990 0.815 3.409 0.276
Discount rate Habit parameter Relative utility weight of labor Inverse Frisch elasticity of labor supply
0.381 0.002 0.972
Fraction of capital that can be diverted Proportional transfer to the entering bankers Survival rate of the bankers
0.330 1.000 0.025 7.200
Effective capital share Steady state capital utilization rate Steady state depreciation rate Elasticity of marginal depreciation with respect to utilization rate
1.728
Inverse elasticity of net investment to the price of capital
4.167 0.779 0.241
Elasticity of substitution Probability of keeping prices fixed Measure of price indexation
1.5 0.50/4 0.8 0.200
Inflation coefficient of the Taylor rule Output gap coefficient of the Taylor rule Smoothing parameter of the Taylor rule Steady state proportion of government expenditures
Financial Intermediaries
l
o y Intermediate good firms
a U dðUÞ
z Capital Producing Firms
Zi Retail firms
e g gP Government
kp ky ri G Y
leverage ratio is trickier to calibrate. For investment banks and commercial banks, which were at the center of the crisis, leverage ratios (assets to equity) were extraordinarily high: typically in the range of 25 30 for the former and 15 20 for the latter. Much of this leverage reflected housing finance. For the corporate and non corporate business sectors this ratio is closer to two in the aggregate. Ideally one would like to extend the model to a multi sector setting which accounts for the differences in leverage ratios. In the interest of tractability, however, we stick with our one sector setting and choose a leverage ratio of four, which roughly captures the aggregate data.11
3.2. Experiments We begin with several experiments designed to illustrate how the model behaves. We then consider a ‘‘crisis’’ experiment that mimics some of the basic features of the current downturn. We then consider the role of central bank credit policy in moderating the crisis. Finally, we explore the implications of the zero lower bound on nominal interest rates. Fig. 1 shows the response of the model economy to three disturbances: a technology shock, a monetary shock, and shock to intermediary net worth. In each case, the direction of the shock is set to produce a downturn. The figure then shows the responses of three key variables: output, investment and the premium. In each case the solid line shows the response of the baseline model. The dotted line gives the response of the same model, but with the financial frictions removed. The technology shock is a negative one percent innovation in TFP, with a quarterly autoregressive factor of 0.95. The intermediary balance sheet mechanism produces a modest amplification of the decline in output in the baseline model relative to the conventional DSGE model. The amplification is mainly the product of a substantially enhanced decline in investment: on the order of 50 percent relative to the frictionless model. The enhanced response of investment in the baseline model is a product of the rise in the premium, plotted in the last panel on the right. The unanticipated decline in investment reduces asset prices, which produces a deterioration in intermediary balance sheets, pushing up the premium. The increase in the cost of capital, further reduces capital demand by non financial firms, which enhances the downturn in investment and asset prices. In the conventional model without financial frictions, of course, the premium is fixed at zero. 11 Note that the calibration implies that the fraction of assets the banker can divert is high, more than 30 percent. This is because the target steady state leverage ratio that helps pin down this parameter is relatively low. With modest elaborations of the model it is possible to make this value much lower. The key is to have the leverage ratio high in sectors that are investing (see Gertler and Kiyotaki, 2010).
M. Gertler, P. Karadi / Journal of Monetary Economics 58 (2011) 17–34
−1.5
0
−0.2 −0.4 −0.6 −0.8
0
20
−0.1 −0.15 −0.2
0
20 Quarters
40
40
−1 −2 0
20
40
I
0
−1 0
20 Quarters
Financial Accelerator
0.4 0.2 0 −0.2
0
20
40
k
0.8 0.6 0.4 0.2 0
0
20
40
E[R ]−R
−0.5
−1.5
k
0.6
E[R ]−R
I
0.5 %Δ from ss
−0.05
20
0
−3
40
Y
0
0
1 %Δ from ss
m %Δ from ss
−6
40
Y
0
N %Δ from ss
20
−4
Annualized %Δ from ss
−1
0 −2
Annualized %Δ from ss
%Δ from ss
a %Δ from ss
−0.5
E[R ]−R
I
2
40
Annualized %Δ from ss
Y
0
27
k
0.4 0.3 0.2 0.1 0
0
20 Quarters
40
DSGE
Fig. 1. Responses to Technology (a), Monetary (m) and Wealth (w) Shocks.
The monetary shock is an unanticipated 25 basis point increase in the short term interest rate. The effect on the short term interest rate persists due to interest rate smoothing by the central bank. Financial frictions lead to greater amplification relative to the case of the technology shock. This enhanced amplification is due to the fact that, everything else equal, the monetary policy shock has a relatively large effect on investment and asset prices. The latter triggers the financial accelerator mechanism. At the core of the amplification mechanism in the first two experiments is procyclical variation in intermediary balance sheets. To illustrate this, we consider a redistribution of wealth from intermediaries to households. In particular, we suppose that intermediary net worth declines by one percent and is transferred to households. In the model with no financial frictions, this redistribution has no effect (it is just a transfer of wealth within the family). The decline in intermediary in our baseline model, however, produces a rise in the premium, leading to a subsequent decline in output and investment. 3.2.1. Crisis experiment We now turn to the crisis experiment. The initiating disturbance is a decline in capital quality. What we are trying to capture is a shock to the quality of intermediary assets that produces an enhanced decline in the net worth of these institutions, due to their high degree of leverage. In this rough way, we capture the broad dynamics of the sub prime crises. Note that there will be both an exogenous and endogenous component to the decline in asset values that the shock generates. The initial decline in capital reduces asset values by reducing the effective quantity of capital. There is, however, also a second round effect: due to the leverage ratio constraint, the weakening of intermediary balance sheets induces a drop in asset demand, reducing the asset price Qt (the price per effective unit of capital) and investment. The endogenous fall in Qt further shrinks intermediary balance sheets. The overall contraction is magnified by the degree of leverage. It is best to think of this shock as a rare event. Conditional on occurring, however, it obeys an AR(1) process. We fix the size of the shock so that the downturn is of broadly similar magnitude to the one we have recently experienced. The initiating shock is a five percent decline in capital quality, with a quarterly autoregressive factor of 0.66. Absent any changes in investment, the shock produces a roughly 10 percent decline in the effective capital stock over a two year period. The loss in value of the housing stock relative to the total capital stock was in this neighborhood. Later we consider an ‘‘unrealized’’ news shock, where the private sector expects a deterioration of capital quality that is never materialized. This will allow us to make clear that the source of the financial crisis is the decline in asset values, as opposed to the physical destruction of capital. We first consider the disturbance to the economy without credit policy and then illustrate the effects of credit policy. For the time being, we ignore the constraint imposed by the zero lower bound on the nominal interest, but then turn to this consideration.
M. Gertler, P. Karadi / Journal of Monetary Economics 58 (2011) 17–34
ξ
−2 −4 0
20
5 0 −5
40
0
−5 K
%Δ from ss 0
20
40
−2 −4
4 2 0 −2 −4
0
Annualized %Δ from ss
−50
0
20 Quarters
40
20
0
20 π
20 Quarters
Financial Accelerator
−20 20
40
Q
10 0 −10
0
20
40
i
0
0
0
0
40
5
−5
40
20
40
L
N
0
−100
0
−6
20 I
Annualized %Δ from ss
%Δ from ss %Δ from ss
40
−10 −20
0 %Δ from ss
0
0
0 −5
40
%Δ from ss
5
20
20
5
C %Δ from ss
%Δ from ss
Y
0
E[Rk]−R
R Annualized %Δ from ss
0
Annualized %Δ from ss
%Δ from ss
28
40
5 0 −5
0
20 Quarters
40
DSGE
Fig. 2. Responses to a Capital Quality Shock.
As Fig. 2 illustrates, in the model without financial frictions, the shock produces only a modest decline in output. Output falls a bit initially due to the reduced effective capital stock. Because capital is below its steady state, however, investment picks up. Individuals consume less and eventually work more. By contrast, in the model with frictions in the intermediation process, there is a sharp recession. The deterioration in intermediary asset quality induces a firesale of assets to meet balance sheet constraints. The market price of capital declines as result. Overall, on impact intermediary capital drops more than 50 percent, which is more than 10 times the initial drop in capital quality. As we noted earlier, the enhanced decline is due to the combination of the endogenous decline in Qt and the high degree of intermediary leverage. Associated with the drop in intermediary capital, is a sharp increase in the spread between the expected return on capital and the riskless rate. Both investment and output drop as a result. Output initially falls about three percent relative to trend and then decreases to about six percent relative to trend. Though the model does not capture the details of the recession, it does produce an output decline of similar magnitude. Recovery of output to trend does not occur until roughly five years after the shock. This slow recovery is also in line with current projections. Contributing to the slow recovery is the delayed movement of intermediary capital back to trend. It is mirrored in persistently above trend movement in the spread. Note that over this period the intermediary sector is effectively deleveraging: it is building up equity relative to assets. Thus the model captures formally the informal notion of how the need for financial institutions to deleverage can slow the recovery of the economy. 3.2.2. Credit policy response We now consider credit interventions by the central bank. Fig. 3 considers several different intervention intensities. In the first case, the feedback parameter n in the policy rule given by Eq. (39) equals 10. At this value, the credit intervention is roughly of similar magnitude to what has occurred in proactive (based on assets absorbed by the Federal Reserve on its balance sheet, as a fraction of total assets in the economy). The solid line portrays this case. In the second, the feedback parameter is raised to 100, which increases the intensity of the response, bringing it closer to the optimum (as we show in the next section). The dashed line portrays this case. Finally, for comparison, the dashed and dotted line portrays the case with no credit market intervention. In each instance, the credit policy significantly moderates the contraction. The prime reason is that central intermediation dampens the rise in the spread, which in turn dampens the investment decline. The moderate intervention (n ¼ 10) produces an increase in the central bank balance sheet equal to approximately seven percent of the value of the capital stock. This is roughly in accord with the degree of intervention that has occurred in practice. The aggressive
ξ
−2 −4 0
20
0 −5
40
0
−5 40
0
0
20
40
20
20
20 0 −20
40
0 %Δ from ss
0
20
40
20
0 −10
0
20
0 −5 0
40
5 0 −5 0
20 Quarters
40
ψ
20 Percent
20
40
i
5
40
40
Q
10
Annualized %Δ from ss
Annualized %Δ from ss 20 Quarters
40
I
π
−50 0
0
L 4 2 0 −2 −4
N
0
−100
0 −2 −4 −6
K %Δ from ss
%Δ from ss %Δ from ss
20
−10 −20
0 −5
40
%Δ from ss
%Δ from ss
%Δ from ss
0
0
20
5
C
5
0
k
5
Y
29
E[R ]−R
R Annualized %Δ from ss
0
Annualized %Δ from ss
%Δ from ss
M. Gertler, P. Karadi / Journal of Monetary Economics 58 (2011) 17–34
10 0
0
Baseline Credit Policy (ν=10)
20 Quarters
40
Aggressive Credit Policy (ν=100)
DSGE
Fig. 3. Responses to a Capital Quality Shock with Credit Policy.
intervention further moderates the decline. It does so by substantially moderating the rise in the spread. Doing so, however, requires that central bank lending increase to approximately 15 percent of the capital stock. Several other points are worth noting. First, in each instance the central bank exits from its balance sheet slowly over time. In the case of the moderate intervention the process takes roughly five years. It takes roughly three times longer in the case of the aggressive intervention. Exit is associated with private financial intermediaries re capitalizing. As private intermediaries build up their balance sheets, they are able to absorb assets off the central bank’s balance sheet. Second, despite the large increase in the central bank’s balance sheet in response to the crisis, inflation remains largely benign. The reduction in credit spreads induced by the policy provides sufficient stimulus to prevent a deflation, but not enough to ignite high inflation. Here it is important to keep in mind that the liabilities the central bank issues are government debt (financed by private assets), as opposed to unbacked high powered money.
3.2.3. Impact of the zero lower bound Next we turn to the issue of the zero lower bound on nominal interest rates. The steady state short term nominal interest rate is four hundred basis points. As Fig. 2 shows, in the baseline crisis experiment, the nominal rate drops more than 500 basis points, which clearly violates the zero lower bound on the nominal rate.12 In Fig. 4 we re create the crisis experiment, this time imposing the constraint that the net nominal rate cannot fall below zero. As the figure illustrates, with this restriction, the output decline is roughly 25 percent larger than in the case without. The limit on the ability to reduce the nominal rate to offset the contraction leads to an enhanced output decline. Associated with the magnified contraction is greater financial distress, mirrored by a larger movement in the spread. 12
For an early analysis of the implications of the zero lower bound for monetary policy, see Eggertsson and Woodford (2003).
M. Gertler, P. Karadi / Journal of Monetary Economics 58 (2011) 17–34
% Δ from ss
−5 −10 0 0
20 K
−20 0 0
20 N
−100 0
20 Quarters
40
0
Financial Accelerator
20 C
40
−5
5
20 L
0
5
20 π
0
20 Quarters
40
E [Rk]−R
5 0
0
20 I
40
20 Q
40
20 i
40
20 Quarters
40
0 −50 0
40
0 −5
50
40
0 −5
40
−50
0
−10 0
40
−10
−5
Annualized % Δ from ss
40 % Δ from ss
0
20 Y
% Δ from ss
% Δ from ss
% Δ from ss
0
0
% Δ from ss
−4
10
% Δ from ss
Annualized % Δ from ss
−2
R
5
Annualized % Δ from ss
ξ
0
Annualized % Δ from ss
% Δ from ss
30
20 0 −20 0 10 0 −10 0
Financial accelerator with the zero lower bound
Fig. 4. Impulse responses to the capital quality shock with and without the zero lower bound (ZLB).
In Fig. 5 we re consider the credit policy experiments, this time taking explicitly into account the zero lower bound restriction. As the figure makes clear, the relative gains from the credit policies are enhanced in this scenario. The reduced output contraction and the smaller drop in the inflation rate also shortens the period during which the interest rate policy is constrained by the zero lower bound restriction.
3.2.4. ‘‘News’’ as a source of asset price variation We introduce the capital quality shock to provide an exogenous source of variation in asset values. Mixed in with this shock, however, is variation in the effective quantity of physical capital. While in the current crisis there was ‘‘destruction’’ in asset values initiated by a contraction of housing prices, there was not effective destruction of physical capital. It is beyond the scope of this paper to incorporate housing and a boom bust cycle in either house prices or asset prices more generally. However, we can do a simple experiment that separates the effect of a contraction in asset values from the effect of an effective loss of physical capital. In particular, suppose the economy is hit by news that a capital shock is likely to hit the economy in the subsequent period with probability s. The expected size and duration of the shock (s times the realization of the shock in each of the subsequent periods) is the same in magnitude as the shock considered in the previous experiments. Suppose further that shock is never actually realized but that for a number of periods the private sector continues to believe it will arise with probability s. After a point it begins to revise down the likelihood of the shock. In this case, the news will reduce asset values, but because the shock is never realized, there is not a direct impact on the effect quantity of capital. Thus, we can disentangle the ‘‘asset value’’ effect from the physical quantity of capital effect. The experiment we consider proceeds as follows. We suppose the economy begins with the capital stock five percent above steady state (due perhaps to past ‘‘overoptimism’’ about the returns to investment). A wave of pessimism then sets in. For four straight quarters the private sector believes a capital quality shock will hit that is in expected value of the same magnitude as the autoregressive shock considered in the previous section. After the shock is not realized, the private sector then revises down the likelihood by a factor of 0.5 each period. Fig. 6 shows the results. The news shock triggers a financial crisis and a collapse in output much like the one following the capital quality shock studied in the previous section. Asset values collapse and the spread increases, which leads to the fall in output and investment. Overall, the collapse in output is nearly double what occurs in the frictionless benchmark. In contrast to the case of the realized capital quality shock, the news shock does not directly alter the stock of capital. Thus the crisis in this case is triggered purely by a loss in asset values.
20
40
Y %Δ from ss
0 −5 −10
0
20
40
K %Δ from ss
0 −10 −20
0
20
40
N
0 −50 −100
0
20
Annualized %Δ from ss
20
40
C
0 −5 0
20
40
0 −5
0
20
40
0 −5
0
20 Quarters
40
0
0
20
40
I
0 −50
0
20
40
Q
20 0 −20
π
5
5
50
L
5
E[Rk]−R
10
0
20
40
i
5 0 −5
0
20 Quarters
40
Financial Accelerator with ZLB Baseline Credit Policy with ZLB, ν=10 Aggressive Credit Policy with ZLB, ν=100
10 0
0
31
ψ
20 Percent
40
−5
−10
Annualized %Δ from ss
%Δ from ss
%Δ from ss
%Δ from ss
0
0
%Δ from ss
−4
%Δ from ss
−2
R
5
Annualized %Δ from ss
ξ
0
Annualized %Δ from ss
%Δ from ss
M. Gertler, P. Karadi / Journal of Monetary Economics 58 (2011) 17–34
0
20 Quarters
40
Fig. 5. Impulse responses to the capital quality shock with the zero lower bound (ZLB) with and without credit policy.
It is interesting to note that the employment drop is of the same magnitude as the output drop. As in the current crisis, labor productivity does not fall. At the same time, once it is realized that the shock will likely not happen, there is a fairly rapid bounce back of output and employment. In reality expectations were likely slower to adjust. We save a richer model of belief formation for subsequent research. 4. Optimal policy and welfare We now consider the welfare gains from central bank credit policy and also compute the optimal degree of intervention. We take as the objective the household’s utility function. We start with the crisis scenario of the previous section. We take as given the Taylor rule (without interest rate smoothing) for setting interest rates. This rule may be thought of as describing monetary policy in normal times. We suppose that it is credit policy that adjusts to the crisis. We then ask what is the optimal choice of the feedback parameter n in the wake of the capital quality shock. In doing the experiment, we take into account the efficiency costs of central bank intermediation, as measured by the parameter t. We consider a range of values for t. Following Faia and Monacelli (2007), we begin by writing the household utility function in recursive form:
Ot ¼ UðCt ,Lt Þ þ bEt Ot þ 1
ð40Þ
We then take a second order approximation of this function about the steady state. We next take a second order approximation of the whole model about the steady state and then use this approximation to express the objective as a second order function of the predetermined variables and shocks to the system. In doing this approximation, we take as given the policy parameters, including the feedback credit policy parameter n. We then search numerically for the value of n that optimizes Ot as a response to the capital quality shock. To compute the welfare gain from the optimal credit policy we also compute the value of Ot under no credit policy. We then take the difference in Ot in the two cases to find out how much welfare increases under the optimal credit policy. To convert to consumption equivalents, we ask how much the individuals consumption would have to increase each period in
32
M. Gertler, P. Karadi / Journal of Monetary Economics 58 (2011) 17–34
R
−2 −4 0
20
40
Annualized %Δ from ss
10 Annualized %Δ from ss
%Δ from ss
E[ξ] 0
0 −10
0
10 0 −10
0
20
20 Quarters
40
40
0 −10
Annualized %Δ from ss 0
20 L
10
40
0 −100
0
N
100
−2
40
K %Δ from ss
%Δ from ss %Δ from ss
20
%Δ from ss
−5
0
%Δ from ss
0
2
0
10
20 π
40 Annualized %Δ from ss
5
0
40
C %Δ from ss
%Δ from ss
Y
20
0 −10
0
20 Quarters
Financial Accelerator
40
E[Rk]−R
10 0 −10
40 20 0 −20 −40
0
20
40
I
0
20
40
Q
20 0 −20
0
20
40
i
20 0 −20
0
20 Quarters
40
DSGE
Fig. 6. Impulse responses to unrealized news shocks in a model with financial accelerator (FA) and without it (SDGE).
the no credit policy case to be indifferent with the case under the optimal credit policy. Because we are just analyzing a single crisis and not an on going sequence, we simply calculate the present value of consumption equivalent benefits and normalize it by one year’s steady state consumption. Put differently, we suppose the economy is hit with a crisis and then ask what are the consumption equivalent benefits from credit policy in moderating this single event. Since we are analyzing a single event, it makes sense to us to cumulate up the benefits instead of presenting them as an indefinite annuity flow, where most of the flow is received after the crisis is over. Finally, we abstract from considerations of the zero lower bound (due to the complications from computing the second order approximation of the model in this case.) In this regard, our results understate the net benefits from credit policy. Fig. 7 presents the results for a range of values of the efficiency cost t. In the baseline case with no efficiency cost (t ¼ 0), the benefit from credit policy of moderating the recession is worth roughly 8.50 percent of one years recession. At a value of 10 basis points, which is probably quite large for assets like agency backed mortgage securities and commercial paper, the efficiency gain is on the order of 7.0 percent of steady state consumption. At this value of efficiency costs, the optimal credit policy comes close to fully stabilizing the markup. The net benefits from the credit policy reduces below 1% of yearly steady state consumption when t reaches roughly 80 basis points. For high grade securities, however, this value for efficiency costs would be astronomical. Our analysis suggests that for reasonable values of efficiency costs (less than 10 basis points) the net gains from responding to the crisis with credit policy may be large.
5. Concluding remarks We developed a quantitative monetary DSGE model with financial intermediaries that face endogenously determined balance sheet constraints. We then used the model to evaluate the effects of expanding central bank credit intermediation to combat a simulated financial crisis. Within our framework the central bank is less efficient than private intermediaries at making loans. Its advantage is that it can elastically obtain funds by issuing riskless government debt. Unlike private intermediaries it is not balance sheet constrained. During a crisis, the balance sheet constraints on private intermediaries tighten, raising the net benefits from central bank intermediation. We find that the welfare benefits from this policy may be substantial during a crisis if the relative efficiency costs of central bank intermediation are within reason.
M. Gertler, P. Karadi / Journal of Monetary Economics 58 (2011) 17–34
33
10
Ω (%)
8 6 4 2 0 0
20
40
60 80 τ (basispoints)
100
120
140
0
20
40
60 80 τ (basispoints)
100
120
140
3000 2500
ν
2000 1500 1000 500 0
Fig. 7. One year consumption equivalent net welfare gains from optimal credit policy (O) and optimal credit policy coefficient (n) as a function of efficiency costs t.
Importantly, as we showed, in a financial crisis there are benefits credit policy even if the nominal interest has not reached the zero lower bound. In the event the zero lower bound constraint is binding, however, the net benefits from credit policy may be significantly enhanced. Conversely, as financial intermediaries re capitalize and the economy returns to normal, the net benefits from unconventional monetary policy diminish. Given this consideration and some considerations outside the model (e.g. the ‘‘politicization’’ of credit allocation in normal times), we interpret our analysis as suggesting that unconventional monetary policy should be used only in crisis situations. Our analysis focused on direct lending activities by the central bank. An alternative type of credit intervention in our model would be direct equity injections into financial intermediaries. Expanding equity in these institutions would of course expand the volume of assets that they intermediate. In our view, a key factor in choosing between equity injections and direct lending involves the relative efficiency cost of the policy action. For certain types of lending, e.g. securitized high grade assets such as mortgage backed securities or commercial paper, the costs of central bank intermediation might be relatively low. In this case, direct central bank intermediation might be highly justified. In other cases, e.g. C&I loans that require constant monitoring of borrowers, central bank intermediation may be highly inefficient. In this instance, capital injections may be the preferred route. By expanding our model to allow for asset heterogeneity, we can address this issue. Within our framework leverage plays a key role in the dynamics of the crisis. Leverage ratios are endogenous in the dynamics about the steady state, but the steady state leverage ratio is effectively determined exogenously. It would be interesting to endogenize the steady state leverage ratio and, in particular, try to account for what led the financial system to such a vulnerable state at the onset of the crisis. Undoubtedly, the long history of protection of large financial institutions (i.e., moral hazard stemming from too big to fail) has played a role. More generally, anticipation of government credit market interventions to dampen a crisis can lead private financial institutions to take on more leverage. By extending the analysis in this direction we can explore quantitatively how moral hazard considerations might factor into the analysis of government credit policies.13 Along these lines, it might also be interesting to think about capital requirements in this framework, following Lorenzoni (2008). Within our framework as within his, in making capital structure decisions, individual intermediaries do not account for the spillover effects of high leverage on the volatility of asset prices. Finally, we considered a one time crisis and evaluated the policy response. In subsequent work we plan to model the phenomenon as an infrequently occurring rare disaster, in the spirit of Barro (2009) and Gourio (2010). In this literature,
13
See Gertler et al. (2010) for an attempt along these lines.
34
M. Gertler, P. Karadi / Journal of Monetary Economics 58 (2011) 17–34
the disaster is taken as a purely exogenous event. Within our framework, the magnitude of the disaster is endogenous. We can, however, use the same tools as applied in this literature to compute welfare. References Acharya, V., Almeida, H., Campello, M., 2010. Aggregate risk and the choice between cash and lines of credit. NBER WP #16122. Adrian, T., Shin, H., 2009. Money, liquidity and monetary policy, American Economic Review. Barro, R., 2009. Rare disasters, asset prices and welfare costs, American Economic Review. Bernanke, B., Gertler, M., 1989. Agency costs, net worth and business fluctuations. American Economic Review. Bernanke, B., Gertler, M., Gilchrist, S., 1999. The financial accelerator in a quantitative business cycle framework. In: Taylor, J., Woodford, M. (Eds.), Handbook of Macroeconomics. Brunnermeier, M., Sannikov, Y., 2010. A macroeconomic model with a financial sector, Mimeo. Carlstrom, C., Fuerst, T., 1997. Agency costs, net worth and business fluctuations: a computable general equilibrium analysis. American Economic Review. Christiano, L., Eichenbaum. M., Evans. C., 2005. Nominal rigidities and the dynamics effects of a shock to monetary policy, Journal of Political Economy. Christiano, L., Motto, L., Rostagno, R., 2010. Financial factors in business cycles, Mimeo. Curdia, V., Woodford, M., 2010. Conventional and Unconventional Monetary Policy, Mimeo. Eggertsson, G., Woodford, M., 2003. The zero lower bound on interest rates and optimal monetary policy, Brookings Papers on Economic Activity. Faia, E., Monacelli, T., 2007. Optimal interest rate rules, asset prices and credit frictions, Journal of Economic Dynamics and Control. Gertler, M., Kiyotaki, N., 2010. Financial Intermediation and Credit Policy in Business Cycle Analysis, Mimeo. Gertler, M., Kiyotaki, N., Queralto, A., 2010. Financial crises, bank risk exposure and government financial policy, Mimeo. Gilchrist, S., Tankov, V., Zakrajsek, E., 2009. Credit market shocks and economic fluctuations: evidence from corporate bond and stock markets. Goodfriend, M., McCallum, B., 2007. Banking and interest rates in monetary policy analysis. Journal of Monetary Economics. Gourio, F., 2010. Disaster Risk and Business Cycles, NBER WP 15399. Holmstrom, B., Tirole, J., 1998. Private and public supply of liquidity. Journal of Political Economy. Iacovello, M., 2005. House prices, borrowing constraints and monetary policy in the business cycle. American Economic Review. Jermann, U., Quadrini, V., 2010. Macroeconomic effects of financial shocks, Mimeo. Justiniano, A., Primiceri, G., Tambalotti, A., 2010a. Investment shocks and business cycles, Journal of Monetary Economics. Justiniano, A., Primiceri, G,. Tambalotti, A., 2010b. Investment shocks and the relative price of investment, Review of Economic Dynamics. Kiyotaki, N., Moore, J., 2008. Liquidity, business cycles and monetary policy. Mimeo. Lorenzoni, G., 2008. Inefficient credit booms. Review of Economic Studies. Mendoza, E., 2010. Sudden stops, financial crises and leverage, American Economic Review, forthcoming. Merton, R.C., 1973. An intertemporal capital asset pricing model. Econometrica. Primiceri, G., Schaumburg, E., Tambalotti, A., 2006. Intertemporal disturbances. NBER WP 12243. Sargent, T.J., Wallace, N., 1981. The real bills doctrine versus the quantity theory of money. Journal of Political Economy. Smets, F., Wouters, R., 2007. Shocks and frictions in U.S. business cycles: a Bayesian DSGE approach. American Economic Review. Wallace, N., 1981. A Miller-Modigliani theorem for open market operations. American Economic Review. Woodford, M., 2003. Interest and Prices. Princeton University Press.
Further reading Brunnermeier, M., 2009. Deciphering the liquidity and credit crunch 2007–2008, Journal of Economic Perspectives. Kiyotaki, N., Moore, J., 2007. Credit cycles. Journal of Political Economy.
Journal of Monetary Economics 58 (2011) 35–38
Contents lists available at ScienceDirect
Journal of Monetary Economics journal homepage: www.elsevier.com/locate/jme
Discussion
Discussion of Gertler and Karadi: A model of unconventional monetary policy Harold Cole University of Pennsylvania, Department of Economics, Bunche Hall, 19104 Philadelphia, PA, United States
a r t i c l e in f o Article history: Received 22 October 2010 Accepted 25 October 2010 Available online 28 October 2010
Gertler and Karadi have developed a very interesting quantitative model of unconventional monetary policy. Their model shows us some of the ingredients that can lead to unconventional monetary policy having an important quantitative role. In addition, the model provides a clear rationale for the policies that we have ended up adopting. In doing both of these things, they are helping to fill an important hole in the literature on monetary policy.1 The fact that the authors have done so in a fairly elegant model which builds nicely on some pre existing work (Mark’s in particular), and is computationally fairly tractable is to be commended. However, since I see my role here as poking around under the hood of their model, in what follows, I discuss what I see as the questionable aspects of the ingredients they use to get all of this. The model of Gertler and Karadi has three key features which operate to make financial variables, and the net worth of the intermediaries in particular, an important constraint on the economy’s level of production. First, each period the intermediate goods firms in effect refinance themselves from scratch in order to acquire the capital they need to produce. They do this despite the fact that the implicit return to funds within the firm will generally be higher than the interest rate. Second, the financial intermediaries which supply the funds to intermediate goods producers are constrained to start out small and die off randomly at a fast enough rate to ensure that the overall financial sector never gets large enough to relax the constraint (at least in the steady state). And third, financial intermediaries never seek to hedge their risks, across states of the world and hence shocks to their net worth go directly into their ability to intermediate. In what follows, I discuss these various assumptions in turn. I start this discussion with the goods producers and their role in the Gertler and Karadi model. The intermediate goods producers issue claims in period t which enable them to acquire capital to be used in period in the following period. Then, in period t + 1 they hire workers on a spot labor market and choose the utilization level of capital which in turn determines the extent to which it depreciates. Their left over capital is subject to a quality shock, which is really a form of stochastic depreciation. This leads to the return to capital equation (25) which determines Rkt + 1. To the extent that Rkt þ 1 4 Rt þ 1 , the gross real interest rate the intermediary pays for its funds, then the fact that the firm cannot use retained earning to finance investment is increasing the costs of the firm and lowering its expected profits. However, this device has the effect of making the behavior of the firm more sensitive to movements in this wedge. If instead some fraction of the firms were able to pile up
E-mail address:
[email protected] Kocherlakota (2009) develops a nonquantitative model of the housing bubble, in which a collapse in the price of land leads to shortage of collateral. Other recent quantitative models that emphasize the role of quantitative easing in monetary policy include Curdia and Woodford (this issue) and Del Negro et al. (2010). 1
0304-3932/$ - see front matter & 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jmoneco.2010.10.008
36
H. Cole / Journal of Monetary Economics 58 (2011) 35–38
enough internal resources so that they could self finance their investments, then the relevant cost of funds would be the economy wide interest rate Rt + 1 since this is the rate at which the future payout of capital would be discounted. If Gertler and Karadi’s model of firm financing was correct, then firm balance sheets should show large amounts of financial liabilities being offset by net worth and the value of their capital. However, this is not exactly what we see. The recent flow of funds accounts put the financial assets of the nonfarm nonfinancial corporate business sector at 13.5 trillion while their tangible assets were 13.9 trillion in 2008. Since the net worth of these corporations was 14 trillion, this implies that their financial assets were essentially equal to the value of their financial liabilities. These numbers are aggregates, and do not imply that no firms are financially constrained, however they strongly suggest that many firms are not. Chari and Kehoe (2009) and Ohanian (2010) utilize Compustat data to examine the extent of cash reserves relative to investment at the firm level. They estimate that on average firms self finance 80% of their investment. Moreover, Bates et al. (2007) document that there has a been a large secular increase in firm cash holdings since 1980, and that once these cash holdings are taken account of, net debt to equity measures have fallen sharply. My skepticism about extent to which nonfinancial firms are cash constrained is further strengthened by observations from the great depression. The most important of which concerns the behavior of firms and in particular the extent to which they paid out dividends even during this extreme downturn. Cole and Ohanian (2000) developed this line of argument. Back then, the situation was reversed, and Mark was our discussant, and he expressed his reservations about our argument in his discussion (see Gertler, 2000). When we compare our standard business cycle models to the data, one of the consistent findings is that our models have trouble properly accounting for the variations in labor (taking productivity as given), but do relatively well in accounting for investment. Several researchers have considered adding tax style wedges to enable the standard real business cycle model to fit the U.S. business cycle data. They have consistently found a large role for the labor wedge and a small role for the capital wedge (see Mulligan, 2002; Chari et al., 2007). Surprisingly Ohanian (2010) shows that even during the current Great Recession, the role of the capital wedge remains small and while role of the labor wedge is even larger than normal. This finding by Ohanian is consistent with a slew of newspaper reports of firms piling up cash, or borrowing cheap in order to expand their cash holdings. Some of these articles are quite recent, but others date back to 2008.2 The Gertler and Karadi model does have a labor wedge coming through its sticky price assumption, but the impact of the financial friction they introduce is to generate a capital wedge in their model relative to its perfect capital markets version. These wedge results, while not uncontroversial, do raise questions about the true extent of the role played by financial frictions. However, their model also implies that a very large nonconventional intervention by Fed would reduce the wedge, so one might interpret the recent findings in a more positive light. Additionally, Gertler and Karadi might want to consider adding in an assumption that firms must borrow in order to finance some fraction of their labor as well as their capital. If the firm had to finance the fraction t of their labor bill. In this case, the first order condition for labor choice by the intermediate goods producers (24) would become Pmt ð1 aÞ
Yt ¼ Wt ð½1 þ tðRkt Rt Þ: Lt
Now increases in the financial friction will generate a labor wedge too through increases in the return gap Rkt Rt. I turn next in my discussion to the financial sector. Financial intermediaries in the model have a simple life. A household member at some t gets chosen to be an intermediary and is given some start up funding which is a fixed fraction of the funds coming from the closing of other financial intermediaries owned by the household. Assuming that the incentive constraint (9) binds in every period, then the internal rate of return exceeds the external cost of funds and the intermediary optimally chooses to borrow up to the constraint and pile up as much net worth as possible before the random termination date. At each point in time, the value of the intermediary, Vjt, is determined by the level of net worth. Gertler and Karadi show that the linearity of their problem implies equation (10), or Vjt ¼ ut Qt Sjt þ Zt Njt , where ut and Zt are time varying factors, Qt is the price of a claim to capital, Sjt is the quantity of claims to capital held by intermediary j, and Njt is its net worth. The intermediary can steal a fraction of these claims and this leads to the incentive constraint (13) which I will assume binds and leads to Qs Sjt ¼ ft Njt , where ft is the private leverage ratio. Taken together, these expressions imply that Vjt ¼ ½ut ft þ Zt Njt :
ðAÞ
If we make use of (13) in the transition equation for net worth (7),we get Njt þ 1 ¼ ðRkt þ 1 Rt þ 1 Þft Njt þRt þ 1 Njt :
2 For example: ‘‘Unlike consumers, companies are piling up cash,’’ New York Times, March 4, 2008. ‘‘Cheap debt for corporations fails to spur economy,’’ New York Times, October 3, 2010.
ðBÞ
H. Cole / Journal of Monetary Economics 58 (2011) 35–38
37
Gertler and Karadi assume that financial intermediaries are initially established at a suboptimally small size and cannot acquire new funds other than by borrowing. They also assumed that the financial intermediary could abscond with the fraction l of their funds and turn them over to their own household. Hence, other households face the same incentive issue as lenders if they sought to buy an equity stake in the financial intermediary. However, the family of the financial intermediary could invest in the firm without facing this incentive issue. For the family of the intermediary, the opportunity cost of these funds would be Rt + 1, however the state contingent return would be Rkt + 1, and to the extent that the financial constraint bound, the expected net payoff would be positive, or Et bLt,t þ 1 ðRkt þ 1 Rt Þ 4 0: The fact that this margin is positive implies that financial intermediaries are strictly on their incentive constraint and hence are investing all available funds. One way to investigate the reliance of these assumptions to the real world is to try and examine the extent to which they seem consistent with the data during the recent financial crisis. On the one hand, if we look at risk spreads and the volume of lending in certain markets, we do see a picture that appears to be consistent with the Gertler and Karadi model. Risk spreads opened up sharply in many bond markets, and the volume dropped, particularly in the asset backed commercial paper market. On the other hand, in the traditional intermediation sector, the banks, did not appear to have been constrained in their lending or investing activities. First, on the household side, the crisis did not lead household depositors to withdraw their funds from the banking system. Total deposits at U.S. banks actually grew by 5% in 2008 relative to 2007. One potential explanation for this is that we have deposit insurance in the real world and hence the increase in the risk associated with the bond market led people to invest in bank deposits. Second, the Federal Reserve’s statistics show that not only were excess reserves positive during 2008, the main year of the crisis, but these excess reserves were extremely large after Fed’s intervention. In September of 2008, the total seasonally adjusted reserves of the banking system shot up from 46 billion in the prior month to over 103 billion. They then rose to 315 billion in October, to 609 billion in November, and finally to 820 billion in December; a level that they stayed around or above throughout the next year. At same time required reserves only rose very slightly during this period. They went from 44 billion in August to 53 billion in December of 2008. Overall, one is struck by the extent to which some sort of financial market segmentation must be an important part of the story that Gertler and Karadi want to tell. However, a segmentation based story raises the question of how long the segmentation would last in the face of large interest rate differentials. If this segmentation broke down quickly, then one would expect only a fairly temporary impact from a fall in the net worth of the affected portion of the intermediation sector. While this segmentation seems to have lasted for a while during the current crisis, it may well be the case that the next time around it is more porous as financial agents learn how to overcome this segmentation in response to large profit opportunities. Another critical assumption by Gertler and Karadi is that financial intermediaries do not efficiently hedge their risk.3 To understand the role of this assumption, consider allowing intermediaries to enter into state contingent contracts which allow them to transfer net worth across states of the world tomorrow. For now, assume that these contracts are enforceable but cannot serve as a source of new loans. Let ot þ 1 denote the quantity of these state contingent claims (where we have followed Gertler and Karadi’s notation and repressed the representation of the state variable. We will require that these state contingent contracts break even in a present value sense to ensure that they do not result in the injection of additional funds into the intermediary: Et bfLt,t þ 1 ot þ 1 g ¼ 0:
ðCÞ
Then, note that we can combine Eqs. (A) and (B) to form the expected value of an intermediary today as Et bfLt,t þ 1 ½ut þ 1 ft þ 1 þ Zt þ 1 ½ðRkt þ 1 Rt þ 1 Þft Njt þ Rt þ 1 Njt þ ot þ 1 g:
ðDÞ
The intermediaries problem with respect to the optimal choice of ot þ 1 is to maximize (D) subject to (C). One can trivially see that this will involve positive transfers in states in which the return to net worth, ½ut þ 1 ft þ 1 þ Zt þ 1 , is high, and the reverse when low. Note also that the fact that the intermediaries are risk neutral will mean that their optimal allocation of net worth will be bang bang; i.e. put all of it at the highest payout state. The addition of these contracts to their model would change the distribution of net worth across states and prevent the ability of shocks to create states with a relative shortage of net worth. My conjecture is that this could severely dampen the impact of shocks in their model and bring the impact closer to the perfect capital markets model. I suspect this would be true even though there could be an overall shortage of net worth, since much of the impact of a negative shock probably comes from creating relative shortages of net worth. One might be concerned that these insurance contracts would lead to violations of the intermediaries incentive constraint. In this case, an alternative way to implement the efficient outcome would be to have the goods producing firms themselves transform their payout per unit of capital from Rkt þ 1 to something like Rkt þ 1 þ ot þ 1 (where we might have to worry about the 3 This assumption of a lack of efficient risk hedging is fairly pervasive in the literature on endogenous financial frictions. It shows up, for example, in Bernanke et al. (1999) and Carlstrom and Fuerst (1997), two of the classic papers in this literature.
38
H. Cole / Journal of Monetary Economics 58 (2011) 35–38
scaling of ot þ 1 relative to Rkt þ 1 in order to replicate the efficient insurance contract). The firms could do this by changing the state contingency of wages Wt + 1 and payouts to capital Rkt + 1 to yield the desired result. One way to defend the assumptions of Gertler and Karadi would be to assume that there are very rare shocks which create large relative shortages of net worth, and that the rarity of these shocks might lead to their not being insured against. The current financial crisis might well fit that mold. Of course one might also fear that the response of the Fed along with the TARP bailout would create a moral hazard problem going forward. Finally, let me note that one aspect of the modeling bothered me. The friction on financial intermediaries that in this paper comes from a fear that they will steal. However, I would argue that the real concern with financial intermediaries is that they will take on excessive risk because of the skewed nature of their compensation. Substantively, a change in the friction from stealing to gambling may not change the results very much. But, it may make it seem less plausible that the government has a real advantage in recovery funds from defaulting intermediaries if they have gambled them away. Since this recovery advantage is important to generate the positive impact of nonconventional monetary policy, rethinking the friction to make it more realistic may reduce the ability of the government to stimulate the economy. Overall, this is a very nice contribution to an important new area in monetary policy. It sets out a clear and tractable quantitative model that makes the case that quantitative easing is important. I suspect that it will have considerable impact on our thinking going forward. References Bates, T., Kahle, K. Stulz, R., 2007. Why do U.S. firms hold so much more cash than they use to? Dice Center for Research in Financial Economics Working paper. Bernanke, B., Gertler, M., Gilchrist, S., 1999. The financial accelerator in a quantitative business cycle framework. In: Taylor, J., Woodford, M. (Eds.), Handbook of Macroeconomics, vol. 1. Elsevier Science BV. Chari, V.V., Kehoe, P., Slide presentation, Federal Reserve Bank of St. Louis Policy Form, 2009. Chari, V.V., Kehoe, P., McGrattan, E., 2007. Business cycle accounting. Econometrica 75 (3), 781–836. Carlstrom, C., Fuerst, T., 1997. Agency costs, net worth, and business fluctuations: a computational general equilibrium analysis. American Economic Review 87 (5), 893–910. Cole, H. Ohanian, L., 2000. Re-examining the contributions of money and banking shocks to the U.S. great depression. In: Bernanke, B., Rogoff, K. (Eds.), NBER Macro Annual 2000, National Bureau of Economic Research, pp. 183–226. Curdia, V., Woodford, M., this issue. The central-bank balance sheet as an instrument of monetary policy, doi:10.1016/j.jmoneco.2010.09.011. Del Negro, M., Eggertsson, G., Ferrero, A., Kiyotaki, N., 2010. A quantitative evaluation of the Fed’s non-standard policies. CEPR Working paper. Gertler, M., 2000. Comment. In: Bernanke, B., Rogoff, K. (Eds.), NBER Macro Annual 2000, National Bureau of Economic Research, pp. 237–258. Kocherlakota, N., 2009. Bursting bubbles: consequences and cures. Research Memo. Mulligan, C., 2002. A century of labor-leisure. NBER Working paper 8774. Ohanian, L., 2010. The economic crisis from a neoclassical perspective. Journal of Economic Perspectives 24 (4), 1–24.
Journal of Monetary Economics 58 (2011) 39–48
Contents lists available at ScienceDirect
Journal of Monetary Economics journal homepage: www.elsevier.com/locate/jme
Politics and the Fed Allan H. Meltzer a,b,n a b
Carnegie Mellon University, Tepper School of Business, Pittsburgh, PA 15213, USA The American Enterprise Institute, USA
a r t i c l e in fo
abstract
Article history: Received 10 May 2010 Received in revised form 10 September 2010 Accepted 17 September 2010 Available online 7 October 2010
In the standard policy model, a policymaker optimizes the welfare of a representative agent. In practice, policies are chosen in a political process by agents elected by voters. Drawing on evidence from my two volume history of the Federal Reserve, the paper reports many examples of decisions influenced by political pressures. The history shows that the meaning of the independence of the Federal Reserve changed over time reflecting political influences. & 2010 Elsevier B.V. All rights reserved.
0. Introduction In the standard economic model of policy choice, choices are made by analyzing the social welfare implications to find welfare optimizing policy choices. In this paper, I argue that in a democratic government, and perhaps elsewhere, policy choices emerge from a political process. People vote; the voting process chooses elected representatives. With their appointees the representatives make policy choices. The appropriate model for policy analysis is a political economy model. This conclusion is reinforced by my 15 years of studying the history of the Federal Reserve. From the start, that study reflected the influence of James Buchanan and many others who followed him by analyzing policy choices made by individuals or committees motivated by considerations other than optimal welfare. Some of my prior work with Alex Cukierman and Scott Richard is fully in that tradition. Consider recent Federal Reserve policy. The Federal Reserve has more than doubled the size of its balance sheet. It has more than $1 trillion of excess reserves and more than $1 trillion of long term, relatively illiquid mortgage backed securities. No central bank in a developed country ever had so much long term debt. In 1932, a much less informed Federal Reserve Board refused a request from Congress to support the mortgage market. Congress responded by creating the Home Loan Banks, a fiscal solution. The current Board undertook the fiscal operations. This is a political decision. Who would claim it is optimal policy? In the late 1970s, Congress gave the Federal Reserve a dual mandate. Except for the years of ‘‘moderation’’, it has acted lexicographically. From 1979 to 1982, it pursued disinflation, letting unemployment rise to 10.8 percent. Before 1979, it concentrated on unemployment letting measured rates of inflation reach double digits. Currently, reducing unemployment is the main goal; inflation can wait. Who would claim that the optimal way to achieve a dual goal is always to work on one objective at a time? I cannot establish that the Federal Reserve follows inefficient procedures for political reasons, but I believe that the absence of a rule or quasi rule and reluctance to ask Congress to approve so called inflation targeting is best explained as a political decision. Policy and politics have a common Greek root.
n
Correspondence address: Carnegie Mellon University, Tepper School of Business, Pittsburgh, PA 15213, USA. Tel.: +1 412 268 3383. E-mail address:
[email protected]
0304-3932/$ - see front matter & 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jmoneco.2010.09.009
40
A.H. Meltzer / Journal of Monetary Economics 58 (2011) 39–48
Current politicized actions are not a rare event. Purchases, basically fiscal actions, are larger than in the past, but through most of its 96 year history, the Federal Reserve has often responded to political pressures. Federal Reserve actions affect economic outcomes. Voters hold elected officials, not Federal Reserve officials, responsible. The Federal Reserve is a major economic policy institution, but it is also a political institution. When Congress created the Federal Reserve in 1913, it gave no thought to optimal welfare policy. Its central concern was how the benefits of having a central bank would be distributed. Bankers wanted a bank modeled on the Bank of England; decisions would be made by bankers and they would benefit from having a place to rediscount loans and an opportunity to compete more effectively with London banks for the financing of agricultural exports. The principal opposition to the bankers came from farmers and others who feared that control by bankers would not be in their interest. They preferred government control. President Woodrow Wilson arranged a compromise. Congress made the 12 regional banks ‘‘semi autonomous.’’ The Board in Washington had a supervisory role. From the very start, the Board and the Reserve Banks differed over decision making. One of the earliest tussles came when the banks organized a Governors Conference in 1915 that began to meet at one of the banks to decide on policy and other matters. The Board decided that the banks had ‘‘assumed powers which they did not possess’’ (Meltzer, 2003, 80). They ordered the banks to meet in Washington no more than four times a year, to hold informal discussions, and it refused to approve spending for a staff. This was the first of many political disputes about governance. The Board won the early conflict, but the conflict continued. Prior to the Great Depression, Reserve Bank directors could approve or refuse to accept their share of securities purchases or sales. In the 1920s, the banks agreed to coordinate purchases and sales under the leadership of Benjamin Strong and the New York bank. They created an open market committee that the 1913 act did not authorize. The Board had no vote on open market purchases or sales. It used its supervisory authority to veto decisions it opposed. Several Board members disliked that arrangement. They found it too restrictive of their power. Soon after Benjamin Strong retired in 1928, the Board expanded the size of the open market committee. In 1935, Congress amended the Federal Reserve Act to authorize a Federal Open Market Committee (FOMC) and gave the Board a majority of votes. New York continued to carry out policy operations, but Board Chairman Eccles did not like the New York president, so until 1942 New York was not given a permanent seat on the FOMC. From 1935 to 1942, Boston declined rotation to permit New York to serve. The 1935 amendments to the 1913 act changed the formal locus of power in the system. The Board did not act to take control until the 1950s. In a series of decisions, Chairman Martin ended arrangements that allowed New York to dominate FOMC purchase and sale decisions. By 1955 at the latest, the Board had taken charge. Changes in the locus of power affect the choice of policies. It is a political act aimed at concentrating power at the Board. Soon after, members of Congress began to express greater interest in monetary policy. They pressed for greater transparency, audits of the FOMC, changes in the membership of bank boards to include labor, minorities, and women, Senate confirmation of Reserve Bank presidents, and changes in the power of Reserve Bank presidents. The system opposed many of the changes often calling on bankers to oppose changes. When Chairman Arthur Burns encouraged banker pressure on members of Congress, Congress reacted angrily. These differences often reflect the same division that President Wilson tried to settle, differences between bankers and regional interests versus national political interests represented in the Congress or usually the Chairs of the Banking Committees. Many of these issues and conflicts are present currently. Behind these disputes over power and influence, there are three main issues active at different times. First is the economic model or framework that the FOMC uses. The second, often related to the first, is the weights given to inflation, the exchange rate, and the unemployment rate in policy decisions. Third, Chairman Martin often said that the Congress often adopted a budget with a deficit. He believed that an independent Federal Reserve had to assist in financing the debt. It is hard to find evidence of optimal policy in the resolution of these issues. 1. Early political intervention The original Federal Reserve Act was based on the gold standard and the real bills doctrine. The latter authorized discounting of commercial paper used to finance agriculture, industry, and trade. That restriction ruled out purchases of government securities, direct finance of the Treasury, loans to brokers and dealers operating on exchanges, or mortgage securities. Real bills advocates claimed that by restricting the quality of credit they could prevent inflation. The flawed argument was that discounts of commercial paper increased when producers increased production and inventories. The increase was temporary, repaid when the producer sold the product, so prices would not rise. The Governor of the New York bank, Benjamin Strong, recognized the fallacy in the real bills doctrine.1 Restricting discounts to real bills restricted the portfolios of the Federal Reserve banks but not the end use that banks made of credit. Board members continued to insist on applying the real bills doctrine. The difference was central to the policy disputes between the Board and the banks in the 1920s and 1928 29. Board members would not approve increases 1 Until 1935, the operating head of each Reserve Bank had the title governor. After 1935, Board members became governors and heads of Reserve Banks became presidents.
A.H. Meltzer / Journal of Monetary Economics 58 (2011) 39–48
41
in the discount rate requested by several Reserve Banks, especially New York. Board members did not forget Congressional ire at the use of discount rate increases in the 1920 21 deflation. Instead, the Board sent a letter to the member banks urging them to prevent the use of credit for stock exchange speculation. The real bills doctrine was one reason for the Board’s action. Reluctance to raise the discount rate above 6 percent was a political reason that supplemented real bills. Financing World War I gave the system a political reason for circumventing the real bills doctrine. Instead of buying government securities as in World War II, the system ensured the success of four Treasury bond drives by offering short term loans at preferential discount rates that financed commercial bank purchases of Treasury certificates during the intervals between bond drives. During bond drives, the Federal Reserve adopted the so called ‘‘borrow and buy’’ policy. The policy permitted banks to defer payments for bond purchases up to a year from the time of purchase. Instead of raising market rates, the Board and the Treasury chose to subsidize the banks. Optimal or political? Many in the system understood that circumventing the real bills doctrine to finance government bond sales was inflationary. After the war, some but not all members wanted to eliminate the preferential discount rate that banks used to finance Treasury purchases. The Treasury did not agree until much later. In the classic pattern followed by debt managers, the Treasury wanted to keep discount rates low to prevent an increase in the rates it paid to lenders. A compromise with the Treasury permitted a rate increase after the Treasury sold certificates. This was not optimal for the buyers. And the Treasury did not accept an increase in the preferential rate for borrowers who purchased Treasury certificates. The Federal Reserve found itself unwilling to raise interest rates when faced with Treasury opposition. The Treasury’s preferential rate remained until the end of the Wilson presidency. The same problem reoccurred after World War II. One difference after the two wars was that in 1918 and 1919 the Treasury Secretary and the Comptroller of the Currency were Board members. The Board waited for a signal from the Treasury before notifying the banks about rate changes. The preferential rate remained until inflation and loss of gold threatened the gold reserve requirement. Experience following World War I shows that political influence had a powerful effect at the time. It was a prelude to many other periods when political influence prevented the system from taking appropriate action. The real bills doctrine was not the only source of disagreement between the Board and the Reserve banks. Congress had criticized the Board for increasing interest rates to 6 percent in 1921 and authorizing penalty interest rates on marginal increases in borrowing.2 The highest rate was 81.5 percent on a small loan at the Atlanta bank; it does not require deep knowledge of politics to recognize what politicians and the press did to an 81.5 percent rate. Maximum rates at St. Louis, Dallas, and Kansas City were 16, 7 and 22.5 percent respectively. The fact that the four Reserve Banks that introduced marginal rates were in agricultural districts reopened a recurrent political charge. Critics claimed that just as they feared the system worked to raise interest rates to farmers and merchants above the rates paid by banks and brokers in Wall Street. Congress responded by amending the Federal Reserve Act. An additional member to represent agriculture joined the Federal Reserve Board. There were now eight Board members including the two ex officio members. Deflation was the next major problem. Prices increased during the war and early postwar. Governor Strong and his British counterpart, Montagu Norman, wanted to restore the prewar gold standard. They decided that both countries had to deflate to the prewar price level, so they undertook deflationary policies. The wholesale price index shows that prices in the United States had increased in response to the prewar gold inflow and wartime monetary expansion. The deflation from 1920 to 1922 lowered the wholesale price index almost 40 percent to the prewar level.3 Strong knew that deflation would be costly. He expected unemployment to rise, but he (and Norman) regarded the cost as necessary to restore the gold standard. Deflation and the high interest rates in agricultural regions brought a Congressional inquiry. Strong succeeded in shifting the blame for interest rates onto the Treasury and making the case that deflation had to follow inflation. The lasting effect of the Congressional inquiry and criticism throughout the country was reluctance to increase the discount rate above six percent and a firm belief that the Federal Reserve could not follow Bank of England discount rate policy. To control credit, the system began to rely on open market purchases and sales. Instead of encouraging and discouraging borrowing by changing the discount rate, the system encouraged borrowing or forced repayment by discounting. It nurtured the belief that banks were reluctant to borrow, so it said credit contracted following increased borrowing because banks did not want to be in debt. This ‘‘needs and reluctance’’ explanation persisted for decades long after facts showed that banks responded to profit opportunities reflected in interest rates above the discount rate. The severe 1921 deflation brought short term real interest rates to 30 percent or more. Gold flowed in. Monetary base growth rose and recovery followed. The 1920 deflation was more severe than many others, but economic expansion always followed monetary expansion. The severe 1929 32 deflation is not an exception. It differed mainly because money growth declined much more than the rate of price change. Until President Roosevelt stopped the monetary decline by devaluing the gold dollar, the expectation of further deflation remained. After the dollar was devalued in January 1934, the political decision to devalue the dollar ended deflation.
2 Each member bank had a ‘‘normal’’ borrowing amount based mainly on size. Borrowing in excess of normal could be charged a higher, progressive rate. Only four districts, Atlanta, St. Louis, Dallas and Kansas City introduced progressive discount rates. 3 Other broad-based price indexes became available much later. They show a comparable reduction to pre-1917 levels.
42
A.H. Meltzer / Journal of Monetary Economics 58 (2011) 39–48
2. The federal reserve and the treasury, 1934–1951 During the long recovery from the Great Depression, the Federal Reserve either did nothing or remained subservient to the Treasury much of the time. Treasury Secretary Morgenthau wanted low interest rates to finance deficits. At times, he threatened to use the Exchange Stabilization Fund to conduct open market purchases. The Federal Reserve did not want the Treasury to take over its responsibilities, so they agreed to Morgenthau’s demands. The Federal Reserve’s decision to slow reserve growth in 1936 and 1937 was not an independent action. The Treasury made the decision to sterilize gold inflows urged on by President Roosevelt’s wish to reduce speculative gold inflows (Meltzer, 2003, 505). Sterilization worked like an open market sale of securities combined with a gold purchase. The Federal Reserve contributed increases in reserve requirement ratios. Before the second and third increases in reserve requirement ratios, the Board got Treasury agreement to the change. In January 1937, the Board met with Morgenthau. Although Morgenthau expressed reservations, he gave his approval to the second increase (Meltzer, 2003, 507 508). The Treasury also approved the decision to divide the remaining change in reserve ratios into two parts, one in February and one in May. Between the two increases, rates rose in the bond market. Although the rise was small, Morgenthau was very upset. He described the increase as a panic when rates reached 2.52 percent on March 13, 1937. He blamed the Federal Reserve and demanded that the Federal Reserve purchase enough to increase reserves and raise the bond rate to par at 2.50 percent. The Board then met in Morgenthau’s office. The Federal Reserve agreed to hold an FOMC meeting on March 15. The meeting agreed to purchase bonds but offset the purchases by selling bills. Eccles and Morgenthau wanted larger purchases, up to $250 million, a 10 percent increase in Federal Reserve holdings, but the FOMC, remembering it is supposed to be an independent agency, did not agree. Independence did not last. Morgenthau threatened to use the Exchange Stabilization Fund. At an evening meeting of the FOMC at Morgenthau’s home, Morgenthau warned that either the FOMC would act or the government would. The next day, April 4, the FOMC agreed to purchase $25 million at once and $250 million by May 1. Purchases started the next day. These were the first open market purchases since November 1933. The 1937 38 recession began soon after these discussions ended. Eccles urged increased government spending, but did not make any net purchases until 1938. Morgenthau and the Treasury wanted to end gold sterilization and reduce reserve requirement ratios. In February 1938, Roosevelt agreed to end gold sterilization. Eccles was, at best, cool to the idea. At the FOMC’s March 1 meeting, he favored continuing the Federal Reserve’s program of selling long term debt and purchasing bills. Eccles opposed proposals for net sales to reduce excess reserves, but did not urge net purchases. The Federal Reserve remained on the sidelines. The Treasury’s decision to end gold sterilization beginning in February 1938 expanded the monetary base and helped to end the recession. In April, after the recession ended, the Federal Reserve reduced reserve requirement ratios. The Board’s minutes refer to the change as part of the president’s program (Board of Governors of the Federal Reserve System, April 13, 1938, 1 2). Shortly after the United States entered World War II, on April 30, 1942, the Federal Reserve announced that it would purchase all 90 day Treasury bills offered at a rate no higher than 0.375 percent per annum. Initially it did not peg rates on other securities explicitly, but it maintained a pattern of rates throughout the war. The highest rate was 2.5 percent on bonds with initial maturity of 25 years. During the war and early postwar, the duration of debt at the largest banks declined, but the maximum yield remained at 2.5 percent. The Federal Reserve did not expect to maintain the wartime rate structure after the war. The Treasury agreed to increase some of the short term rates, but it wanted the 2.5 percent long rate to remain. Political concerns limited the Federal Reserve’s ability to respond to postwar inflation. At times, they complained to each other about the restriction, but they did not believe they had political support for an independent policy. The Federal Reserve shared the widespread concern among economists that the presence of a large, outstanding public debt limited the role of monetary policy. Increases in interest rates would reduce the value of outstanding debt. President Truman was not alone in his concern that a rise in interest rates would reproduce the severe deflation of 1920 21 that caused his haberdashery to fail. Chairman Eccles used this mix of political and economic argument to conclude that the Federal Reserve had to wait until the public supported higher interest rates. The Treasury’s position was more extreme than after World War I. Then, Treasury wanted no changes as long as it had to undertake large scale financing. After World War II, the budget had a surplus by spring 1946, so Treasury could retire debt. Nevertheless, it opposed any change in wartime interest rates. Two political events brought change. Senator Paul Douglas became chairman of the Joint Economic Committee. Douglas, a former distinguished economics professor at the University of Chicago, wanted the Federal Reserve to be more independent of the Treasury. Joined by Senator Ralph Flanders and Congressman Jesse Wolcott, Douglas gave the Federal Reserve an opportunity to speak publicly in favor of independence. Chairman Eccles had often said that the system could not act independently without Congressional support. The Douglas hearings showed that leaders in the Congress supported an end to pegged interest rates. Still, the Federal Reserve delayed. Testimony by Federal Reserve officials did not present a uniform view of the system’s role. Eccles saw the Federal Reserve as mainly a government agency regulating the financial industry and carrying out government policy. Alan Sproul, President of the New York bank, saw the Federal Reserve as mainly a financial institution, blending public and private control. Although much had changed since President Wilson’s 1913 compromise, the split between political and non political control remained.
A.H. Meltzer / Journal of Monetary Economics 58 (2011) 39–48
43
Testimony by Sproul and Eccles shows that both recognized a political constraint. Sproul said that the system had to balance the political concerns from Washington with financial concerns (Subcommittee on Monetary, Credit, and Fiscal Policies, 1950, 444 445).4 Eccles held a more political view of the Federal Reserve as an agent of Congress. ‘‘Congress appropriates the money; they levy the taxes; they determine whether or not there should be deficit financing. The Treasury is charged with the responsibility of raising whatever funds the government needs to meet its requirements. y I do not believe it is consistent to have an agent so independent that it can undertake, if it chooses, to defeat the financing of a large deficit, which is a policy of the Congress’’ (Subcommittee on Monetary, Credit, and Fiscal Policies, 1950, 231). The Korean War was the second event fostering policy change. The Federal Reserve succeeded in getting the Treasury to end the 2.5 percent interest rate ceiling. Shortly after the Korean War started late in June 1950, the FOMC met to discuss war finance. The war increased military spending and awakened fears of inflation. Instead of voting to raise interest rates, the FOMC drafted a letter to Secretary Snyder asking him to agree to an issue of 2.5 percent bonds ineligible for bank purchase. The committee hoped to reduce short term issues and raise the short term rate. Snyder’s reply emphasized the need to keep rates stable and the importance of leaving short term rates unchanged. The Board informed Snyder that it had approved an increase in discount rates to 1.75 percent, the first increase in 2 years. Snyder would not yield. He responded by announcing sale of a 13 month 1.75 percent note, clearly in conflict with the Board’s discount rate decision. The Treasury issue failed, so the Board bought $8 billion of the $13 billion refunded. Throughout the fall, the Federal Reserve repeatedly urged the Treasury to issue the 2.5 percent bond, and the Treasury refused. Allan Sproul urged the FOMC to increase the 1 year rate to 1.75 percent while holding the long rate at the 2.5 percent ceiling. Fed officials warned Snyder about rising inflation and the cost of inflation for government procurement. Snyder continued to oppose rate increases. Looking back on this period after a year, the Board wrote that it had started a credit restraint program. It failed because the ‘‘policy could not be followed far enough to make the discount rate effective’’ (quoted in Meltzer, 2003, 697). In other words, the Board was not willing to insist on an independent policy. Politics overrode anti inflation policy. Conflict between the Federal Reserve and the Treasury increased and became open during the fall of 1950. The issue reached a climax after Snyder and Federal Reserve Chairman McCabe met with President Truman. After the meeting, Snyder in a speech in New York claimed that the 2.5 percent ceiling would remain. His claim went far beyond McCabe’s statements to President Truman. The speech greatly strengthened the position of those like Sproul who wanted to end support for the interest rate peg. On January 31, the entire FOMC met with President Truman. The FOMC members met before the meeting but did not agree on a common position. They let Chairman McCabe speak for the group. Neither the president nor McCabe mentioned the dispute with the Treasury. Following the meeting, the White House released a statement saying that the Federal Reserve agreed to maintain stability of the government securities market. Treasury followed with a statement that interest rate levels would be maintained during the Korean War. The Treasury lied. The FOMC had not agreed to maintain the rate, and the president had not asked for a commitment. Most of the FOMC members were angry. One of the members had kept notes of the meeting with the president. Marriner Eccles released the notes to the press. The Sunday papers reported that the president and Secretary Snyder had lied. The conflict moved toward a political settlement known as the Accord that ended pegged interest rates. Four factors worked for the Federal Reserve’s benefit. First, Defense Secretary Wilson shared concern about inflation and the rising cost of fighting the war. Second, the financial press sided with the Federal Reserve and opposed the Treasury. Third, some Senators, led by Douglas, wanted more Federal Reserve independence. Fourth, inflation was rising. In December 1950, consumer prices rose at a 14 percent annual rate. In the subsequent negotiation between the Federal Reserve and the Treasury, the Treasury representatives accepted the Federal Reserve proposal. The Accord in March 1951 was a political agreement. No one mentioned optimal policy. The Accord did not grant unlimited independence. The Federal Reserve agreed to share responsibility for the orderly marketing of government debt, and it agreed to make no change in discount rates for a year without Treasury approval. The Federal Reserve also agreed to purchase up to $200 million of 2.5 percent bonds during the next refunding. The Treasury agreed to let short term rates rise and to permit an exchange of longer term debt at a rate above 2.5 percent. 3. The Martin years, 1951–1970 Once the chance of agreement rose, Thomas McCabe resigned as Chairman. President Truman appointed William McChesney Martin. Martin negotiated the Accord for the Treasury, so he was familiar with the central issue. Also, his father served as Governor of the St. Louis Reserve Bank for many years. Martin inherited an agency that had not pursued an independent policy since 1933. In the intervening years, the Federal Reserve had become the world’s principal monetary authority. After the war, Congress passed two major pieces of legislation that affected the Federal Reserve. The Employment Act of 1946 contained a vague mandate to direct 4 All the materials referenced in my volumes are available on the website of the Federal Reserve Bank of St. Louis. I am grateful to them for making all the materials available.
44
A.H. Meltzer / Journal of Monetary Economics 58 (2011) 39–48
government policy to achieving ‘‘maximum employment and purchasing power.’’ The 1944 Bretton Woods Agreement established a system of fixed but adjustable exchange based on the $35 dollar gold price. If anyone in government saw a potential for conflict between the two laws, I have not found it. In practice, Congress learned that the Federal Reserve could increase employment by reducing interest rates. The Federal Reserve gradually accepted responsibility for maintaining ‘‘full employment.’’ Until 1976, when Alan Greenspan was Chairman of the Council of Economic Advisers, the Federal Reserve and other agencies defined full employment as a 4 percent unemployment rate. They did not adjust for demographic or other changes. The Bretton Woods commitment to maintain the $35 gold price called for actions to prevent inflation. The Federal Reserve decided that the Treasury was responsible for international policy; its role was secondary or tertiary. Chairman Martin explained many times that the Federal Reserve was independent within the government, not independent of the government. Like Eccles he explained that the Federal Reserve could raise interest rates to prevent a boom in private investment spending, and he suggested that it was obligated to do so. But government budget deficits were different. Congress approved the budget and authorized the deficit. The Federal Reserve had to help finance the deficit. Many economists praised this conclusion. They wanted coordinated monetary and fiscal policies. They ignored politics, but politics was a dominant influence on Federal Reserve policy. The 1951 Accord reduced the Treasury’s influence over the Federal Reserve’s actions but did not eliminate it. The Federal Reserve agreed to help with Treasury financings. Assistance took the form of ‘‘even keel’’ policy. The Federal Reserve kept interest rate constant for a week or two before and after a sale of Treasury notes or bonds. Even keel did little harm during the Eisenhower and Kennedy administrations because, except in recession, the budget had a surplus or a relatively small deficit. Both presidents pursued budget policies that did not cause excessive monetary expansion. The Johnson administration followed a much more expansive fiscal policy with large deficits to finance the war in Vietnam and the Great Society. Chairman Martin’s policies produced inflation. He coordinated policy by financing a large part of the budget deficit with new money. By 1966, economic expansion had restored high employment. Annual base money growth rose to 6 percent or above, and the funds rate reached 4.9 percent in May 1966. After May, base growth declined but the funds rate continued to rise reaching 534 percent in November. Regulation Q ceilings restricted banks from raising time deposit rates. As open market rates rose, holders of thrift association certificates of deposit transferred their deposits to banks. The Board made a political decision to keep the ceiling rate fixed. Faced with declining deposit growth, the thrifts curtailed lending to home builders. Building activity declined by 40 percent in 1966. Bank loans went mostly to business customers, not to home builders. Further, banks sold their state and local government securities to service business customers. This raised rates on tax exempt debt angering an important political group. Thrifts and home builders are found in every Congressional district. They complained that they were being forced to bear the cost of financing the war and reducing the 4.8 percent inflation in early 1966. The thrifts were unwilling to raise deposit rates because they would have to pay the higher rate on all deposits. State and local governments complained to Congress about higher interest rates. After Congress held hearings, the Federal Reserve looked for ways to control inflation that did not require higher interest rates at thrifts and small banks. In response to Congressional criticism, the Board raised the reserve requirement ratio for time deposits from 4 to 5 percentage points for banks with more than $5 million of time deposits but not for smaller banks. And it wrote a letter to member banks restricting the use of discounts by banks financing the run off of certificates of deposit. With the funds rate at 5.75 percent, in July the Board rejected requests for a 5 percent discount rate from New York, Cleveland, Chicago and St. Louis. Governor Daane gave his reason for voting no. ‘‘Every economic ground said to increase the rate; his reluctance was because such an action would be harmful to relationships with the Administration’’ (Meltzer, 2010, Book 1, 505). A year later, Treasury Secretary Fowler and Council Chairman Ackley lobbied Board members Daane, Brimmer, Robertson, and Maisel not to increase the discount rate. They voted against the increase and it did not pass. Many in Congress disliked the rise in market interest rates. In August 1967 the House Banking Committee approved a bill putting a 4.5 percent ceiling on interest paid on time deposits under $100,000 and a 5.5 percent ceiling over that limit. Higher rates could be offered only if the president approved. The bill failed in the Senate, but the threat remained. Future Board actions could not ignore the political response to higher interest rates. Although Chairman Martin made frequent speeches about inflationary dangers, he left office at the end of January 1970 with the twelve month rate of increase in the consumer price index at 6 percent. Much higher inflation came later, but 6 percent was considered a high rate of peacetime inflation. Three political discussions dominated Federal Reserve policy during the Martin chairmanship. First, Chairman Martin and many others accepted the restricted meaning of independence that Martin often repeated and also the threat of Congressional interfering. Second, the Federal Reserve opposed securities auctions, and the Treasury did not begin auctions for notes and bonds until the 1970s. The Federal Reserve’s even keel periods came very frequently. The system supplied reserves to maintain interest rates and bought unsold bonds when an issue failed to attract buyers. The reserves remained on the Federal Reserve balance sheet permitting excessive money growth. Third, the Federal Reserve accepted academic arguments supporting policy coordination. To the administration, coordination meant that the Federal Reserve would keep
A.H. Meltzer / Journal of Monetary Economics 58 (2011) 39–48
45
interest rates from rising when budget deficits increased. Chairman Martin and many others at the Federal Reserve believed that coordination also required the administration to increase tax rates to reduce deficits instead of asking for Federal Reserve purchases. From 1966 to 1968, Chairman Martin repeatedly urged President Johnson to adopt a surtax. Johnson did not act, in part because Congress demanded spending reductions on Great Society programs. The president would not tell Martin or the public how much war spending and the budget deficit would increase. Martin learned from his own sources and again urged the president to ask for a surtax. He made a speech at Columbia University in June 1967 comparing the then current policy problem to 1929. Although he warned about the dangers of inflation, in September he voted against his colleagues on FOMC who urged an interest rate increase. Martin explained that he had worked for collaboration and did not want to oppose the president. Political concerns again dominated policy action. President Johnson often told Martin that he disliked ‘‘high’’ interest rates. Chairman Martin often told him that interest rates would be lower if Congress reduced the deficit by passing the surtax. In the spring the president agreed to reduce spending by $6 billion in fiscal 1969 as part of a bill to impose a ten percent surtax. Within days following passage, Council Chairman Arthur Okun warned the president about ‘‘fiscal overkill.’’ Okun and others tried to convert Martin’s claim that interest rates would decline once the surtax passed into a decision to lower the funds rate. The administration and the Board’s staff forecast a marked slowing in the economy. The Board’s staff report to the FOMC in July urged lower interest rates to prevent slow growth. FOMC members who favored policy coordination urged action to offset some of the contractive effects of the surtax. President Galusha (Minneapolis) thought it would be ‘‘unwise y both economically and (perhaps more importantly) politically to delay easing’’ (quoted in Meltzer, 2010, Book 1, 544). In August, the Board sent Governor Daane to the Minneapolis Board meeting to urge a reduction in the discount rate from 5.5 to 5 percent. The directors voted for 5.25 percent. Directors at Philadelphia and New York would not go along. Some Philadelphia directors said ‘‘they would be willing to countenance a certain degree of economic downturn in order to bring inflation under control’’ (Meltzer, 2010, 545). This was heresy for a majority of the Board. On a 4 to 3 vote, the Board approved the 5.25 percent rate. Martin asked for, and got, a unanimous vote for the change. Several banks refused to follow. Despite rising inflation and little evidence of slower growth, Okun tried to pressure the system to lower rates. He suggested the president appoint a commission to recommend changes in the Federal Reserve Act. In November, the commission reported. It recommended excluding the Reserve Banks from the FOMC. That would remove the strong proponents of anti inflation policy. The FOMC did not reverse policy until after the November election. Faced with market rates of 6.25 percent and 4.6 percent reported inflation, the Board approved a 5.5 percent discount rate, restoring the earlier 0.25 percentage point reduction. The forecasts of ‘‘fiscal overkill’’ were wrong. Okun later acknowledged the error. Asked about coordinated monetary policy, Okun replied: ‘‘If anything, we were pushing them harder for more easing, and we were off’’ (Hargrove and Morley, 1984, 307). Some of the Reserve Bank presidents resisted the pressure, reducing the Board’s ability to respond to political pressure. Other political interventions in Federal Reserve policy affected minor as well as major decisions. In 1969, some members of Congress wanted the Federal Reserve to assist housing by buying agency securities issued to finance government mortgage purchases. The FOMC was reluctant but feared that Congress would mandate action. Chairman Martin said he was ‘‘saddened by the fact that the System was involved in a political matter whether it liked it or not’’ (quoted in Meltzer, 2010, Book 1, n. 133). Purchases were substantially smaller than the more than $1 trillion of mortgage backed securities bought by the Federal Reserve in the year to January 2010. Two differences are that Chairman Martin limited the purchases and recognized that central banks in developed countries did not load their balance sheet with illiquid long term debt. The current Federal Reserve is more influenced by political pressures.
4. The Burns years In February 1970, Arthur Burns replaced Martin as Board chairman. At his swearing in ceremony, President Nixon acknowledged Federal Reserve independence but added that he hoped they would independently decide to support him. The president left no doubt that he expected the system to do as he wished. Burns met with the president frequently. Early in his tenure at the Board, the economy was in the 1969 70 recession. The recession did not eliminate inflation. Burns concluded that economic laws had changed. He blamed monopoly unions, oligopolies, and the welfare state. In my history, I discuss these errors. One of the main errors was neglect of expectations. Markets learned that the Federal Reserve responded quickly and decisively to unemployment but slowly and hesitantly to inflation. That produced the phenomenon known as stagflation. Burns and many others said, economics failed because inflation declined very little during recessions. In the 1960s, Burns criticized the use of wage price guideposts. In office, he became the leading proponent. Eventually, President Nixon imposed price and wage controls for political reasons. The Democrats in Congress authorized him to use controls. He believed they would criticize his failure to use them in the 1972 election. Burns participated in the meeting at Camp David in August 1971 at which the president prepared for a 90 day price and wage freeze and an end to
46
A.H. Meltzer / Journal of Monetary Economics 58 (2011) 39–48
convertibility of the dollar into gold. Why was the chairman of the independent Federal Reserve at a meeting to decide on economic policy to prepare for the 1972 election? Burns answered the question in a meeting with the president recorded on President Nixon’s tapes. Burns said: I am a dedicated man to serve the health and strength of our national economy, and I have done everything in my power, as I see it, to help you as president, your reputation and standing in American life and history. y I want you to know this y the moment a conflict arises, I’m going to be right here. I’ll tell you about it, and we’ll talk it out and try to decide together where to go next (quoted in Meltzer, 2010, Book 1, 635). During the fall of 1971, the president urged Burns repeatedly to increase money growth. Burns never said, yes I will. The closest he came was to tell the president that he would not repeat the Federal Reserve’s 1960 monetary restriction. Money growth rose in 1972. All the remaining members of the Board of Governors had been appointed by Presidents Kennedy and Johnson. They had no desire to help President Nixon’s re election. They voted for the policy to reduce the unemployment rate. Burns shared their concern about unemployment, and President Nixon said that no election was lost because of inflation. He said repeatedly that he lost the close 1960 election to Kennedy because unemployment rose in October 1960. Federal Reserve actions in 1972 helped to lower the unemployment rate before the election. Burns acted for political reasons and from a desire to lower the unemployment rate for reasons he believed were consistent with increased economic welfare. After President Carter did not reappoint Burns, Burns gave the Per Jacobson lecture at the 1979 IMF meeting in Belgrade. His speech, ‘‘The Agony of a Central Banker’’ contained this explanation of his actions as chairman and a message for economists who disregard political influence on policy decisions: Viewed in the abstract, the Federal Reserve had the power to abort the inflation at its incipient stage fifteen years ago [1964] or at any later point, and it has the power to end it today [1979]. At any time within that period, it could have restricted the money supply and created sufficient strains in financial and industrial markets to terminate inflation with little delay. It did not do so because the Federal Reserve was itself caught up in the philosophic and political currents that were transforming American life and culture (quoted in Meltzer, 2010, Book 1, 676. Emphasis added). Volcker left the Belgrade meeting to begin the policy change that would lower inflation by 1982. He had made the decision to let the market determine the funds rate, and he explained the decision to two administration officials on the trip to Belgrade. The political constraints that Burns emphasized had much less force for two reasons. First, Volcker believed inflation was costly. He rejected the Phillips curve model because inflation and the unemployment rate rose together, contrary to the model. From his experience on the FOMC he knew that it would be difficult, probably impossible to raise the funds rate high enough soon enough by FOMC action. Second, for the first time in the 1970s, the public told opinion pollsters that inflation was the most important domestic problem. President Carter and key members of the two banking committees supported the policy change. The president and many members of Congress faced the 1980 election in which inflation was certain to be a major issue. After interest rates rose, labor union leaders and some members of Congress urged President Carter to use credit controls instead of higher interest rates. Existing law gave the president authority to request the Federal Reserve to control credit. Most of the president’s economic advisers opposed his decision. The decision was made for political reasons, to provide an alternative to high interest rates (Biven, 2002, 247). In a televised address to the nation, he announced spending reductions and called on the Federal Reserve to control credit. Volcker had participated in the meetings leading up to the president’s announcement. He did not want to use credit controls but he, and all but one Board member, voted to accede to the president’s request. The governors decided to make a modest effort. The public’s response surprised everyone. Although the new regulations did not restrict the use of credit cards, people cut their cards and mailed them to the administration and the Federal Reserve. Businesses cancelled loans. In the second quarter, aggregate demand fell more than in any quarter up to that time. The public signaled they wanted lower inflation. They sent the same signal in the fall by electing Ronald Reagan who promised to reduce inflation and restore economic growth. Credit controls ended in the summer and money growth rose. The Federal Reserve had to restart its anti inflation policy in the fall of 1980. To his credit, President Carter did not interfere again. After January 1981, President Reagan supported the anti inflation policy. History provides many additional instances of political influence on Federal Reserve policy. Recent history is perhaps two well known to require comment. The Federal Reserve has undertaken fiscal actions, lending to AIG, investment banks and others. Perhaps one example will suffice. Interested economists can find many examples in my Federal Reserve history books. In 1932, Congressional leaders urged the Federal Reserve Board to assist housing and the mortgage market, then as now in dire straits. The Board declined, saying that housing was not its responsibility. Congress responded with a fiscal solution by authorizing the Federal Home Loan Banks. In contrast, the current Board has undertaken large fiscal policy actions. On January 21, 2010, 42 percent of the Federal Reserve’s assets were for mortgage backed securities guaranteed by Fannie Mae and other government housing lenders. Nearly 4 percent more consisted of loans to AIG or special Maiden Lane facilities.
A.H. Meltzer / Journal of Monetary Economics 58 (2011) 39–48
47
The mortgage backed securities had an average maturity in excess of 10 years. No central bank in developed countries has held so much illiquid long term debt. I have not mentioned international policy. Abandoning the gold standard was a political decision. Moving responsibility for international economic policy from the New York bank to the Board was a political decision, Senator Carter Glass’s punishment for promising a loan to the Bank of England after England returned to the gold standard was also a political decision. And the mistaken and unsuccessful policy to produce SDRs instead of revaluing exchange rates was a political decision.
5. Conclusion Federal Reserve governors are appointed to 14 year non renewable terms to reduce political influence. The Federal Reserve is said to be independent to protect it from pressure to finance the government’s borrowing. Like many other regulations, protection is circumvented. The Federal Reserve lends very small amounts directly to the Treasury, but it ‘‘warehouses’’ loans for foreign exchange intervention and purchases Treasury debt in the open market. In the current crisis, it has undertaken substantial fiscal actions, unlike any actions in its more than 95 year history. Congress granted Federal Reserve independence under the gold standard. That imposed restrictions on Federal Reserve actions. Once the gold standard ended, those restrictions ended. Congress or an administration can get a compliant Federal Reserve to accommodate to political pressure. The historical examples cited here are a small part of a large sample. Pressures on the Federal Reserve in the postwar years have varied. Presidents Eisenhower, Ford, Reagan and Clinton did not interfere. President Kennedy repeatedly urged the Federal Reserve to ‘‘twist the yield curve’’ and wanted to trade some inflation for lower unemployment. Presidents Johnson and Nixon interfered most until the current administration. History does not offer evidence of the government seeking to optimize policy in the interests of consumer welfare. Much of my Federal Reserve history is about the Great Inflation, 1964 82, the Great Depression, 1929 41, or wartime finance. Periods of low inflation and low unemployment are rare. The years 1923 1929, 1953 63, and the recent great moderation exhaust the years of sustained economic growth with low inflation. Political intervention, analytic error, excessive concern for near term events, and forecast errors are main reasons. One can look through the records of FOMC meetings without finding any discussion by members of the longer term consequences of policy actions. Of course, Board staff forecasts include forecasts of longer term outcomes, but the Reserve banks use different models. There is no evidence of an effort to reconcile differences in the 73 years to 1986 covered in my history. Further, discussion of forecast differences are so rare as to be non existent. Perhaps this changed after 1986. I conjecture that excessive concern for the near term and neglect of the longer term reflects perceived political pressures on policy to respond to current events. I have not produced evidence for that conjecture, but I cannot think of another reason why the Federal Reserve acts that way repeatedly. Washington has many so called independent agencies. They are removed from politics by making fixed term appointments of members and allowing many decisions to be made without Congressional or administration interference. A small number, like the Federal Reserve finance operations without Congressional appropriations. Congressman Patman tried several times to make the Federal Reserve subject to a congressionally approved budget. In periods of stress and public discomfort, Congress, the administration or both offer the Federal Reserve and others increased scrutiny and direction. The Great Depression resolved the long dispute about the locus of power in the system by choosing the Board, more responsive to political pressure than the banks. The Great Inflation brought Humphrey Hawkins hearings, closer monitoring, legislation restructuring directors of Reserve banks, and the release of FOMC minutes. The current crisis has not produced legislation at this time, but there are active proposals. One of them would allow the Government Accounting Office to audit open market decisions. Others review proposals to reduce the role of the reserve banks and reduce or eliminate regulatory responsibility. These and many other changes show that the Federal Reserve officials face political constraints. The constraints can be reduced by agreeing on a rule, or quasi rule, with Congress. The Federal Reserve has the authority to create inflation or recession, but the public blames its elected officials for outcomes they do not like. Years ago I proposed that the Federal Reserve should announce the inflation and unemployment rate they expect to achieve in the medium term. If they failed to reach their stated goals, they could offer an explanation along with an offer to resign. Congress could accept the explanation or the resignation. Authority and responsibility would be brought closer together. In 1988, I met with officials at the Reserve Bank of New Zealand and made my proposal. They improved on it by setting an inflation target with the political authorities and offering an explanation for target misses along with an offer to resign. Other countries have followed. Policy targets are chosen with political officials in all these countries and others. That does not eliminate political influence; it constrains it. It is rare, even unusual, to find political pressures mentioned in FOMC records. But it would be no less surprising if they were totally absent from the minds of the members. The Board and the FOMC sit in political Washington. Presidents and members of Congress have to seek re election when unemployment, inflation, or financial fragility are prominent issues. The Federal Reserve cannot and does not ignore the pressures. Its best protection is a rule, or quasi rule, that Congress accepts.
48
A.H. Meltzer / Journal of Monetary Economics 58 (2011) 39–48
References Biven, W.Carl, 2002. In: Jimmy Carter’s Economy. University of North Carolina Press, Chapel Hill. Board of Governors of the Federal Reserve System (various dates). Minutes of the Federal Open Market Committee, Washington, Unpublished. Hargrove, Edwin C., Morley, Samuel A., 1984. In: The President and the Council of Economic Advisers, Interviews with CEA Chairmen. Westview Press, Boulder, CO. Meltzer, Allan H., 2003. A History of the Federal Reserve, vol. 1. University of Chicago Press, Chicago 1913–51. Meltzer, Allan H., 2010. In: A History of the Federal Reserve, vol. 2, Book 1. University of Chicago Press, Chicago 1951–1969. Subcommittee on Monetary, Credit, and Fiscal Policies, (1950). Monetary, credit and fiscal policies. Joint Committee on the Economic Report. 81st Congress, 2nd session (November–December). Government Printing Office, Washington.
Journal of Monetary Economics 58 (2011) 49–53
Contents lists available at ScienceDirect
Journal of Monetary Economics journal homepage: www.elsevier.com/locate/jme
Discussion
Comment on: ‘‘Politics and the Fed’’ by Allan H. Meltzer Gregory D. Hess Robert Day School of Economics and Finance, Claremont McKenna College, 500 E 9th St, Claremont, CA 91711, USA
Allan Meltzer’s paper is a thoughtful historical review that highlights key political clashes over the path of economic policy between the Federal Reserve and the legislative and executive branches of the U.S. government. This paper, based on the research and findings from Meltzer’s (2002, 2010a,b) seminal economic investigation of the Federal Reserve, delivers a powerful message: the Federal Reserve is a political institution, established and institutionally influenced from a political process. As such, Meltzer argues that you cannot understand monetary policy decisions unless you bring to bear the constant political considerations in play. The recent decisions by the Federal Reserve in responding to the recent financial crisis and the Great Recession can be no clearer reminder of the paper’s main thesis. Elegantly characterizing the influence of politics on central bank decision making, however, is non trivial. Meltzer identifies a recipe for these political disputes. These include, from the Federal Reserve’s perspective, its economic framework, its prioritization of inflation and employment outcomes, and its degree of economic and political interest in assisting the management of the Federal Government’s liabilities. Rather than re work Meltzer’s superb sleuthing, it is useful to provide additional structure, albeit heuristic, to place Meltzer’s thesis in context. There are four key questions that need to be answered about the political economy of monetary policy: Who are the players? What are their objectives and responsibilities? What are the constraints? Finally, does the resulting framework help us to better understand historical episodes and outcomes? There are four players involved in the political aspects of monetary policy, presuming each player to be single actor: the Federal Reserve, the Legislature, the Administration, and Voters. What are there objectives and responsibilities? Taken in reverse order, Voters elect the Legislature and the Administration, and they desire price stability and maximum employment the dual mandate. The Administration (hereafter, the President) wants to retain office, it approves or vetoes legislation from the Legislature, and nominates members to the Federal Reserve. The Legislature (hereafter, the Congress) also wants to retain office, and they must legislate fiscal initiatives as well as manage the institutional context of monetary policy.1 Finally, the Federal Reserve (hereafter, the Fed) desires to obtain its dual mandate of price stability and maximum employment, provide macroeconomic and financial stabilization for the economy, and maintain its independence from election pressures and partisan concerns. The players collectively face several constraints. First, the players face a government budget constraint which equates the expected present discounted value of government spending, determined by the President and Congress, to the expected present discounted value of labor and business tax receipts, again determined by the President and Congress, and through an expected inflation tax, financed by money creation by the Fed. Second, the players are constrained by the current state of macroeconomic understanding that is available at the time. Model mis specification is inherent to economic decision making, and policy economists must recognize this inevitable constraint faced by our profession. Meltzer’s paper clearly identifies undue attention to the Real Bills doctrine as one example, but other examples will be noted below. Finally, social currents that affect policy see Burns (1979) discussed below also constrain and influence decision making. Overall, considering the political economy of monetary policy through this framework captures salient characteristics of events that are pointed to in Meltzer’s paper. Let’s look more closely at the Great Inflation. As noted by Meltzer, Chairman
Tel.: + 1 909 607 3686.
E-mail address:
[email protected] Issues involving the central bank often require quick decision making, and hence this rapid ‘inside lag’ for decision making is one reason that the Legislature delegates to the Federal Reserve the conduct of monetary policy. 1
0304-3932/$ - see front matter & 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jmoneco.2010.11.001
50
G.D. Hess / Journal of Monetary Economics 58 (2011) 49–53
Burns believed that the Fed was unduly influenced by the political current of the Great Society, which encouraged aggregate demand expansion to lower the rate of unemployment and create more economic opportunity. In addition, reelection pressures on the President and Congress, and the policy prescriptions from Samuelson and Solow (1960) and others of an exploitable and inherently mis specified Phillips Curve relationship, placed further demands on the Fed to run expansionary policies. In his Per Jacobsson lecture, Burns (1979) articulates his anguish: My conclusion that it is illusory to expect central banks to put an end to the inflation that now afflicts the industrial democracies does not mean that central banks are incapable of stabilizing actions; it simply means that their practical capacity for curbing an inflation that is continually driven by political forces is very limited. p. 21. Chairman Paul Volcker, an attendee at the lecture by Burns (1979), used circumstance and his transformational leadership to shift the Federal Reserve’s stance and ability to effectively fight inflation. Fortunately, by this time, policy economists became aware of the model mis specification of confusing short run versus long run Phillips Curves, so that no longer did policymakers believe there was a long run tradeoff between lower unemployment and higher inflation. Moreover, Congress, the President and Voters had become sufficiently educated and motivated to make lowering inflation a priority. The result was a painful lesson and pair of recessions, but ultimately a low inflation environment helped to lay the foundation for the Great Moderation. One area that Meltzer’s paper does not directly address is how a political economy framework can be used to better understand the run up to the Great Recession. Again, the key ingredients are model mis specification, lexicographic preferences that demoted inflation concerns relative to enhanced levels of economic growth, and politics. Interestingly, the political economy aspects of the Great Recession played out differently than they did in the Great Inflation and the Great Depression. First, the model mis specification by economic policymakers in the current crisis was the de emphasis of banking, credit and financial markets in simple monetary policy models. Indeed, key contributions by Clarida et al. (1999) and Woodford (2003) formed the basis for work horse models for applied monetary economics where financial, intermediation and credit factors were absent. Second, the Fed’s policy preferences shifted in the aftermath of the decline of technology stocks at the turn of the millennium, so that fear of deflation became their primary policy objective and risk management strategy. This fear was fed in part by the persistent deflation and economic stagnation that Japan has faced since the beginning of the 1990’s.2 This undue concern about potential deflation was, in large part, responsible for the policy of an exceptionally low nominal federal funds rate, and negative real federal funds rate, that the Fed pursued from 2002 through 2005. Third, the philosophical and political influence that pervaded the decade of the 2000’s was the Great Homeownership Society. This ‘philosophic and political current’ of rising rates of homeownership were facilitated and enhanced by the role of the GSE’s Fannie Mae and Freddie Mac. Clearly, members of the U.S. Congress and Senate were open to lobbying and contributions from these two organizations. They were also pleased to attend ribbon cutting ceremonies for first time home buyers. However, politics between the Legislature, the Administration and the Fed played out very differently during the Great Recession as compared to other historical episodes. While all were pleased by the robust housing market of the early 2000’s, the Administration and the Federal Reserve teamed up to take the strong political position that the GSE’s needed to be seriously reformed, their activities curtailed, and given sounder oversight and regulation. Chairman Greenspan (2004) was consistent on this point: The Federal Reserve is concerned about the growth and the scale of the GSEs’ mortgage portfolios, y which concentrate interest rate and prepayment risks at these two institutions. Unlike many well capitalized savings and loans and commercial banks, Fannie and Freddie have chosen not to manage that risk by holding greater capital. Instead, they have chosen heightened leverage, which raises interest rate risk but enables them to multiply the profitability of subsidized debt in direct proportion to their degree of leverage. Without the expectation of government support in a crisis, such leverage would not be possible without a significantly higher cost of debt. February 24, 2004. Other FOMC members were also vocal in their concerns about the GSE’s. In a question and answer article, President William Poole (2004) of the Federal Reserve Bank of St. Louis responded to a question as to what his worries were for the two mortgage companies. I believe that Fannie and Freddie expose the economy to a systemic risk [by borrowing short to lend long]. I don’t regard the probability [of them disrupting the financial system] as being very high, but I regard the probability as being high enough. ... I’m also concerned that, whatever may be the probability today, I would hate to see that probability rise. And I fear that it could because the capital positions [of Fannie and Freddie] are thin, and because these two firms have very aggressive growth objectives. Therefore, the problem is going to be larger going forward. July 5, 2004 2
See, for example, the speech by Governor Ben S. Bernanke (2002).
G.D. Hess / Journal of Monetary Economics 58 (2011) 49–53
51
The Administration was also vocal in its concerns about the GSE’s. Treasury Secretary John Snow (2005) repeated the Administration’s call for GSE reform, saying: Events that have transpired since I testified before this Committee in 2003 reinforce concerns over the systemic risks posed by the GSE’s and further highlight the need for real GSE reform to ensure that our housing finance system remains a strong and vibrant source of funding for expanding homeownership opportunities in America. Half measures will only exacerbate the risks to our financial system. April 13, 2005 The arrival of Ben Bernanke as Federal Reserve chairman in February of 2006 did not change the Federal Reserve’s position. As Shin (2006) reports: In his first appearance before Congress as Federal Reserve chairman, Ben S. Bernanke picked up where Alan Greenspan left off, criticizing a House bill that would strengthen oversight of Fannie Mae and Freddie Mac for not going far enough to rein in the companies’ investment portfolios. Bernanke said the House bill would not give a new regulator for the two companies ‘sufficiently strong guidance’ on how to regulate the portfolios, which he and others think have grown so large that they threaten the stability of the financial system. February 16, 2006 What political lessons were learned from the President and the Fed teaming up against Congress? First, the Fed was not influential with Congress outside its immediate areas of responsibility, even with the support of the Administration. Second, Congress originates legislation about the Fed, oversees the Fed’s institutional framework, and requires the Fed to provide regular financial and economic testimony not the other way around. Finally, lobbying matters. Fannie Mae and Freddie Mac were generous and welcome donors with many legislator on both sides of the aisle, and the Fed’s case against the GSE’s, even with the Adminstration’s full support, was certainly no match. From a political economy perspective, the above example as well as recent events have clarified that the Federal Reserve is overseen by the legislative branch of government, and that this relationship is subject to change by will of the Congress. Indeed, Congress has given the Federal Reserve the dual mandate of price stability and maximum employment, which is consistent with economic outcomes desired by voters. Is there empirical evidence that the Congress manages its oversight of the Fed in response to inflation and unemployment? Alternatively, Meltzer argues that there is evidence that Congress manages its oversight of the Fed when the level of government debt is high, or when interest rates are high. Following Kettl’s (1988) seminal contribution on Federal Reserve leadership, one can construct a proxy measure of Congressional interest in the Fed based on the number of House bills and resolutions introduced in a given year that pertain to the Fed. Indeed, we define BILLSt as the number of such bills and resolutions that contain the terms ‘Federal Reserve’ or ‘Federal Open Market Committee’ in a given year. Fig. 1 provides a graph of BILLSt from 1973 to 2009.3 Notice the jagged two year pattern of the data, reflecting the fact that a term of Congress lasts two years, and that more bills and resolutions are introduced in the first year of a Congressional term owing to the fact that bills in Congress take a substantial amount of time to move forward in order for them to potentially become law. Taking the Fed’s dual mandate at face value, one would consider annual measures of CPI inflation, pt , and unemployment, ut, as relevant explanatory factors in Congressional interest in the Fed.4 The regression estimate is equal to (robust standard errors in parentheses) BILLSt ¼ 13:914 27:179 EY t þ 3:163 pt þ3:982 ut ð9:405Þ ð3:839Þ
ð0:855Þ
ð1:523Þ
where EYt is a dummy variable that takes the value of 1 in even years when Congress has an election, and is 0 otherwise. Consistent with Fig. 1, Congressional interest as measured by BILLS drops significantly in election years since it is more difficult to move forward with legislation in the last year of a session. The equation has an R2 = 0.712, with 37 observations from 1973 to 2009.5 Interestingly, the coefficients on both inflation and unemployment are both statistically significant at below the 0.01 and 0.05 levels (denoted by nnn and nn), respectively, and the estimated coefficients are remarkably close in value. The F statistic for the test that the coefficients on inflation and unemployment are equal to 0.17, which has a p value of 0.679. These results lead us to estimate the following specification: BILLSt ¼ a EY t þ b MISERY t þ gXt þ et
ð1Þ
where the misery index ðMISERY t pt þut Þ is the simple sum of CPI inflation plus the unemployment rate, and we include additional explanatory variables, Xt, as alternative explanations for Congressional interest in the Fed. These additional explanatory variables include annual measures of the federal funds rate, the 10 year Treasury rate, the real federal funds rate, the slope of the yield curve, the government debt to nominal GDP ratio, the average maturity of government debt, the number 3 Data for BILLS were obtained from an electronic search of The Library of Congress’ Thomas (Jefferson) advanced bill summary and search for bills and resolutions actions introduced in the House from 1973 through 2009. Thus, we have 37 observations. The mean and standard deviation of BILLS are 39.57 and 21.37, respectively, with a high of 86 in 1975 and a minimum of 8 in 2006. 4 The macroeconomic and financial data are from the Economic Report of the President, 2010 Edition, Tables B1, B2, B42, B73, B88, and B96 and based on calculations from FDIC Table BF02. 5 Note that if we exclude inflation and unemployment as explanatory variables the R2 is equal to 0.41.
52
G.D. Hess / Journal of Monetary Economics 58 (2011) 49–53
80
BILLS
60 40 20 0 1970
1980
1990
2000
2010
YEAR Fig. 1. Congressional interest: 1973–2009.
Table 1 BILLSt a EY t þ b MISERY t þ gXt þ et . (1) Election Year Misery
(2) nnn
27.33 (3.857) 3.363nnn (0.628)
Federal funds rate
(3) nnn
27.32 (3.901) 3.437nnn (0.956) 0.101 (0.931)
Treasury 10 year rate
(4) nnn
27.28 (3.921) 3.503nnn (0.905)
(5) nnn
27.33 (3.914) 3.365nnn (0.634)
(6) nnn
27.32 (3.932) 3.343nnn (0.676)
(7) nnn
27.33 (3.903) 3.225nnn (0.828)
(8) nnn
27.13 (3.778) 2.365nn (0.875)
(9) nnn
27.63 (3.718) 3.317nnn (0.632)
0.255 (1.054)
Real federal funds rate
0.0982 (0.771)
Yield curve slope
0.109 (1.335)
Gov Debt to GDP ratio
8.603 (34.04)
Ave maturity of Gov Debt
4.381 (2.981) 0.027n (0.014)
FDIC failures
13.921n (8.056)
FDIC insured real failed assets Observations R2 p-value
27.32nnn (3.810) 3.435nnn (0.641)
37 0.711 0.679
37 0.711 0.659
37 0.711 0.657
37 0.711 0.704
37 0.711 0.460
37 0.712 0.543
37 0.737 0.715
37 0.733 0.559
37 0.728 0.958
Regressions include a constant. Robust standard errors in parentheses. nnn, nn, and n represent statistical significance at the 0.01, 0.05 and 0.1 levels, respectively. Data are from 1973 to 2009. p-value refers to the test that the coefficients on CPI-INFLATION and UNEMPLOYMENT are equal. BILLS refers to House Bills and Resolutions introduced that contain ‘‘Federal Reserve’’ or ‘‘Federal Open Market Committee’’. Data are from the Economic Report of the President, 2010 Edition, Tables B1, B2, B42, B73, B88, and B96 and based on calculations from FDIC Table BF02.
of FDIC insured institution failures (FDIC FAILURES), and the real value of the assets of the failed financial institutions that are FDIC insured.6 Based on Meltzer’s thesis, one would expect that there would be higher Congressional interest in the Fed, and hence more political pressure on the Fed, when interest rates are higher, or when the yield curve is flatter (that is, short rates are high relative to long rates and monetary policy is presumed to be tight). Moreover, based on Meltzer’s thesis, one would also expect that Congress pressures the Fed when the level of government debt is high, when the average maturity of Government debt is high, or when financial institutions are failing.7 Table 1 presents the regression results for expression (1). Column (1) presents the baseline regression results when Congressional election year effects and the misery index are included as explanatory factors. The row at the bottom of the
6 The real federal funds rate is calculated using urban CPI inflation, unemployment is for all civilian workers, the average maturity of government debt is measured in fractions of a year, the slope of the yield curve is the 10 year Treasury Bond rate less 1 year Treasury Bill rate, and the estimated real value of the assets of failed institutions that are FDIC insured is measured in trillions of 1982 dollars (FDIC INSURED REAL FAILED ASSETS). 7 When the average maturity structure of debt is high, one would surmise that the Government does not have to reissue debt as often and hence an inflation tax can more effectively shift the fiscal burden.
G.D. Hess / Journal of Monetary Economics 58 (2011) 49–53
53
40
BILLS
20 0 −20 −40 5
10
15
20
MISERY INDEX Fig. 2. Congressional interest and misery.
table labeled ‘p value’ is from an F test of the null hypothesis that the coefficients on unemployment and inflation are equal. As the results in the column indicate, MISERY is a very significant predictor of BILLSt. Indeed, Fig. 2 presents a cross plot of BILLS and MISERY, where election year effects and a constant are eliminated from the former. The figure demonstrates a consistent positive relationship between the two variables. The results reported in columns (2) (7) of Table 1 demonstrate a consistent negative result: that is, after controlling for Congressional election year affects and the misery index, interest rate and government debt measured are not significant explanatory factors of Congressional interest in the Fed. However, the final two columns provide limited and contradictory evidence on the contribution of bank failures to Congressional interest in the Fed. As demonstrated in column (8), an increase in the number of failed institutions that are FDIC insured actually leads to a significant decrease in Congressional interest in the Fed. By contrast, the results in column (9) suggest that the real value of the assets of failed FDIC insured institutions is significantly and positively related to Congressional interest in the Fed.8 Given the coefficient values, a trillion dollars (1982 basis) of failed assets in FDIC insured institutions is equivalent to a 4 percentage point rise in the misery index (4.03=13.92/3.45). Note, however, that the estimated coefficient on real failed assets in column (9) is extremely sensitive to the observation for 2009. What are the takeaways from this discussion? First, Meltzer’s contribution highlights that applied monetary economists neglect the role of politics in the Fed’s decision making at their own peril. Second, as the Fannie Mae and Freddie Mac episode discussed above suggests, the Fed is on the receiving end of political pressure, not on the giving end, even when it teams up with the Administration. Third, our empirical evidence does not provide a consistent paradigm for relating measures of Congressional interest in the Fed with interest rate and government debt measures. Of course, this simple test does not negate the historical examples pointed in Meltzer’s paper, particularly some aspects of the Fed’s handling of the recent financial crisis. Finally, Meltzer’s historical work and the empirical evidence demonstrated in this paper speak to the importance and primacy of the dual mandate to Congress and the Fed. It is unlikely, however, that there is a truly common understanding by the Fed and Congress of the dual mandate. From the Fed’s perspective, a monetary policy that delivers long run price stability, and that credibly insures against both deflation and inflation is the path for maximum employment. The Federal Reserve should use its leadership to solidify this understanding with Congress, and thereby limit the political pressure it will face to contort policy in the face of high levels of unemployment. References Bernanke, B., 2002. Deflation: Making Sure it Doesn’t Happen Here. Remarks Before the National Economics Club, Washington, DC. Burns, A., 1979. The Anguish of Central Banking. The 1979 Per Jacobsson Lecture, Belgrade, Yugoslavia, September 30. Clarida, R., Gali, J., Gertler, M., 1999. The science of monetary policy: a new Keynesian perspective. Journal of Economic Literature 37, 1661–1707. Greenspan, A., 2004. Testimony of Chairman Alan Greenspan Before the Committee on Banking, Housing, and Urban Affairs, U.S. Senate: Governmentsponsored Enterprises, February 24. Kettl, D., 1988. Leadership at the Fed. Yale University Press, New Haven. Meltzer, A., 2002. A History of the Federal Reserve, volume 1: 1913–1951. University of Chicago Press, Chicago. Meltzer, A., 2010a. A History of the Federal Reserve, volume 2: Book 1, 1952–1969. University of Chicago Press, Chicago. Meltzer, A., 2010b. A History of the Federal Reserve, volume 2: Book 2, 1970–1986. University of Chicago Press, Chicago. Poole, W., 2004. My Goal for Inflation is Zero. Business Week, July 5. Samuelson, P., Solow, R., 1960. Analytical aspects of anti-inflation policy. American Economic Review 50 (May), 177–184. Shin, A. 2006. Bernanke Criticizes Fannie, Freddie Bill. The Washington Post, February 16, D.03. Snow, John W., 2005. Testimony Before The U.S. House Financial Services Committee, April 13. Woodford, M., 2003. Interest and Prices: Foundations of a Theory of Monetary Policy. Princeton University Press, Princeton.
8
I thank Mark Gertler for this suggestion.
Journal of Monetary Economics 58 (2011) 54–79
Contents lists available at ScienceDirect
Journal of Monetary Economics journal homepage: www.elsevier.com/locate/jme
The central-bank balance sheet as an instrument of monetary policy Vasco Cu´rdia a,, Michael Woodford b,1 a b
Federal Reserve Bank of New York, 33 Liberty Street, New York, NY 10045, United States Columbia University, United States
a r t i c l e i n f o
abstract
Article history: Received 30 April 2010 Received in revised form 2 September 2010 Accepted 9 September 2010 Available online 16 October 2010
We extend a standard New Keynesian model to allow an analysis of ‘‘unconventional’’ dimensions of policy alongside traditional interest rate policy. We find that quantitative easing in the strict sense is likely to be ineffective, but that targeted asset purchases by a central bank can instead be effective when financial markets are sufficiently disrupted, and we discuss the conditions under which such interventions increase welfare. We also discuss optimal policy with regard to the payment of interest on reserves. Published by Elsevier B.V.
JEL classification: E50 E52 E58 Keywords: Credit policy Quantitative easing Zero lower bound Interest on reserves
1. Introduction The recent global financial crisis has confronted central banks with a number of questions beyond the scope of standard accounts of the theory of monetary policy. Monetary policy is ordinarily considered solely in terms of the choice of an operating target for a short term nominal interest rate, such as the federal funds rate in the case of the Federal Reserve. Yet during the recent crisis, other dimensions of policy have occupied much of the attention of central bankers. One is the question of the appropriate size of the central bank’s balance sheet. In fact, the Fed’s balance sheet has grown dramatically in size since the fall of 2008 (Figs. 1 and 2). As shown in Fig. 1, the component of the Fed’s liabilities constituted by reserves held by depository institutions has changed in an especially remarkable way: by the fall of 2008 reserves were more than 100 times larger than they had been only a few months earlier. This explosive growth has led some commentators to suggest that the main instrument of US monetary policy has changed, from an interest rate policy to one often described as ‘‘quantitative easing.’’ Does it make sense to regard the supply of bank reserves (or perhaps the monetary base) as an alternative or superior operating target for monetary policy? Does this (as some would argue) become the only important monetary policy decision once the overnight rate (the federal funds rate) has reached the zero lower bound, as it effectively has in the US since December 2008? And now that the Federal Reserve has legal authorization to pay interest on reserves (under the Emergency Economic Stabilization Act of 2008), how should this additional potential dimension of policy be used?
Corresponding author. Tel.: + 1 212 720 5994; fax: + 1 212 720 1844.
E-mail addresses:
[email protected],
[email protected] (V. Cu´rdia). Prepared for the 75th Carnegie-Rochester Conference on Public Policy, ‘‘The Future of Central Banking,’’ April 16–17, 2010.
1
0304-3932/$ - see front matter Published by Elsevier B.V. doi:10.1016/j.jmoneco.2010.09.011
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
56
evaluate the effects of alternative prescriptions for monetary policy have little to say about them. Many models used for monetary policy analysis both theoretical models used in normative discussions of ideal monetary policy commitments, and quantitative models used for numerical simulation of alternative policies abstract altogether from the central bank’s balance sheet, simply treating a short term nominal interest rate as if it were under the direct control of the monetary authorities, and analyzing how that interest rate should be adjusted.2 But such a framework rules out the kinds of questions that have recently preoccupied central bankers from the start. In this paper, we extend a basic New Keynesian model of the monetary transmission mechanism to explicitly include the central bank’s balance sheet as part of the model. In addition to making more explicit the ways in which a central bank is able to (indirectly) exert control over the policy rate, the extended model allows us to address questions about other dimensions of policy of the sort just posed. In order to make these questions non trivial, we also introduce non trivial heterogeneity in spending opportunities, rather than adopting the familiar device of the ‘‘representative household,’’ so that financial intermediation matters for the allocation of resources; we introduce imperfections in private financial intermediation, and the possibility of disruptions to the efficiency of intermediation, for reasons taken here as exogenous, so that we can examine how such disturbances affect the desirability of central bank credit policy; and we allow central bank liabilities to supply transactions services, so that they are not assumed to be perfect substitutes for privately issued financial instruments of similar maturity and with similar state contingent payoffs. Finally, we consider the conduct of policy both when the zero lower bound on the policy rate is not a binding constraint, and also when it is. Section 2 outlines the structure of our model, with primary attention to the way that we model financial intermediation and the policy choices available to the central bank. Section 3 then uses the model to discuss changes in the supply of bank reserves as a dimension of policy, and the related question of the rate of interest that should be paid on reserves. Section 4 then turns to the question of the optimal composition of the central bank’s asset portfolio, and Section 5 concludes.
2. A model with multiple dimensions of monetary policy Here we sketch the key elements of our model, which extends the model introduced in Cu´rdia and Woodford (2009) to introduce the additional dimensions of policy associated with the central bank’s balance sheet. (The reader is referred to our earlier paper, and especially its technical appendix, for more details.) 2.1. Heterogeneity and the allocative consequences of credit spreads Our model is a relatively simple generalization of the basic New Keynesian model used for the analysis of optimal monetary policy in sources such as Goodfriend and King (1997) and Woodford (2003). The model is still highly stylized in many respects; for example, we abstract from the distinction between the household and firm sectors of the economy, and instead treat all private expenditure as the expenditure of infinite lived household firms, and we similarly abstract from the consequences of investment spending for the evolution of the economy’s productive capacity, instead treating all private expenditure as if it were non durable consumer expenditure (yielding immediate utility, at a diminishing marginal rate). We begin by summarizing, in this section, the equilibrium relations in the present model that are the same as those in the simpler model of Cu´rdia and Woodford (2009), and that would also apply to a model with heterogeneity but no credit frictions. We then (in the next subsection) discuss in more detail the part of the model that is new here, namely the equilibrium relations involving intermediary behavior and the central bank balance sheet. We depart from the assumption of a representative household in the standard model, by supposing that households differ in their preferences. Each household i seeks to maximize a discounted intertemporal objective of the form " # Z 1 1 X bt utt ðiÞ ðct ðiÞ; xt Þ vtt ðiÞ ðht ðj; iÞ; xt Þ dj , ð1Þ E0 t
0
0
where tt ðiÞ 2 fb,sg indicates the household’s ‘‘type’’ in period t. The period utility from consumption is assumed to be of the form t
1
c1st ðC t Þst , 1 s1 t 1
ut ðc; xt Þ
ð2Þ
for t ¼ b,s, where for each type, st 4 0 would be the intertemporal elasticity of substitution of expenditure if one’s type t were expected to remain unchanged, and C t is an exogenous type specific disturbance, indicating variation in aggregate spending opportunities. As in the basic NK model, there is assumed to be a continuum of differentiated goods, each produced by a monopolistically competitive supplier, with a production technology for each good j of the form yt ðjÞ ¼ At ht ðjÞ1=f ,
2
This approach is developed in detail in Woodford (2003).
ð3Þ
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
57
where f 4 1 indicates the degree of diminishing returns, and At is an exogenous productivity factor, common to all goods. The index ct(i) is a Dixit Stiglitz aggregator of the household’s purchases of these differentiated goods, with elasticity of substitution y 4 1 between any two goods. The household similarly supplies a continuum of different types of specialized labor, indexed by j, that are hired by firms in different sectors of the economy. The disutility of work for each type of labor is of the form vt ðh; xt Þ
ct 1þ n
n
h1 þ n H t ,
ð4Þ
for t ¼ b,s, where n 4 0 is the inverse of the Frisch elasticity of labor supply for both types, and H t is an exogenous disturbance, also common to both types. Each agent’s type tt ðiÞ evolves as an independent two state Markov chain. Specifically, we assume that each period, with probability 1 d (for some 0 r d o 1) an event occurs which results in a new type for the household being drawn; otherwise it remains the same as in the previous period. When a new type is drawn, it is b with probability pb and s with probability ps , where 0 o pb , ps o 1, pb þ ps ¼ 1. (Hence the population fractions of the two types are constant at all times, and equal to pt for each type t.) We assume moreover that ubc ðc; xÞ 4 usc ðc; xÞ
ð5Þ
for all levels of expenditure c in the range that occur in equilibrium. Hence a change in a household’s type changes its relative impatience to consume; in addition, the current impatience of households to consume is changed by the t exogenous disturbances C t . The coexistence of the two types with differing impatience to consume creates a social function for financial intermediation. In the present model, as in the basic New Keynesian model, all output is consumed either by households or by the government; hence intermediation serves an allocative function only to the extent that there are reasons for the intertemporal marginal rates of substitution of households to differ in the absence of financial flows. The present model reduces to the standard representative household model in the case that one assumes that ub ðc; xÞ ¼ us ðc; xÞ and vb ðh; xÞ ¼ vs ðh; xÞ. We assume that most of the time, households are able to spend an amount different from their current income only by depositing funds with or borrowing from financial intermediaries, that the same nominal interest rate idt is available to all savers, and that a (possibly) different nominal interest ibt is available to all borrowers,3 independent of the quantities that a given household chooses to save or to borrow. For simplicity, we also assume that only one period riskless nominal contracts with the intermediary are possible for either savers or borrowers. The assumption that households cannot engage in financial contracting other than through the intermediary sector represents one of the key financial frictions.4 We also allow households to hold one period riskless nominal government debt, but since government debt and deposits with intermediaries are perfect substitutes as investments, they must pay the same interest rate idt in equilibrium, and the decision problem of the households is the same as if they have only a decision about how much to deposit with or borrow from the intermediaries. Aggregation is simplified by assuming that households are able to sign state contingent contracts with one another, through which they may insure one another against both aggregate risk and the idiosyncratic risk associated with a household’s random draw of its type, but that households are only intermittently able to receive transfers from the insurance agency; between the infrequent occasions when a household has access to the insurance agency, it can only save or borrow through the financial intermediary sector mentioned in the previous paragraph. The assumption that households are eventually able to make transfers to one another in accordance with an insurance contract signed earlier means that they continue to have identical expectations regarding their marginal utilities of income far enough in the future, regardless of their differing type histories. It then turns out that in equilibrium, the marginal utility of a given household at any point in time depends only on its type tt ðiÞ at that time; hence the entire distribution of marginal utilities of income at any time can be summarized by two b s state variables, lt and lt , indicating the marginal utilities of each of the two types. The expenditure level of type t is similarly the same for all households of that type, and can be obtained by inverting the marginal utility functions to yield an expenditure demand function ct ðl; xt Þ for each type. Aggregate demand Yt for the Dixit Stiglitz composite good can then be written as X pt ct ðltt ; xt Þ þGt þ Xt , ð6Þ Yt ¼ t
where Gt indicates the (exogenous) level of government purchases and Xt indicates resources consumed by intermediaries (the sum of two components, Xpt representing costs of the private intermediaries and Xcb t representing costs of central bank activities, each discussed further below). Thus the effects of financial conditions on aggregate demand can be summarized 3
Here ‘‘savers’’ and ‘‘borrowers’’ identify households according to whether they choose to save or borrow, and not by their ‘‘type’’. In our simple model, this is the only reason for the (possible) existence of a credit spread. One can, of course, alternatively motivate spreads as resulting from risk premia, even when all investors choose from among a spectrum of assets with the same state-contingent returns. But without market segmentation of the kind that we assume, the central bank’s balance sheet would play no role in equilibrium determination, owing to a ‘‘Modigliani– Miller’’ theorem of the kind first discussed by Wallace (1981). See Cu´rdia and Woodford (2010b, Section 1) for further discussion. 4
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
58
t
b
s
by tracking the evolution of the two state variables lt . The marginal utility ratio Ot lt =lt Z1 provides an important measure of the inefficiency of the allocation of expenditure owing to imperfect financial intermediation, since in the case of frictionless financial markets we would have Ot ¼ 1 at all times. In the presence of heterogeneity, instead of a single Euler equation each period, relating the path of the marginal utility of income of the representative household to ‘‘the’’ interest rate, we instead have two Euler equations each period, one for each of the two types, and each involving a different interest rate ibt in the case of the Euler equation for type b (who choose to borrow in equilibrium) and idt in the case of the Euler equation for type s (who choose to save). These are of the form 1 þ itt t t ltt ¼ bEt f½d þ ð1 dÞpt lt þ 1 þ ð1 dÞpt lt þ 1 g , ð7Þ
Pt þ 1
for each of the two types t ¼ b,s, where for either type t, we use the notation t to denote the opposite type, and Pt þ 1 Pt þ 1 =Pt (where Pt is the Dixit Stiglitz price index) is the gross rate of inflation. These two equations determine the two marginal utilities of expenditure and hence aggregate demand, using (6) as a function of the expected forward paths of the two real interest rates ð1 þ itt Þ=Pt þ 1 and the expected average marginal utility of expenditure far in the future. This generalizes the relation between real interest rates and aggregate demand in the basic New Keynesian model. Note that in the generalized model, the paths of the two different real interest rates (those faced by borrowers and those faced by savers) are both relevant for aggregate demand determination; alternatively, the forward path of the credit spread matters for aggregate demand determination, in addition to the forward path of the general level of interest rates, as in the basic model. (See Cu´rdia and Woodford, 2009, for further discussion.) We assume Calvo style staggered price adjustment, under which the price of each good remains unchanged from one period to the next with probability a. From this we obtain structural relations linking the dynamics of inflation and real activity that are direct generalizations of those implied by the basic New Keynesian model (as presented, for example, in Benigno and Woodford, 2005). As in the representative household model, inflation is determined by a relation of the form5
Pt ¼ PðZt Þ,
ð8Þ 6
where Zt is a vector of two forward looking endogenous variables, determined by a pair of structural relations that can be written in recursive form as b
s
Zt ¼ zðYt , lt , lt ; xt Þ þ Et ½FðZt þ 1 Þ,
ð9Þ
where zðÞ and FðÞ are each vectors of two functions, and the vector of exogenous disturbances xt now includes shocks to technology and tax rates, in addition to preference shocks. (The relations (9) reduce to precisely the equations in Benigno t and Woodford, 2005, in the case that the two marginal utilities of income lt are equated.) This set of structural equations makes inflation a function of the expected future path of output, generalizing the familiar ‘‘New Keynesian Phillips curve’’; but in addition to the expected paths of aggregate output and of various exogenous disturbances, the expected future path of the marginal utility gap fOt g also matters,7 and hence the expected future path of the credit spread (which determines the marginal utility ratio). Thus this part of the model is completely standard, except that ‘‘cost push’’ effects of credit spreads are taken into account. Cu´rdia and Woodford (2009) show that Eqs. (8) (9) can be log linearized to yield a relation identical to the standard ‘‘New Keynesian Phillips curve,’’ except with additional additive terms for the effects of credit spreads. Finally, our model of the effects of the two interest rates on the optimizing decisions of households of the two types imply an equation for aggregate private borrowing. The effects of interest rates both on expenditure and on labor supply (and hence on labor income) can be summarized by the effects of the expected paths of interest rates on the two marginal utilities of income. In the case of an isoelastic disutility of labor effort function, the degree of asymmetry between the amount by which expenditure exceeds income for type b relative to type s households can be written as a function b s Bðlt , lt ,Yt , Dt ; xt Þ, where Dt is an index of price dispersion and xt includes disturbances to both technology and preferences. The index of price dispersion is a positive quantity, equal to 1 if and only if all goods prices at that date are identical, and higher than 1 when prices are unequal. Price dispersion matters because total hours worked (and hence the wage income of both types), for any given quantity of demand Yt for the composite good, is proportional to Dt ; greater price dispersion results in a less efficient composition of output and hence an excess demand for inputs relative to the quantity consumed of the composite good. Real per capita private debt bt then evolves in accordance with a law of motion of the form b
s
ð1 þ pb ot Þbt ¼ pb ps Bðlt , lt ,Yt , Dt ; xt Þ pb bgt þ d½bt1 ð1 þ ot1 Þ þ pb bgt1
1þ idt1
Pt
,
ð10Þ
5 The definition of the function PðÞ, and similarly of the functions referred to in the remaining equations of this section, are given in the Appendix, available online. 6 These are the variables denoted Kt and Ft in Benigno and Woodford (2005) and similarly in Cu´rdia and Woodford (2009). b s 7 Note that using (6) and the definition of Ot , one observes that the values of Yt , Ot and the exogenous disturbances determine the values of lt , lt at any point in time.
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
59
where ot is the short term credit spread defined by 1 þ ot
1þ ibt , 1 þ idt
ð11Þ
and bgt is real per capita government debt (one of the exogenous disturbance processes in our model). The supply of government debt matters for the evolution of private debt because it is another component (in addition to the deposits with intermediaries that finance their lending) of the financial wealth of type s households; because in our model government debt and deposits are substitutes from the standpoint of type s households (who hold positive quantities of both in equilibrium), in equilibrium government debt must also earn the interest rate idt.8 Hence the interest rates idt and ibt (or alternatively, idt and the spread ot ) are the only ones that matter for the evolution of private debt. (Note that Eq. (10) does not correspond to any equation of the basic New Keynesian model, as there can be no private debt in a representative household model.) Finally, as in Benigno and Woodford (2005), the assumption of Calvo style price adjustment implies that the index of price dispersion evolves according to a law of motion of the form
Dt ¼ hðDt1 , Pt Þ,
ð12Þ
where for a given value of Dt1 ,hðDt1 ,Þ has an interior minimum at an inflation rate that is near zero (Pt ¼ 1) when initial price dispersion is negligible (Dt1 near 1), and the minimum value of the function is itself near 1 (i.e., price dispersion continues to be minimal). Relative to this minimum, either inflation or deflation results in a greater degree of price dispersion; and once some degree of price dispersion exists, it is not possible to achieve zero price dispersion again immediately, for any possible choice of the current inflation rate. The system of equations consisting of (6) (10) together with (12) specifies eight equations per period, to determine the b s eight endogenous variables fPt ,Yt , lt , lt ,Zt ,bt , Dt g, given three more equations per period to determine the evolution of the interest rates {idt ,ibt } (and hence of the credit spread) and of the resources fXt g consumed in financial intermediation. The latter equations follow from the decisions of private intermediaries and of the central bank. 2.2. Financial intermediaries We assume an intermediary sector made up of identical, perfectly competitive firms. Intermediaries take deposits, on which they promise to pay a riskless nominal return idt one period later, and make one period loans on which they demand a nominal interest rate of ibt. An intermediary also chooses a quantity of reserves Mt to hold at the central bank, on which it will receive a nominal interest yield of im t . Each intermediary takes as given all three of these interest rates. We assume that arbitrage by intermediaries need not eliminate the spread between ibt and idt, for either of two reasons. On the one hand, resources are used in the process of loan origination; and on the other hand, intermediaries may be unable to tell the difference between good borrowers (who will repay their loans the next period) and bad borrowers (who will be able to disappear without having to repay), and as a consequence have to charge a higher interest rate to good and bad borrowers alike. We suppose that origination of good loans in real quantity Lt requires an intermediary to also originate bad loans in quantity wt ðLt Þ, where wut , w00t Z0, and the function wt ðLÞ may shift from period to period for exogenous reasons. (While the intermediary is assumed to be unable to discriminate between good and bad loans, it is able to predict the fraction of loans that will be bad in the case of any given scale of lending activity on its part.) This scale of operations also requires the intermediary to consume real resources Xpt ðLt ; mt Þ in the period in which the loans are originated, where mt Mt =Pt , and Xpt ðL; mÞ is a convex function of its two arguments, with XpLt Z 0, Xpmt r0, XpLmt r0. We further suppose that for any scale of operations L, there exists a finite satiation level of reserve balances m t ðLÞ, defined as the lowest value of m for which Xpmt ðL; mÞ ¼ 0. (Our convexity and sign assumptions then imply that Xpmt ðL; mÞ ¼ 0 for all m Z m t ðLÞ.) We assume the existence of a finite satiation level of reserves in order for an equilibrium to be possible in which the policy rate is driven to zero, a situation of considerable practical relevance at present that raises interesting theoretical issues. Given an intermediary’s choice of its scale of lending operations Lt and reserve balances mt to hold, we assume that it acquires real deposits dt in the maximum quantity that it can repay (with interest at the competitive rate) from the anticipated returns on its assets (taking into account the anticipated losses on bad loans). Thus it chooses dt such that ð1 þ idt Þdt ¼ ð1 þibt ÞLt þ ð1 þ im t Þmt :
ð13Þ
The deposits that it does not use to finance either loans or the acquisition of reserve balances, dt mt Lt wt ðLt Þ Xpt ðLt ; mt Þ,
ð14Þ
8 Thus we abstract from any transactions role for the deposits that type s households hold with intermediaries. The model can easily be extended to allow deposits to supply transactions services, at the cost of introducing an additional interest-rate spread into the model. Note, however, that neither our account of the way in which the central bank controls short-term interest rates nor our account of the role of credit in macroeconomic equilibrium depends on any monetary role for the liabilities of private intermediaries.
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
60
are distributed as earnings to its shareholders. The intermediary chooses Lt and mt each period so as to maximize these earnings, given idt ,ibt ,im t . This implies that Lt and mt must satisfy the first order conditions
XpLt ðLt ; mt Þ þ wLt ðLt Þ ¼ ot Xpmt ðLt ; mt Þ ¼ dm t
ibt idt , 1 þ idt
idt im t : 1 þ idt
ð15Þ
ð16Þ
Eq. (15) can be viewed as determining the equilibrium credit spread ot as a function ot ðLt ; mt Þ of the aggregate volume of private credit and the real supply of reserves. As indicated above, a positive credit spread exists in equilibrium to the extent that Xpt ðL; mÞ, wt ðLÞ, or both are increasing in L. Eq. (16) similarly indicates how the equilibrium differential dm t between the interest paid on deposits and that paid on reserves at the central bank is determined by the same two aggregate quantities. In addition to these two equilibrium conditions that determine the two interest rate spreads in the model, the absolute level of (real) interest rates must be such as to equate the supply and demand for credit. Market clearing in the credit market requires that bt ¼ Lt þLcb t ,
ð17Þ
Lcb t
represents real lending to the private sector by the central bank, as discussed next. Eqs. (15) (17) then provide three where more equilibrium conditions per period, to determine the three additional endogenous variables fıdt ,ibt ,Lt g along with those cb discussed in the previous section, given paths (or rules for the determination of) the central bank policy variables {Mt,im t ,Lt }. 2.3. Dimensions of central bank policy In our model, the central bank’s liabilities consist of the reserves Mt (which also constitute the monetary base in our simple model), on which it pays interest at the rate im t . These liabilities in turn fund the central bank’s holdings of government debt, and any lending by the central bank to type b households. We let Lcb t denote the real quantity of lending by the central bank to the private sector; the central bank’s holdings of government debt are then given by the residual cb mt Lcb t . We can treat mt (or Mt) and Lt as the bank’s choice variables, subject to the constraints 0 r Lcb t r mt :
ð18Þ
It is also necessary that the central bank’s choices of these two variables satisfy the bound g mt o Lcb t þ bt ,
ð19Þ
bgt
is the total outstanding real public debt, so that a positive quantity of public debt remains in the portfolios of where households. In the calculations below, however, we shall assume that this last constraint is never binding. (We confirm this in our numerical examples.) We assume that central bank extension of credit other than through open market purchases of Treasury securities consumes real resources, just as in the case of private intermediaries, and represent this resource cost by a function cb Xcb ðLcb t Þ, that is increasing and at least weakly convex, with X uð0Þ 4 0, as is discussed further in Section 4. The central bank has one further independent choice to make each period, which is the rate of interest im t to pay on reserves. We assume that if the central bank lends to the private sector, it simply chooses the amount that it is willing to lend and auctions these funds, so that in equilibrium it charges the same interest rate ibt on its lending that private intermediaries do; this is therefore not an additional choice variable for the central bank. Similarly, the central bank receives the market determined yield idt on its holdings of government debt. The interest rate idt at which intermediaries are able to fund themselves is determined each period by the joint inequalities m
mt Z mdt ðLt , dt Þ,
ð20Þ
dm t Z 0,
ð21Þ
together with the ‘‘complementary slackness’’ condition that at least one of (20) and (21) must hold with equality each m period. Here mdt ðL, d Þ is the demand for reserves defined by (16), and defined to equal the satiation level m t ðLÞ in the case m that d ¼ 0. (Condition (20) may hold only as an inequality, as intermediaries will be willing to hold reserves beyond the m satiation level as long as the opportunity cost dt is zero.) We identify the rate idt at which intermediaries fund themselves with the central bank’s policy rate (e.g., the federal funds rate, in the case of the US). The central bank can influence the policy rate through two channels, its control of the supply of reserves and its control m of the interest rate paid on them. By varying mt, the central bank can change the equilibrium differential dt , determined as d , it can change the level of the policy rate i that corresponds to a given the solution to (20) (21). And by varying im t t differential. Through appropriate adjustment on both margins, the central bank can control idt and im t separately (subject to d the constraint that im t cannot exceed it ). We also assume that for institutional reasons, it is not possible for the central bank to pay a negative interest rate on reserves. (We may suppose that intermediaries have the option of holding currency,
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
61
earning zero interest, as a substitute for reserves, and that the second argument of the resource cost function Xpt ðb; mÞ is actually the sum of reserve balances at the central bank plus vault cash.) Hence the central bank’s choice of these variables is subject to the constraints d 0 r im t r it :
ð22Þ
There are thus three independent dimensions along which central bank policy can be varied in our model: variation in the quantity of reserves Mt that are supplied; variation in the interest rate im t paid on those reserves; and variation in the breakdown of central bank assets between government debt and lending Lcb t to the private sector. Alternatively, we can specify the three independent dimensions as interest rate policy, the central bank’s choice of an operating target for the policy rate idt; reserve supply policy, the choice of Mt, which in turn implies a unique rate of interest that must be paid on reserves in order for the reserve supply policy to be consistent with the bank’s target for the policy rate9; and credit policy, 10 the central bank’s choice of the quantity of funds Lcb t to lend to the private sector. We prefer this latter identification of the three dimensions of policy because in this case our first dimension (interest rate policy) corresponds to the sole dimension of policy emphasized in many conventional analyses of optimal monetary policy, while the other two dimensions are additional dimensions of policy introduced by our extension of the basic New Keynesian model.11 Changes in central bank policy along each of these dimensions has consequences for the bank’s cash flow, but we abstract from any constraint on the joint choice of the three variables associated with cash flow concerns. (We assume that seignorage revenues are simply turned over to the Treasury, where their only effect is to change the size of lump sum transfers to the households.) Given that central bank policy can be independently varied along each of these three dimensions, we can independently discuss the criteria for policy to be optimal along each dimension. Here we restrict our analysis to the two ‘‘unconventional’’ dimensions of policy, reserve supply policy and credit policy. The consequences of heterogeneity and credit frictions for interest rate policy (i.e., conventional monetary policy) are addressed in Cu´rdia and Woodford (2009, 2010c). Of course, we have to make some assumption about interest rate policy when considering adjustments of policy along the other two dimensions; in some of the analysis reported below, we assume that interest rate policy is optimal (despite not seeking here to characterize optimal interest rate policy), while in other places we assume a simple conventional specification for interest rate policy (a ‘‘Taylor rule’’). It is also true that the changes in reserve supply policy and credit policy have consequences for optimal interest rate policy; but these are not the concern of the present study, except to the extent that they influence the optimal use of the unconventional policies themselves. 2.4. The welfare objective In considering optimal policy, we take the objective of policy to be the maximization of average expected utility. Thus we can express the objective as maximization of Et0
1 X t
btt0 Ut ,
ð23Þ
t0
where the welfare contribution Ut each period weights the period utility of each of the two types by their respective population fractions at each point in time. As shown in Cu´rdia and Woodford (2009),12 this can be written as Ut ¼ UðYt , Ot , Xt , Dt ; xt Þ,
ð24Þ
9 We might choose to call the second dimension variation in the interest rate paid on reserves, which would correspond to something that the Board of Governors makes an explicit decision about under current US institutional arrangements, as is also true at most other central banks. But description of the second dimension of policy as ‘‘reserve-supply policy’’ allows us to address the question of the value of ‘‘quantitative easing’’ under this heading as well. 10 Here we only consider the kind of credit policy that involves direct lending by the central bank to ultimate borrowers, or (equivalently, in our model, since the loan market is competitive) targeted asset purchases. Thus our ‘‘credit policy’’ is intended to represent, in a stylized way, the kind of programs that became an important part of Fed policy after September 2008, such as the Commercial Paper Funding Facility or the Fed’s purchases of mortgage-backed securities. We do not take up the separate question of what might be accomplished by central-bank lending to intermediaries rather than to ultimate borrowers, as under the ‘‘liquidity facilities’’ that played such an important role in the Fed’s response to the financial crisis up until September 2008. In our model, central-bank lending to intermediaries can also be effective; and as with our analysis of credit policy below, the welfare consequences of such intervention will depend crucially on whether financial disruptions involve increases in the real resource costs of private intermediation or not. Other analyses that distinguish the effects of these two types of credit policy include Reis (2009) and Gertler and Kiyotaki (2010). 11 Goodfriend (2009) similarly describes central-bank policy as involving three independent dimensions, corresponding to our first three dimensions, and calls the first of those dimensions (the quantity of reserves, or base money) ‘‘monetary policy.’’ We believe that this does not correspond to standard usage of the term ‘‘monetary policy,’’ since the main focus of policy deliberations at many central banks prior to the crisis was frequently the choice of an operating target for the policy rate. Reis (2009) also distinguishes among the three dimensions of policy in terms similar to ours. 12 Cu´rdia and Woodford (2009) analyze a special case of the present model, in which central-bank lending and the role of central-bank liabilities in reducing the transactions costs of intermediaries are abstracted from. However, the form of the welfare measure (24) depends only on the nature of the heterogeneity in our model, and the assumed existence of a credit spread and of resources consumed by the intermediary sector; the functions that determine how Ot and Xt are endogenously determined are irrelevant for this calculation, and those are the only parts of the model that are generalized in this paper. Hence the form of the welfare objective in terms of these variables remains the same.
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
62
b
s
where Ot is again the marginal utility gap, Ot lt =lt ; Xt is the total resources consumed in financial intermediation (also including resources used by the central bank, to the extent that it lends to the private sector, as discussed further in Section 4); Dt is the index of price dispersion appearing in (10) (12); and xt is a vector of exogenous disturbances to preferences, technology, and government purchases. It may be useful to briefly explain why the arguments in (24) suffice, and the way each of them affects welfare. In order to derive the period objective in (24), we sum the utility of consumption and disutility of labor for the two types, weighting each P type t by its population fraction pt . The average utility of consumption is equal to t pt ut ðctt ; xt Þ, which depends on cbt ,cst, and exogenous shocks to preferences (i.e., to spending opportunities). But it is possible to solve uniquely for cbt ,cst given values for Yt , Ot , Xt , and the exogenous disturbances (including Gt), using (6) and the definition of Ot ; hence the arguments of (24) suffice to determine this component of average utility. The total disutility of work can be written as a product of factors ~ t ; xt ÞDt , LðOt ÞvðY
ð25Þ
~ t ; xt Þ would be the disutility of supplying quantity Yt of the composite good, if a common quantity yt(j)=Yt were where vðY produced of each of the individual goods, and if the labor effort involved in producing them were efficiently divided between households of the two types13; and LðOt Þ is a distortion factor that arises as a result of differing marginal utilities of income of the two types (which means that their relative wages no longer correctly reflect their relative marginal disutilities of work), leading to an inefficient division of equilibrium work effort across the two types of households. Given these other two factors, the total disutility of work is proportional to Dt because greater price dispersion results in a less efficient composition of the output that comprises a quantity Yt of the composite good. (Note that except for the presence of the factor LðOt Þ, the total disutility of work is the same as in the representative household model.) Hence the arguments of (24) also suffice for the calculation of this term, and so for the calculation of the period t contribution to average utility. From this discussion, it should be evident how each of the four endogenous variables in (24) affects welfare. Given the values of the other variables, increasing Yt increases the average utility of consumption but also the total disutility of work, as in the representative household model. Under standard assumptions about preferences and technology, there is diminishing marginal utility from consumption and increasing marginal disutility of work as Yt increases, so that there should be an interior maximum for Ut as a function of Yt (the location of which will depend on preferences, technology, government purchases, etc.). And given the values of the other variables, Ut is monotonically decreasing in Dt . Both of these effects are similar to those in the representative household model, where fYt , Dt g are the only endogenous variables that matter for welfare.14 With heterogeneity and credit frictions, additional variables become relevant as well. Given values of the other variables, Ut is monotonically decreasing in both Ot and Xt . Average utility is reduced by an increase in Ot because both the efficiency of the allocation of total private expenditure across the two types and the efficiency of the allocation of total work effort across the two types is reduced; it is reduced by an increase in Xt because, for given values of Yt and Gt, a higher value of Xt means less total private expenditure and hence lower values of both cbt and cst (given a value for Ot ). Other variables matter for welfare purely through their effects on the paths of these four endogenous variables. For example, the level of real bank reserves matters in our model, because of its effect on the resources Xt consumed by financial intermediaries. Central bank credit policy can matter in our model as well, to the extent that it reduces credit spreads and as a consequence the size of the equilibrium marginal utility gap Ot . We turn now to an analysis of the optimal use of these additional dimensions of policy in the light of this objective. 3. Reserve-supply policy We shall first consider optimal policy with regard to the supply of reserves, taking as given (for now) the way in which the central bank chooses its operating target for the policy rate idt , and the state contingent level of central bank lending Lcb t to the private sector. Under fairly weak assumptions, we obtain a very simple result: optimal policy requires that intermediaries be satiated in reserves, i.e., that Mt =Pt Z m t ðLt Þ at all times. For levels of reserves below the satiation point, an increase in the supply of reserves has two effects that are relevant for welfare: on the one hand, the resource cost of financial intermediation Xpt is reduced (for a given level of lending by the intermediary sector); and on the other hand, the credit spread ot is reduced (again, for a given level of lending) as a consequence of (15). Each of these effects raises the value of the objective (23); note that reductions in credit spreads increase welfare because of their effect on the path of the marginal utility gap Ot .15 Hence an increase in the supply of reserves is unambiguously desirable, in any period in which they remain below the satiation level.16 Once reserves are at or above the satiation level, however, further increases reduce neither the resource costs of intermediaries nor equilibrium 13 Note that this disutility will depend both on the state of productivity and on preferences regarding labor supply, both of which are elements of the vector xt . 14 Because the evolution of price dispersion is determined entirely by the path of inflation, as indicated by (12), we can alternatively state that aggregate output and inflation are the only endogenous variables that matter for welfare in the representative-household model. 15 Log-linearization of Eq. (7) can be used to show that, up to this log-linear approximation, log deviations of the marginal-utility gap should equal a forward-looking moving average of expected deviations of the credit spread from its steady-state level; see Cu´rdia and Woodford (2009). 16 The discussion here assumes that the upper bound in (18) is not a binding constraint. But if that constraint does bind, then an increase in the supply of reserves relaxes the constraint, and this too increases welfare, so that the conclusion in the text is unchanged.
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
63
credit spreads (as in this case Xpmt ¼ XpLmt ¼ 0), so that there would be no further improvement in welfare. Hence policy is optimal along this dimension if and only if Mt =Pt Z m t ðLt Þ at all times,17 so that
Xpmt ðLt ; mt Þ ¼ 0:
ð26Þ
This is just another example in which the familiar ‘‘Friedman Rule’’ for ‘‘the optimum quantity of money’’ (Friedman, 1969) applies. Note, however, that our result has no consequences for interest rate policy. While the Friedman rule is sometimes taken to imply a strong result about the optimal control of short term nominal interest rates namely, that the nominal interest rate should equal zero at all times the efficiency condition (26), together with the equilibrium m relation (16), implies only that the interest rate differential dt should equal zero at all times. With zero interest on reserves, this would also require that idt = 0 at all times; but given that the central bank is free to set any level of interest on reserves consistent with (22), the efficiency condition (26) actually implies no restriction upon either the average level of the degree of state contingency of the central bank’s target for the policy rate idt . 3.1. Is a reserve supply target needed? Our result about the importance of ensuring an adequate supply of reserves might suggest that the question of the correct target level of reserves at each point in time should receive the same degree of attention at meetings of the FOMC as the question of the correct operating target for the federal funds rate. But deliberations of that kind are not needed in order to ensure fulfillment of the optimality criterion (26). For the efficiency condition can alternatively be stated (using (16)) as requiring that idt =im t at all times. This requires that reserves be supplied to the point at which the policy rate falls to the level of the interest rate paid on reserves, or, in a formulation that is more to the point, that interest be paid on reserves at the central bank’s target for the policy rate. Given a rule for setting an operating target for idt (discussed in the next section), im t should be chosen each period in accordance with the simple rule d im t ¼ it :
ð27Þ
To put this rule into practice, when the central bank acts to implement its target for the policy rate through open market operations, it will automatically have to adjust the supply of reserves so as to satisfy (26). But this does not require a central bank’s monetary policy committee (the FOMC in the case of the US) to deliberate about an appropriate target for reserves at each meeting; once the target for the policy rate is chosen (and the interest rate to be paid on reserves is determined by that, through condition (27), the quantity of reserves that must be supplied to implement the target can be determined by the bank staff in charge of carrying out the necessary interventions (the trading desk at the New York Fed, in the case of the US), on the basis of a more frequent monitoring of market conditions than is possible on the part of the monetary policy committee. One obvious way to ensure that the efficiency condition (27) is satisfied is to adopt a routine practice of automatically paying interest on reserves at a rate that is tied to the current operating target for the policy rate. This is already the practice of many central banks outside the US. At some of those banks, the fixed spread between the target for the policy rate and the rate paid on overnight balances at the central bank is quite small (for example, 25 basis points in the case of the Bank of Canada); in the case of New Zealand, the interest rate paid on overnight balances is the policy rate itself. There are possible arguments (relating to considerations not reflected in our simple model) according to which the optimal spread might be larger than zero, but it is likely in any event to be desirable to maintain a constant small spread, rather than treating the question of the interest rate to be paid on reserves as a separate, discretionary policy decision to be made at each meeting of the policy committee. Apart from the efficiency gains modeled here, such a system should also help to facilitate the central bank’s control of the policy rate (Goodfriend, 2002; Woodford, 2003, Chapter 1, Section 3). 3.2. Is there a role for ‘‘quantitative easing’’? While our analysis implies that it is desirable to ensure that the supply of reserves never falls below a certain lower bound m t ðLt Þ, it also implies that there is no benefit from supplying reserves beyond that level. There is, however, one important exception to this assertion: it can be desirable to supply reserves beyond the satiation level if this is necessary in order to make the optimal quantity of central bank lending to the private sector Lcb t consistent with (18). This qualification is important in thinking about the desirability of the massive expansion in the supply of reserves by the Fed since September 2008, as shown in Fig. 1. As can be seen from Fig. 2, the increase in reserves occurred only once the Fed decided to expand the various newly created liquidity and credit facilities beyond the scale that could be financed simply by reducing its holdings of Treasury securities (as had been its policy over the previous year).18 17 To be more precise, policy is optimal if and only if (26) is satisfied and the upper bound in (18) does not bind. Both conditions will be satisfied by any quantity of reserves above some finite level. 18 Bernanke (2009) distinguishes between the Federal Reserve policy of ‘‘credit easing’’ and the type of ‘‘quantitative easing’’ practiced by the Bank of Japan earlier in the decade, essentially on this ground. Shiratsuka (2009) distinguishes between ‘‘pure credit easing’’ and ‘‘pure quantitative easing’’ in the way that we have proposed, but argues that the actual ‘‘unconventional policies’’ of central banks, including both the Bank of Japan in 2001–2006 and the
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
64
Some have argued, instead, that further expansion of the supply of reserves beyond the level needed to bring the policy rate down to the level of interest paid on reserves is an important additional tool of policy in its own right one of particular value precisely when a central bank is no longer able to further reduce its operating target for the policy rate, owing to the zero lower bound (as at present in the US and many other countries). It is sometimes proposed that when the zero lower bound is reached, it is desirable for a central bank’s policy committee to shift from deliberations about an interest rate target to a target for the supply of bank reserves, as under the Bank of Japan’s policy of ‘‘quantitative easing’’ during the period between March 2001 and March 2006. Our model provides no support for the view that such a policy should be effective in stimulating aggregate demand. Indeed, it is possible to state an irrelevance proposition for quantitative easing in the context of our model. Let the three dimensions of central bank policy be described by functions that specify the operating target for the policy rate, the supply of reserves, the interest rate to be paid on reserves, and the quantity of central bank credit as functions of macroeconomic conditions. For the sake of concreteness, we may suppose that each of these variables is to be determined by a Taylor type rule, id
idt ¼ f ðpt ,Yt ,Lt ; xt Þ, m
Mt =Pt ¼ f ðpt ,Yt ,Lt ; xt Þ, im
im t ¼ f ðpt ,Yt ,Lt ; xt Þ, L
Lcb t ¼ f ðpt ,Yt ,Lt ; xt Þ,
ð28Þ ð29Þ ð30Þ ð31Þ
where the functions are such that constraints (18) (22) are satisfied for all values of the arguments. (Here the vector of exogenous disturbances xt upon which the reaction functions may depend includes the exogenous factors that shift the id im L function Xpt ðL; mÞ.) Then our result is that given the three functions f ðÞ, f ðÞ, and f ðÞ, the set of processes d b fpt ,Yt ,Lt ,bt ,it ,it , Ot , Dt g that constitute possible rational expectations equilibria is the same, independently of the choice of m m the function f ðÞ, as long as the specification of f ðÞ is consistent with the other three functions (in the sense that (18) and (20)) are necessarily satisfied, and that (20) holds with equality in all cases where (21) is a strict inequality).19 m Of course, the stipulation that f ðÞ be consistent with the other functions uniquely determines what the function must m be for all values of the arguments for which the functions id ðÞ and im ðÞ imply that dt 4 0. However, the class of policies considered allows for an arbitrary degree of expansion of reserves beyond the satiation level in the region where those m functions imply that dt ¼ 0, and in particular, for an arbitrary degree of quantitative easing when the zero bound is reached (i.e., when idt = im t =0). The class of policies considered includes the popular proposal under which the quantity of excess reserves should depend on the degree to which a standard Taylor rule (unconstrained by the zero bound) would call for a negative policy rate. Our result implies that there should be no benefits from such policies. Our result might seem to be contradicted by the analysis of Auerbach and Obstfeld (2005), in which an open market operation that expands the money supply is found to stimulate real activity even when the economy is at the zero bound at the time of the monetary expansion. But their thought experiment does not correspond to pure quantitative easing of the kind contemplated in the above proposition, because they specify monetary policy in terms of a path for the money supply, and the policy change that they consider is one that permanently increases the money supply, so that it remains higher after the economy has exited from the ‘‘liquidity trap’’ in which the zero bound is temporarily binding. The contemplated id policy change is therefore not consistent with an unchanged reaction function f ðÞ for the policy rate, and the effects of the intervention can be understood to be the consequences of the commitment to a different future interest rate policy. Our result only implies that quantitative easing should be irrelevant under two conditions: that the increase in reserves finances an increase in central bank holdings of Treasury securities, rather than an increase in central bank lending to the private sector; and that the policy implies no change in the way that people should expect future interest rate policy to be conducted. Our model does allow for real effects of an increase in central bank lending Lcb t financed by an increase in the supply of reserves, in the case that private sector financial intermediation is inefficient20; but the real effects of the increased central bank lending in that case are the same whether the lending is financed by an increase in the supply of reserves or by a reduction in central bank holdings of Treasury securities. Our model also allows for real effects of an announcement that interest rate policy in the future will be different, as in the case where a central bank commits itself not to return immediately to its usual Taylor rule as soon as the zero bound ceases to bind, but promises instead to maintain policy accommodation for some time after it would become possible to comply with the Taylor rule (as discussed (footnote continued) Federal Reserve during the current crisis, are mixtures of the two pure types. Ueda (2009) similarly argues that many central banks represent mixed cases. 19 This result generalizes the irrelevance result for quantitative easing in Eggertsson and Woodford (2003) to a model with heterogeneity and credit frictions. 20 This result differs from that obtained in Eggertsson and Woodford (2003), where changes in the composition of the assets on the central bank’s balance sheet are also shown to be irrelevant. That stronger result depends on the assumption of a representative household as in the earlier paper, or alternatively, frictionless financial intermediation, as discussed in Cu´rdia and Woodford (2010b, Section 1).
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
65 120
400
80
200
40
Trillion Yen
Trillion Yen
600
4 Q
08
4
0
20
4
Q 07
4
Q 06
20
4
Q 05
20
20
4
Q 04
20
4
Q 03
4
Q 02
20
4
Q 01
20
20
4
Q 00
20
4
Q 99
19
4
Q 98
4
Q 97
19
Q 96
4
Monetary Base (right axis)
19
19
4
Q 95
19
4
Q 94
Q 93
19
4
Q 92
19
19
4
Q
Q
91
90
19
19
4
Nominal GDP (left axis)
0
Fig. 3. The monetary base and nominal GDP for Japan (both seasonally adjusted), 1990–2009. The shaded region shows the period of ‘‘quantitative easing,’’ from March 2001 through March 2006. (Sources: IMF International Financial Statistics and Bank of Japan.)
in the next section). But such a promise (if credible and correctly understood by the private sector) should increase output and prevent deflation to the same extent even if it implies no change in policy during the period when the zero lower bound binds. While our definition of quantitative easing may seem a narrow one, the policy of the Bank of Japan during the period 2001 2006 fits our definition fairly closely. The BOJ’s policy involved the adoption of a series of progressively higher quantitative targets for the supply of reserves, and the aim of the policy was understood to be to increase the monetary base, rather than to allow the BOJ to acquire any particular type of assets. The assets purchased consisted primarily Japanese government securities and bills issued by commercial banks; while there were also some more ‘‘unconventional’’ asset purchases under the quantitative easing policy direct purchases of asset backed securities and of stocks the size of these operations was quite small relative to the total increase in the supply of reserves shown in Fig. 3.21 Finally, there was no suggestion that the targets of policy after the end of the zero interest rate period would be any different than before; there was no commitment to maintain the increased quantity of base money in circulation permanently, and indeed, once it was judged time to end the zero interest rate policy, the supply of reserves was rapidly contracted again, as also shown in Fig. 3. Our theory suggests that expansion of the supply of reserves under such circumstances should have little effect on aggregate demand, and this seems to have been the case. For example, as is also shown in Fig. 3, despite an increase in the monetary base of 60 percent during the first two years of the quantitative easing policy, and an eventual increase of nearly 75 percent, nominal GDP never increased at all (relative to its March 2001 level) during the entire five years of the policy.22 Apart from the absence of the effects on aggregate expenditure that simple quantity theoretic reasoning would have predicted, there was also little evidence of effects of the policy on longer term interest rates (the channel through which it might have been expected to ultimately influence aggregate expenditure). Those studies of Japan’s experience with quantitative easing that find some reduction in longer term interest rates attribute this mainly to successful signalling by the BOJ of an intention to maintain the zero interest rate policy, and find little effect on bond yields of the increase in the supply of reserves itself (Okina and Shiratsuka, 2004; Oda and Ueda, 2007; Ugai, 2007; Ueda, 2009).23 Of course, it is
21 According to Bank of Japan statistics, these ‘‘unconventional’’ purchases had a value only slightly greater than 2 trillion yen at their maximum, whereas the total increase in the monetary base during the quantitative easing (QE) period was in excess of 45 trillion yen. For more detailed discussion of the different aspects of the BOJ policy during the period, and an attempt to separate the effects of targeted asset purchases from those of quantitative easing, see Ueda (2009). Shiratsuka (2009) provides additional information. 22 As indicated in Fig. 3, over the first two years of the quantitative easing policy, nominal GDP fell by more than 4 percent, despite extremely rapid growth of base money. While nominal GDP recovered thereafter, it remained below its 2001:Q1 level over the entire period until 2006:Q4, three quarters after the official end of quantitative easing, by which time the monetary base had been reduced again by more than 20 percent. Moreover, even if the growth of nominal GDP after 2003:Q1 is regarded as a delayed effect of the growth in the monetary base two years earlier, this delayed nominal GDP growth was quite modest relative to the size of the expansion in base money. 23 Baba et al. (2006) find some effects of BOJ purchases of commercial paper on the spreads associated with those particular types of paper. But this is really more evidence of the effectiveness of targeted asset purchases than of effectiveness of quantitative easing as such. See also the discussion by Ueda (2009).
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
66
perfectly consistent with our model that signals about future interest rate policy can be an effective channel through which a central bank can seek to provide further monetary stimulus even when the current policy rate is constrained by the zero lower bound.24 4. Optimal credit policy We turn now to another of our three independent dimensions of central bank policy, namely, adjustment of the composition of the asset side of the central bank’s balance sheet, taking as given the overall size of the balance sheet (determined by the reserve supply decision discussed above). In this section, we take for granted that reserve supply policy is being conducted in the way recommended in the previous section, i.e., that the rate of interest on reserves will satisfy (27) at all times. In this case, we can replace the function Xpt ðLt ; mt Þ by p
X t ðLt Þ Xpt ðLt ; m t ðLt ÞÞ
ð32Þ
and the function ot ðLt ; mt Þ, defined by the left hand side of (15), by
o t ðLt Þ ot ðLt ; m t ðLt ÞÞ,
ð33Þ 25
Xpt
and since there will be satiation in reserves at all times. Using these functions to specify the equilibrium evolution of ot as functions of the evolution of aggregate private credit, we can then write the equilibrium conditions of the model without any reference to the quantity of reserves or to the interest rate paid on reserves. We wish to consider alternative possible state contingent evolutions for central bank lending {Lcb t }, under various maintained assumptions about interest rate policy (made explicit below). According to the traditional doctrine of ‘‘Treasuries only,’’ the central bank should not vary the composition of its balance sheet as a policy tool; instead, it should avoid both balance sheet risk and the danger of politicization by only holding (essentially riskless) Treasury securities at all times, while varying the size of its balance sheet to achieve its stabilization goals for the aggregate economy.26 And even apart from these prudential concerns, if private financial markets can be relied upon to allocate capital efficiently, it is hard to argue that there would be any substantial value to allowing the central bank this additional dimension of policy. Eggertsson and Woodford (2003) present a formal irrelevance proposition in the context of a representative household general equilibrium model; in their model, the assets purchased by the central bank have no consequences for the equilibrium evolution of output, inflation or asset prices and this is true regardless of whether the central bank purchases long term or short term assets, nominal or real assets, riskless or risky assets, and so on. And even in a model with heterogeneity of the kind considered here, the composition of the central bank’s balance sheet would be irrelevant if we were to assume frictionless private financial intermediation, since private intermediaries would be willing to adjust their portfolios to perfectly offset any changes in the portfolio of the central bank, as discussed in Cu´rdia and Woodford (2010b, Section 1). This irrelevance result does not hold, however, in the presence of credit frictions of the kind assumed here, as shown by the numerical illustrations below of the difference between the economy’s evolution under an optimal credit policy and under the constraint of ‘‘Treasuries only.’’27 Moreover, a growing empirical literature finds that targeted asset purchases and/or direct extensions of credit to private borrowers by the Fed and other central banks have indeed affected equilibrium rates of return.28 We accordingly consider the optimal use of this additional dimension of policy. In our model, an increase cb in Lcb t can improve welfare on two grounds: for a given volume of private borrowing bt, an increase in Lt allows the volume p of private lending Lt to fall, which should reduce both the resources Xt consumed by the intermediary sector and the equilibrium credit spread ot (due to equilibrium relation (15)). Under plausible conditions, our model implies both a positive shadow value jX,t of reductions in Xt (the Lagrange multiplier associated with the resource constraint (6)) and a positive shadow value jo,t of reductions in ot ; hence an increase in Lcb t should be desirable on both grounds. In the absence of any assumed cost of central bank credit policy, one can easily obtain the result that it is always optimal for the central bank to lend in amount sufficient to allow an equilibrium with Lt =0, i.e., the central bank should substitute for private credit markets altogether. Of course, we do not regard this as a realistic conclusion. As a simple way of introducing into our calculations the fact that the central bank is unlikely to have a comparative advantage at the activity of credit allocation under normal circumstances, we assume that central bank lending consumes real resources in p a quantity Xcb ðLcb t Þ, by analogy with our assumption that real resources Xt are consumed by private intermediaries. The 24
See Eggertsson and Woodford (2003) and Cu´rdia and Woodford (2010a) for further discussion of the effectiveness of and ideal use of this aspect of
policy. Even if at some times mt exceeds m t ðLt Þ, this will not affect the values of Xpt or ot . See Goodfriend (2009) for discussion of this view and a warning about the dangers of departing from it. 27 Of course, one might be interested in the effectiveness of credit policy, quite apart from considerations of the optimal policy from the standpoint of the welfare criterion proposed here. Cu´rdia and Woodford (2010b, Section 4) analyze the consequences of simple rules for credit policy, and show that a sufficiently aggressive credit policy is able to largely insulate aggregate output, the composition of expenditure, and inflation from the effects of ‘‘purely financial’’ disturbances. This result is relatively independent of the nature of the ‘‘purely financial’’ distribution. The question whether it is optimal to stabilize the economy to that extent is instead much more sensitive to the nature of the disturbance, as discussed here. 28 See, e.g., Ashcraft et al. (2010), Baba et al. (2006), Gagnon et al. (2010), Sarkar (2009), Sarkar and Shrader (2010), and the discussion in Cu´rdia and Woodford (2010b, Section 1.3). For skeptical readings of the evidence, see instead Taylor (2009) and Stroebel and Taylor (2009). 25 26
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
67
Table 1 Numerical parameter values used.
pb
0.5
sb
d b
0.975
ss
0.990
n
0.105
sb ss
b
h =h
1
s
b
l =l
0.798
ðy1Þ
0.602
f
a
13.6 s
1
g
1
0.15
t
0.2
0.75
X =Y
0.0003
0.66
w =L 1þo Z
(1.02)1/4
2.72
b =Y
0
1.22
b=Y
3.2
0 51.6
function Xcb ðLÞ is assumed to be increasing and at least weakly convex; in particular, we assume that Xcb uð0Þ 4 0, so that there is a positive marginal resource cost of this activity, even when the central bank starts from a balance sheet made up entirely of Treasury securities. The first order conditions for optimal choice of Lcb t then become p
cb cb jX,t ½X ut ðbt Lcb t Þ X uðLt Þþ jo,t ½X
p 00
t ðbt
00 cb Lcb t Þ þ w t ðbt Lt Þ r0,
ð34Þ
Lcb t Z 0,
ð35Þ
together with the complementary slackness condition that at least one of conditions (34) or (35) must hold with equality in each period. (Here the first expression in square brackets in (34) is the partial derivative of Xt with respect to Lcb t , holding constant the value of total borrowing bt, while the second expression in square brackets is the partial derivative of ot with respect to Lcb t under the same assumption.) A ‘‘Treasuries only’’ policy is optimal in the event of a corner solution, in which (34) is an inequality, as will be the case if Xcb uð0Þ is large enough. In our view, it is most reasonable to calibrate the model so that this is true in steady state. Then not only will the optimal policy involve ‘‘Treasuries only’’ in the steady state, but (assuming that the inequality is strict at the steady state) this will continue to be true in the case of any stochastic disturbances that are small enough. However, it will remain possible for the optimal policy to require Lcb t 4 0 in the case of certain large enough disturbances. This is especially likely to be true in the case of large enough disruptions of the financial sector, of a type that increase the marginal resource p cost of private intermediation (the value of X u) and/or the degree to which increases in private credit require a larger credit spread (the value of o u). p
However, not all ‘‘purely financial’’ disturbances by which we mean exogenous shifts in the functions X t ðLÞ or wt ðLÞ of a type that increase the equilibrium credit spread o t ðLÞ for a given volume of private credit are equally likely to justify an active central bank credit policy on the grounds just mentioned. In fact, the optimal credit policy response to a given shift in the function o t ðLÞ depends greatly on the source of the shift.29 To illustrate this, we consider the quantitative effects of alternative types of purely financial disturbances in a calibrated model. The numerical values for parameters used in our calculations are summarized in Table 1.30 Here bars denote the implied b b steady state values of various state variables: l is the steady state value of lt ,b is the steady state value of bt, and so on. Our quantitative assumptions about the credit frictions are particularly worthy of comment. In the absence of shocks, we assume that wt ðLÞ ¼ 0, so that all loans are expected to be repaid, and the credit spread is due purely to the resource costs of intermediation. We assume an intermediation technology such that when intermediaries hold reserves at or above the satiation level (as occurs in equilibrium under the optimal reserve supply policy assumed in this section), p
p
X ðLÞ ¼ X~ LZ
ð36Þ
in the absence of shocks. (Below, we consider a variety of types of exogenous shifts in these two functions.) We specify the ~ p so that the steady state credit spread o equals 2.0 percentage points per annum, following Mehra unperturbed value of X et al. (2008). Combined with our assumption (implied by the value of d) that ‘‘types’’ persist for 10 years on average, this b
s
implies a steady state ‘‘marginal utility gap’’ O l =l ¼ 1:22, so that there would be a non trivial welfare gain from transferring further resources from savers to borrowers. 29 Our result here is quite different from that in Cu´rdia and Woodford (2009), where the consequence of a ‘‘purely financial’’ disturbance for optimal interest-rate policy, taking as given the path of central-bank lending to the private sector, depends (to a first approximation) only on the size of the shift in o t ðLÞ, which is why the paper does not bother to show the optimal responses to more than one type of purely financial disturbance. The same is true of the calculations reported in Cu´rdia and Woodford (2010b) that illustrate the ability of credit policy to stabilize the economy in response to a purely financial disturbance: we show our results for only one type of disturbance because the figures are fairly similar when we consider alternative disturbances that shift the function o t ðLÞ to the same extent. 30 The parameter values are essentially the same as those used in the numerical analysis in Cu´rdia and Woodford (2009), where they are discussed in greater detail. The one important difference is that here we calibrate the model so that the steady-state real return on deposits (identified with the real federal funds rate, for purposes of the empirical interpretation of the model) is 3.0 percent per annum, rather than 4 percent as in the earlier paper. Several other numbers in Table 1 are also slightly different from those reported in the appendix to Cu´rdia and Woodford (2009), but these are consequences of the change in the calibration target for the steady-state real policy rate, given unchanged numerical values for the other calibration targets discussed in our earlier paper. See the appendix for further discussion of the calibration.
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
68
The value of Z given in Table 1 implies that a 1 percent increase in private lending results in an increase in the marginal cost of intermediation, and hence in the equilibrium credit spread, of one percentage point (per annum). This implies fairly inelastic credit supply by the private sector, but we believe that this is the case of greatest interest for the exercises here, both because private credit supply is often asserted to be quite inelastic during financial crises, and because this is in any event the assumption most favorable to a potential role for central bank credit policy, as in this case a substitution of central bank lending for private lending to even a modest degree can lower the marginal cost of private lending and hence the equilibrium credit spread.31 We are interested in biasing our results in this direction, not to pre judge the desirability of central bank lending, but because our results show in any event that the justification for active credit policy is often fairly modest, even in the case of financial disturbances that increase credit spreads by a significant amount. It is most interesting to observe that this is true even under the assumption of quite a large value for Z; in the case of a more modest value of Z, then both the shadow value of allowing active credit policy (relaxing the constraint that Lcb t =0) and the optimal scale of central bank lending in response to shocks will in all cases be substantially smaller than the values reported below.
4.1. The value of relaxing the ‘‘Treasuries only’’ restriction We first assume that a strict policy of ‘‘Treasuries only’’ is maintained, but seek to determine the shadow value of relaxing this constraint, i.e., the marginal increase in the value of the welfare objective (23) that is achieved by a marginal increase in Lcb t above zero in some period. This shadow value is given by the left hand side of (34); a positive value would 32 Since the shadow value is obviously reduced if we assume a imply that welfare is increased by allowing Lcb t to be positive. cb and since nothing about the equilibrium associated with the higher marginal cost X uð0Þ of central bank lending an alternative ‘‘Treasuries only’’ constraint is affected by the assumed value of Xcb uð0Þ, other than this shadow value measure of the value of relaxing the constraint, that is economically meaningful, is to report the value of Xcb uð0Þ that is required in order for the shadow value of relaxation of the constraint to be exactly zero. (The higher the shadow value of relaxing the constraint, in the case of any fixed value of Xcb uð0Þ, the higher the value of Xcb uð0Þ that would be required to reduce the shadow value to zero; and indeed one of these quantities is a linear function of the other.) Thus the quantity that we report in our figures is the value of p
Xcbu,crit X ut ðLt Þ þ t
jo,t p 00 ½X t ðLt Þ þ w00 t ðLt Þ jX,t
ð37Þ
in an equilibrium where the ‘‘Treasuries only’’ policy is imposed. Note that this quantity exceeds the marginal resource cost p p of lending by private intermediaries (X u) to the extent that either X t ðLÞ or wt ðLÞ is a strictly convex function. Under the p calibration used here, X u is equal to 2.0 percent per annum in the steady state, while the steady state value of Xcbu,crit , the minimum marginal cost of central bank lending required for ‘‘Treasuries only’’ to be optimal, is nearly 3.5 percent per p annum, as shown in Fig. 4. The assumed degree of convexity of the function X ðLÞ makes a substantial difference. Under the calibration assumed, ‘‘Treasuries only’’ will be optimal in the steady state, if we make the further assumption that Xcb uð0Þ is equal to 3.5 percent per annum or more. But what if a financial disturbance causes a significant increase in the size of credit spreads? The answer depends on the nature of the financial disturbance, and not only on the size of the increase in credit spreads. To illustrate this point, let us consider four different possible purely financial disturbances, each of which will be assumed to increase the value of o t ðLÞ by the same number of percentage points. By an additive shock, we mean one that translates the schedule o t ðLÞ vertically by a constant amount; a multiplicative shock will instead multiply the entire schedule o t ðLÞ by some constant factor greater than 1. We shall also distinguish between disturbances that change the function X t ðLÞ (‘‘X shocks’’) and disturbances that change the function wt ðLÞ (‘‘w shocks’’). Thus a ‘‘multiplicative w shock’’ is a change in the function wt ðLÞ as a consequence of which the schedule o t ðLÞ is multiplied by a factor greater than 1 for all values of L, and so on. to each of these four types of purely financial disturbance, when the model is Fig. 4 plots the dynamic response of Xcbu,crit t calibrated as discussed above. (The quantity defined in (37) is multiplied by 400 so that the interest rate differential is expressed in percentage points per annum.) In these simulations, both interest rate policy and reserve supply policy are assumed to be optimal, and each of the four disturbances is of a size that increases the value of ot ðLÞ by 4 percentage points per annum (i.e., from 2.0 percent to 6.0 percent).33 Moreover, the effect of the disturbance is assumed to decay 31 In the case that the private intermediary sector has a constant marginal cost, rather than one increasing with the volume of private lending, then central-bank lending will reduce the equilibrium credit spread only to the extent that it completely replaces private lending, by lending at a rate that is too low to be profitable for private intermediaries at any positive scale of operation. 32 Note that condition (34) requires the shadow value to be non-positive under an optimal policy, as there is assumed to be no practical obstacle (apart from those reflected in the central bank’s resource cost function Xcb ðLÞ) to a positive level of central-bank lending. 33 This is not quite the same thing as defining the shocks so that they all increase equilibrium credit spreads to the same amount, under the assumption of no central-bank lending. Because Lt declines from the value L in response to the shocks, ot o t ðLt Þ does not increase as much as o t ðLÞ. We prefer to measure the size of the shock in terms that do not depend on a calculation of the endogenous response to the shock; but each of these shocks does increase the equilibrium credit spread, and to a roughly similar extent.
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
7
69
Mult Mult Additive Additive
6.5 6 5.5 5 4.5 4 3.5 3 2.5 0
4
8
12
16
20
cb
Fig. 4. Response of the critical threshold value of X uð0Þ for a corner solution, in the case of four different types of ‘‘purely financial’’ disturbances, each of which increases ot ðLÞ by 4 percentage points. Interest-rate policy responds optimally in each case.
exponentially, so that
o t ðLÞ ¼ o þ ½o 0 ðLÞ o rt for all t Z0 for each of the four shocks, and again we assume that r ¼ 0:9. The figure clearly shows that the degree to which a financial disturbance provides a justification for active central bank credit policy depends very much on the reason for the increase in spreads. In fact, the factors that affect the size of Xcbu,crit t p
are not too closely similar to those that affect the size of o t ðLÞ. The increase in o t ðLÞ is the sum of the increases in X uðLÞ and wuðLÞ. The value of
Xcbu,crit is t p 00
p
also increased by an increase in X u, but not by an increase in wu as such. Moreover,
Xcbu,crit t
is
increased by increases in X and w00 , or in the relative shadow price jo,t =jX,t (which is generally increased by a financial disturbance, since an increase in credit spreads increases the marginal distortion associated with a given further increase in spreads), whereas these do not change the value of ot ðLÞ. increases the most if the credit spread increases due to a ‘‘multiplicative X’’ shock, since in The figure shows that Xcbu,crit t p
this case the increase in the spread is due entirely to an increase in X u and X p
p 00
(and hence o u) increases in the same
p
proportion as does X u. Only one of these two effects (the increase in X u or the increase in o u) is present in the case of the ‘‘additive X’’ or ‘‘multiplicative w’’ shocks, and neither is present in the case of an ‘‘additive w’’ shock. In this last case (an increase in the fraction of loans that are expected not to be repaid, that is independent of the volume of private lending), p
p
neither X uðLÞ nor o uðLÞ increases due to the shock, while both X uðLt Þ and o uðLt Þ decrease, owing to the decrease in Lt (as a consequence of the increase in credit spreads).34 Hence in the case of an ‘‘additive w’’ shock, the marginal social value of central bank lending actually decreases at the time of the shock. Clearly, the mere fact that a given disturbance is observed to increase credit spreads does not in itself prove that it would increase welfare for the central bank to lend directly to private borrowers. This is not because conventional monetary policy (i.e., interest rate policy) alone suffices to eliminate the distortions created by such a disturbance (in our model, it cannot, even if it can mitigate the distortions created by the disturbance to some extent); nor is it because central bank credit policy is unable to influence market credit spreads (in our model, active credit policy would reduce the size of the credit spread). But even granting both of these points, if central bank lending is not costless (and we believe that it should not be considered to be), then it is necessary to balance the costs of intervention against the benefits expected to be achieved; and our model implies that there is no simple relation between the outcome of this tradeoff and the degree to which credit spreads increase in response to a shock. Financial disturbances increase the marginal social benefit of central bank credit policy to a greater extent, however (relative to the size of the disturbance), if the zero lower bound on the policy rate prevents the policy rate from being cut in 34 This last effect, which by itself reduces the marginal social value of central-bank lending, is present in the case of all four disturbances, but in the . other three cases this effect is outweighed by the effects discussed in the text that increase the value of Xcbu,crit t
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
70
Opt ma Interest Rate Po cy
Ignoring the ZLB
Accounting for the ZLB
30
30
25
25
20
20
15
15
10
10
5
5
0
0
Tay or Ru e
0
4
8
12
16
20
0
30
30
25
25
20
20
15
15
10
10
5
5
0
4
8
12
16
20
Mult Mult Additive Additive
0 0
4
8
12
16
20
0
4
8
12
16
20
Fig. 5. Response of the critical threshold value of Xcb uð0Þ for a corner solution, in the case of financial disturbances that increase ot ðLÞ by 12 percentage points. Interest-rate policy responds optimally in the panels of the top row, but follows a Taylor rule in the bottom panels. The zero lower bound is assumed not to constrain interest-rate policy in the panels of the left column, while the constraint is imposed in the corresponding panels of the right column.
response to the shock as much as is assumed in Fig. 4. The distortions created by a binding zero lower bound are even greater under the hypothesis that policy is conducted in a forward looking way after the period in which the zero lower bound constrains the policy rate, so that there is no commitment to subsequent reflation of a kind that would mitigate the extent to which the zero bound results in an undesirably high level of the real policy rate. (An optimal interest rate policy commitment, that takes account of the occasionally binding zero lower bound, will include a commitment to history dependent policy of this sort, as discussed in Eggertsson and Woodford, 2003, and Cu´rdia and Woodford, 2010a. However, such policy requires a type of commitment that actual central bankers seem quite reluctant to contemplate, as discussed for example by Walsh, 2009.) If a binding zero lower bound coincides with this kind of expectations about future monetary policy, the marginal social benefit of credit policy may be much greater than would be suggested by Fig. 4. This is illustrated by Fig. 5, where the same four types of financial disturbances are considered, but the disturbances are are shown under two different assumed to be three times as large as in Fig. 4. In the figure, the responses of Xcbu,crit t assumptions about interest rate policy: in the top panels, interest rate policy is assumed to be optimal, while in the bottom row it is assumed to follow a Taylor rule of the form idt ¼ maxfr d þ fp pt þ fy Y^ t ,0g,
ð38Þ
here written so as to respect the zero lower bound on short term nominal interest rates. In this equation, pt logPt is the inflation rate, Y^ t logðYt =Y Þ, and r d is the steady state real policy rate,35 so that the policy rule is consistent with the zero inflation steady state (discussed above) in the absence of disturbances. The Taylor rule coefficients are assigned the values fp ¼ 2, fy ¼ 1=4, in rough accordance with estimates of US monetary policy in recent decades.36 The figure also illustrates the consequences of the zero lower bound on interest rates; the panels in the left column show the response of Xcbu,crit t As explained above, we calibrate the model so that 1 þ r d ð1:03Þ1=4 . These are the baseline parameter values used in the numerical analysis of the representative-household New Keynesian model in Woodford (2003, Chapter 4). We use these parameter values in our analysis in Cu´rdia and Woodford (2009) of the model’s implications for the effects of disturbances when monetary policy follows a Taylor rule, in order to allow direct comparison with the results shown in Woodford (2003) for the basic New Keynesian model. 35 36
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
71
under the assumption that the zero lower bound can be ignored (even though this means that the policy rate is negative in the quarters immediately following the shock, under either assumption about monetary policy), whereas the corresponding panels in the right column show the response when the zero lower bound is imposed as a constraint on monetary policy. In the upper left panel of Fig. 5, interest rate policy is optimal and the zero lower bound is assumed not to bind, as in Fig. 4; hence the responses to all four disturbances are exactly the same as in Fig. 4, except that the scale of the departures from the steady state is multiplied by a factor of three. In the lower left panel, we can see the difference that is made by is at least somewhat greater assuming instead that interest rate policy follows the Taylor rule (38).37 The increase in Xcbu,crit t declines by less in the case of each of the types of financial disturbances; even in the case of the ‘‘additive w’’ shock, Xcbu,crit t than it does under optimal interest rate policy. Thus optimal interest rate policy reduces at least slightly the welfare gain from active credit policy. In both panels of the left column of Fig. 5, however, the policy rate is assumed to become negative in response to the disturbances. The right column shows how the responses are modified when the zero lower bound is respected in our calculations. Because the zero lower bound binds in the quarters immediately following financial disturbances of the magnitude assumed in this figure,38 the marginal social benefit of central bank credit policy is greater than it would be if it were possible for the central bank to lower the policy rate below zero, as assumed in the panels in the left column. Even in the case of an ‘‘additive w’’ shock, it can be optimal for the central bank to lend to the private sector when such a shock occurs, even though it was not optimal in the absence of the shock. (In the case shown in the figure, this would be true in the case of a value for Xcb uð0Þ between 3.5 and 4.0 percent per annum, for example.) The desirability of active credit policy is even greater if the zero lower bound binds and interest rate policy is determined as in (38). In this case, there is a large as a result of a financial disturbance; and interestingly, in this case, the size of the increase is similar increase in Xcbu,crit t regardless of the source of the increase in credit spreads. Under each of the four types of financial disturbance that are considered, it will be optimal for the central bank to lend to private borrowers, at least in the first two quarters, even if the marginal cost of central bank lending is as high as 10 percentage points per annum. Thus there is a clearer case for active central bank credit policy in the case of a financial disturbance that is severe enough to cause the zero lower bound on the policy rate to become a binding constraint; and in this case, it is to a first approximation only the size and persistence of the effects of the disturbance on credit spreads that matters for determining the extent to which credit policy is desirable. We turn next to the characterization of the optimal degree of central bank lending to the private sector in such cases. 4.2. Optimal credit policy under alternative financial disturbances Here we consider the optimal state contingent evolution of central bank lending {Lcb t }, imposing only the constraint that it be non negative, and taking account of the resource costs of loan origination and monitoring by the central bank. We continue to assume the existence of a competitive loan market, so that the central bank lends at the same (market clearing) interest rate ibt as the private intermediaries, who continue to lend even when the central bank lends as well. (Our assumption of a highly convex cost function for private intermediaries ensures that they continue to serve part of the market, even if credit policy lowers the equilibrium credit spread by several percentage points.) For simplicity, we assume ~ cb Lcb (i.e., that the in the remainder of this section that the resource costs of central bank lending are of the form Xcb ðLcb Þ ¼ X ~ cb is at least as large as marginal cost of central bank lending is independent of the scale of its intervention), but that X cbu,crit
, the steady state value of Xcbu,crit , so that ‘‘Treasuries only’’ is an optimal policy in the steady state. t As the results of the previous subsection suggest, the degree to which active credit policy is optimal varies with the nature of the financial disturbance. Our results above already suggest that the case for such intervention is strongest in the case of a ‘‘multiplicative X’’ shock, so we consider this case first. Fig. 6 shows the impulse responses to a ‘‘multiplicative X’’ shock of the magnitude considered in Fig. 4, under the assumption that monetary policy follows the Taylor rule (38), when the model ~ cb ¼ X cbu,crit exactly. The dashed lines show the impulse responses of several is calibrated as discussed above and X
X
endogenous variables in the case that the ‘‘Treasuries only’’ policy is adhered to. The shock increases the equilibrium credit spread by more than 2.5 percentage points,39 and as a consequence, private borrowing (and likewise private credit) contracts by more than 2 percent. This results not only in a contraction of aggregate output and in deflation, but also in increased inefficiency of the composition of private expenditure: spending by type b households contracts by more than 6 percent, while spending by type s households actually increases (owing to the low returns available on their savings). 37 Exactly the same difference would be made in the case of disturbances of the size assumed in Fig. 4, if we assumed that policy followed the Taylor rule. Note that in the case of the smaller disturbances assumed in Fig. 4, there would be no difference between the two columns, as the zero lower bound would not be a binding constraint even when imposed. 38 In the case of optimal interest-rate policy, the lower bound binds in the first three quarters in the case of the X shocks, and in the first four quarters in the case of the w shocks. In the case of the Taylor rule, the lower bound binds in the first two quarters under all four shocks. The policy rate remains at the lower bound for a longer time under an optimal policy commitment, for the same reason as in the numerical example of Eggertsson and Woodford (2003). 39 Note that the spread does not increase by the full 4 percent by which o t ðLÞ increases, because of the contraction of Lt.
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
72
id (level)
Y 0
0
3 2 1 0
0
8
16
24
32
40
0
8
16
24
32
40
0
8
16
24
32
40
Lcb
b 4
0 2
3 2
1
1
0
0 0
8
16
24
32
40
0
8
16
cs
24
32
40
1
1
0.5
0
0 16
16
24
32
40
0
8
16
24
32
40
1.5
2
8
8
cb 0
0
0
24
32
40
0
8
16
Optimal Credit Policy
24
32
40
No Credit Policy
Fig. 6. Impulse responses under optimal credit policy compared to those under a policy of ‘‘Treasuries only,’’ in the case of a ‘‘multiplicative X’’ shock of ~ cb is exactly equal to the steady-state critical threshold. the size considered in Fig. 4, if interest-rate policy follows a Taylor rule and X
Under an optimal credit policy (shown by the solid lines), however, the responses are quite different. Central bank lending to the private sector should increase, to such an extent that there is barely any increase in credit spreads, and barely any decline in private borrowing. This requires considerable lending by the central bank, since private lending (not shown, but equal to bt Lcb t ) contracts even more (nearly twice as much) as in the absence of credit policy; the reason is that private intermediaries contract credit supply even more when they are unable to increase the interest rate at which they lend. The policy also results in practically no contraction of output, practically no decrease in inflation, practically no change in the composition of private spending, and stabilizes the economy without any need for a cut in the policy rate. If we assume instead that interest rate policy is optimal (as in Fig. 7, where the disturbance is the same as in Fig. 6), the effects of the shock on aggregate output and inflation are much smaller even in the absence of credit policy, and the effects on the relative expenditure by the two types are also somewhat smaller than those shown in Fig. 6. Nonetheless, interest rate policy alone cannot prevent an increase in credit spreads and a contraction of credit supply similar to those in Fig. 6 (even though the zero lower bound does not bind for a shock of this size), and there remains a role for credit policy. In fact, the optimal degree of lending by the central bank in response to the shock is essentially the same as in Fig. 6, and again this is the amount needed to virtually eliminate any increase in the credit spread or any reduction in private borrowing. The equilibrium under optimal credit policy is virtually the same as in Fig. 6: for while the Taylor rule is not too close an approximation to optimal policy in the absence of credit policy, there is little need for a more sophisticated interest rate policy if credit policy is used optimally. ~ cb is assumed to be no higher than X cbu,crit is the one most favorable to an argument for active credit The case in which X policy, of course. Fig. 8 shows the responses under active credit policy (and in the absence of credit policy) for the same ~ cb is assumed disturbance as in Fig. 7, and again under the assumption of optimal interest rate policy, but in the case that X to be 10 basis points higher than X
cbu,crit 40
.
It remains optimal for the central bank to begin lending to private borrowers
when a disturbance of this kind occurs, as one would expect from Fig. 4. (Recall that Fig. 4 showed Xcbu,crit rising by much t more than 10 basis points in response to a multiplicative X shock.) 40
Under our calibration, X
cbu,crit
~ cb is equal to 3.48 percent per annum, as shown in Figs. 4 and 5, so this means that X
3:58 percent.
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
73
id (level)
Y 0
0.06
3
0.04
2
0.02
1
0 0
8
16
24
32
40
0 0
8
16
24
32
40
0
8
16
24
32
40
Lcb
b 4
0
3
2
2
1
1
0
0 0
8
16
24
32
40
0
8
16
cs
24
32
40
0
8
16
24
32
40
0
8
16
24
32
40
cb
3
1.5
0
1
2 1
0.5
0
0 0
8
16
24
32
40
0
8
16
Optimal Credit Policy
24
32
40
No Credit Policy
Fig. 7. Impulse responses under optimal credit policy compared to those under a policy of ‘‘Treasuries only,’’ with the same disturbance as in Fig. 6, but ~ cb is exactly equal to the steady-state critical threshold. under optimal interest-rate policy. Again X
However, while the optimal increase in central bank lending at the time of the shock is not greatly less than in Fig. 7, the optimal duration of active credit policy is much less in this case (though the disturbance itself has the same persistence in both cases). In Fig. 7, optimal credit policy requires the central bank’s lending not to have been completely phased out six years and more after the shock; in Fig. 8, central bank lending should already have been sharply contracted in the second year, and credit policy is phased out altogether after a little more than two years. And while the effects of the disturbance on credit spreads, private borrowing, aggregate output, and the composition of expenditure are all mitigated by optimal credit policy, even relative to what can be achieved by optimal interest rate policy (again, in a case where interest rate policy is not constrained by the zero lower bound), the effects are not completely eliminated. If one assumes cbu,crit
, the optimal use of credit policy is further weakened. an event higher value of X The case considered above is also especially favorable to the optimality of active credit policy because the type of financial disturbance considered is a ‘‘multiplicative X’’ shock, which as shown in Figs. 4 and 5 is the type that increases the marginal social benefit of credit policy to the greatest extent, for a given size of increase in o t ðLÞ. If we were to consider instead an ‘‘additive w’’ shock, again of the size assumed in Fig. 4 and if, as in Fig. 8, we assume optimal interest rate ~ cb 10 basis points higher than X cbu,crit then optimal credit policy will instead involve Lcb policy and a value of X t = 0 in all periods. This can be seen from Fig. 4: a disturbance of this kind never raises the value of Xcbu,crit by as much as 10 basis t points above the steady state level.41 Yet this alternative disturbance increases the equilibrium credit spread (under ‘‘Treasuries only’’) by virtually the same amount as the disturbance considered in Fig. 8; and in fact, the size of the contraction in credit, the size of the change in the composition of expenditure, and so on, are essentially the same in this case as in the one shown in Fig. 8. Hence it matters a great deal what causes an increase in credit spreads, in order to judge cb
cbu,crit
cb
cbu,crit
41 The fact that we do assume a value of X slightly above X is, however, important for this result. If we assume that X X exactly, as in Fig. 7, then in the case of an ‘‘additive w’’ shock, it would not be optimal for the central bank to lend at the time that the shock occurs, but it would be optimal for it to commit to lend later. The optimal commitment of this kind would actually involve substantial central-bank lending, lasting for many shown in Fig. 4. But because this conclusion is radically changed by even a small increase in years—something that is not obvious from the path of Xcbu,crit t the assumed cost of lending by the central bank, we do not further discuss this fine point in the theory of optimal credit policy here.
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
74
id (level)
Y 0
0.06
3
0.04
2
0.02
1
0 0
8
16
24
32
40
0 0
8
16
24
32
40
0
8
16
24
32
40
Lcb
b 0 3
2
2 1
1
0
0 0
8
16
24
32
40
0
8
16
cs
24
32
40
0
8
16
24
32
40
0
8
16
24
32
40
cb
3
1.5
0
1
2 1
0.5
0
0 0
8
16
24
32
40
0
8
16
Optimal Credit Policy
24
32
40
No Credit Policy
Fig. 8. Impulse responses under optimal credit policy and under ‘‘Treasuries only,’’ for the same disturbance and interest-rate policy as in Fig. 7, but when cb X~ is 10 bp higher than the steady-state critical threshold.
the appropriate response of credit policy; and the aspects of the disturbance that matter cannot easily be judged from an observation of the aggregate effects of the disturbance alone. As still another example (yielding a somewhat intermediate conclusion), Fig. 9 displays the optimal policy response in the case of a ‘‘multiplicative w’’ shock, also of the size assumed in Fig. 4. Under this hypothesis, credit spreads increase owing to a change in the perceived risk of default by borrowers, but the fraction of loans expected to be bad increases with the scale of lending, at such a rate as to leave unchanged the elasticity of the credit supply curve. It is assumed in this figure, as in Fig. 8, that interest rate ~ cb is 10 basis points higher than X cbu,crit . As Fig. 4 indicates, this type of disturbance, like the policy is optimal and that X ‘‘multiplicative X’’ shock, increases the marginal social value of central bank lending enough to justify active credit policy. However, in this case, the optimal amount of central bank lending is both substantially smaller than in Fig. 8, and much more transitory: the optimal policy only involves lending to private borrowers in the first two quarters. While the credit policy does mitigate the effects of the disturbance to some extent (not too visible in the figure), the economy’s response to the disturbance is not dramatically different than it would be in the absence of credit policy: the credit spread still increases by 2 percentage points, and so on. The size, persistence and effects of optimal credit policy are similarly modest in the case of an ‘‘additive X’’ shock. As Fig. 5 has already shown, the case for active credit policy is somewhat stronger and more robust to alternative assumptions about the nature of the financial disturbance in the case of disturbances large enough to cause the zero lower bound to constrain interest rate policy, and especially if interest rate policy is purely forward looking (as in the case of a Taylor rule). For example, Fig. 10 displays the optimal policy response in the case of a ‘‘multiplicative X’’ shock that is ~ cb is now three times as large as the one considered in Fig. 8 (which is to say, of the size assumed in Fig. 5). As in Fig. 8, X assumed to be 10 basis points higher than X
cbu,crit
; but interest rate policy is now assumed to follow the Taylor rule (38). As ~ cb is the figure indicates, the zero lower bound on the policy rate does bind in the case of a shock of this size. Because X cbu,crit
higher than X , optimal credit policy does not insulate the economy from the effects of the financial disturbance to the extent that was true in Fig. 6. But it does prevent the credit spread from rising nearly as much as it would in the absence of credit policy, and averts a large part of the decline in private borrowing as well; this eliminates the declines in output and inflation almost entirely, and greatly reduces the distortion of the composition of private spending.
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
75
id (level)
Y 3
0.08 0.06 0.04 0.02 0 0.02
0
0
8
16
24
32
40
2 1 0 0
8
16
24
32
40
0
8
16
24
32
40
Lcb
b 0 0.6
2
0.4 1
0.2
0
0 0
8
16
24
32
40
0
8
16
cs
24
32
40
0
8
16
24
32
40
0
8
16
24
32
40
cb
3
1.5
0
1
2 1
0.5
0
0 0
8
16
24
32
40
0
8
16
Optimal Credit Policy
24
32
40
No Credit Policy
Fig. 9. Impulse responses under optimal credit policy and under ‘‘Treasuries only,’’ in the case of a ‘‘multiplicative w’’ shock of the size considered in ~ cb is 10 bp higher than the steady-state critical threshold. Fig. 4. As in Fig. 8, interest-rate policy is optimal and X
Figs. 11 and 12 show the corresponding optimal responses in the case of ‘‘multiplicative w’’ and ‘‘additive w’’ shocks respectively, again of the size assumed in Fig. 5. While the optimal scale and duration of lending by the central bank is considerably smaller in these two cases than in the case of the ‘‘multiplicative X’’ shock shown in Fig. 10, it is still optimal in these cases for the central bank to intervene in a substantial way for several quarters. And credit policy has important effects in each of these cases, significantly reducing the size of the spike in credit spreads at the time of the shock, and considerably weakening the contraction of spending by type b households. Our conclusions about optimal credit policy are also more robust in this case, not only in the sense that the optimal responses to the two different types of w shocks are now more similar, but (more importantly) in that our results are no longer especially sensitive to the precise magnitude of cb X~ , even in the case of the ‘‘additive w’’ shock. At the same time, the optimal responses to the w shocks shown in these last two figures are quite different from the optimal response to a ‘‘multiplicative X’’ shock. Thus even when the zero lower bound binds and interest rate policy can be described by a Taylor rule, the size of the increase in credit spreads alone provides insufficient information to judge the optimal credit policy response with much precision.
4.3. Segmented credit markets In the simple model expounded above, there is a single credit market and single borrowing rate ibt charged for loans in this market; our discussion of central bank credit policy has correspondingly simply referred to the optimal quantity of central bank lending to the private sector overall, as if the allocation of this credit is not an issue. In reality, of course, there are many distinct credit markets, and many different parties to which the central bank might consider lending. Moreover, since there is only a potential case to be made for central bank credit policy when private financial markets are severely impaired, it does not make sense to assume efficient allocation of credit among different classes of borrowers by the private sector, so that only the total credit extended by the central bank would matter. Our simple discussion here has sought merely to clarify the connection that exists, in principle, between decisions about credit policy and the other dimensions of credit policy. An analysis of credit policy that could actually be used as a basis for credit policy decisions would instead have to allow for multiple credit markets, with imperfect arbitrage between them.
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
76
id (level)
Y 0
0
3 2 1 0
0
8
16
24
32
40
0
8
16
24
32
40
0
8
16
24
32
40
Lcb
b 0
10
6 4
5
2 0
0 0
8
16
24
32
40
0
8
16
cs
24
32
40
0
8
16
24
32
40
0
8
16
24
32
40
cb
0
4
4
3 2
2
1
0
0 0
8
16
24
32
40
0
8
16
Optimal Credit Policy
24
32
40
No Credit Policy
Fig. 10. Impulse responses under optimal credit policy and under ‘‘Treasuries only,’’ in the case of a ‘‘multiplicative X’’ shock of the size considered in ~ cb is 10 bp higher than the steady-state critical threshold. Fig. 5. Interest-rate policy follows a Taylor rule and X
We do not here attempt an extension of our model in that direction. (A simple extension would be to allow for multiple types of ‘‘type b’’ households, each only able to borrow in a particular market with its own borrowing rate, and market specific frictions for the intermediaries lending in each of these markets.) We shall simply note that in such an extension, there would be a distinct first order condition, analogous to conditions (34) (35), for each of the segmented credit markets. There would be no reason to assume that the question whether active credit policy is justified should have a single answer at a given point in time: lending might be justified in one or two specific markets while the corner solution remained optimal in the other markets. Thus the main determinants of whether central bank credit policy is justified when it is justifiable to initiate active policy, and when it would be correct to phase out such programs should not be questions such as whether the zero lower bound on interest rate policy binds, or whether the central bank continues to undershoot the level of real GDP that it would like to attain. While aggregate conditions will be one factor that affects the shadow value of marginal reductions in the size of credit spreads (represented by the multiplier jo,t in (34)), the value of this multiplier will likely be different for different markets, and the main determinants of variations in it are likely to be market specific. This will be even more true of the other variables that enter into the first order condition (34). 5. Conclusions We have shown that a canonical New Keynesian model of the monetary transmission mechanism can be extended in a fairly simple way to allow analysis of additional dimensions of central bank policy that have been at center stage during the recent global financial crisis: variations in the size and composition of the central bank balance sheet, and in the interest rate paid on reserves, alongside the traditional monetary policy issue of the choice of an operating target for the federal funds rate (or some similar overnight inter bank rate elsewhere). But we have found that explicitly modeling the role of the central bank balance sheet in equilibrium determination need not imply any role for ‘‘quantitative easing’’ as an additional tool of stabilization policy, even when the zero lower bound on the policy rate is reached. While different results might be obtained under alternative theoretical assumptions, our reading of the Bank of Japan’s experience with quantitative easing leads us to suspect that our theoretical irrelevance result is likely close to the truth.
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
77
id (level)
Y 0
0
3 2 1 0
0
8
16
24
32
40
0
8
16
24
32
40
0
8
16
24
32
40
Lcb
b 0
4
6
3
4
2
2
1
0
0 0
8
16
24
32
40
0
8
16
cs
24
32
40
0
8
16
24
32
40
0
8
16
24
32
40
cb
0
6
4 3
4
2
2
1
0
0 0
8
16
24
32
40
0
8
16
Optimal Credit Policy
24
32
40
No Credit Policy
Fig. 11. Impulse responses in the case of a ‘‘multiplicative w’’ shock of the size considered in Fig. 5, under the same assumptions as in Fig. 10.
Our analysis indicates that there may, instead, be a role for central bank credit policy (or for targeted asset purchases), when private financial markets are sufficiently impaired. It is worth stressing that the central bank’s asset holdings should also be irrelevant for macroeconomic equilibrium in the case of well functioning financial markets that can be accessed at low cost by any economic agents who would benefit from such trades. Hence credit policy is only a relevant additional dimension of central bank policy to the extent that private markets are not already effectively eliminating most of the potential gains from trade in financial instruments; and while we recognize that there are circumstances, such as those arising during the recent crisis, in which it is arguable that financial markets fail to fulfill that function, we are inclined to suspect that it is only at times of unusual financial distress that active credit policy will have substantial benefits. Even when financial markets are seriously disrupted, as indicated by significant increases in interest rate spreads, one must be cautious in drawing conclusions about the welfare consequences of credit policy. While we have shown that it is possible for disturbances originating in the financial sector to create circumstances under central bank lending to the private sector can increase welfare, our analysis has also shown that the mere size of the increase in credit spreads does not provide sufficient information about the nature of the disturbance to judge the benefits of active credit policy. Moreover, while we have shown that credit policy is more likely to be justified when a financial disturbance is severe enough to make the zero lower bound a binding constraint on interest rate policy, the mere fact that the zero bound has been reached is neither necessary nor sufficient for active credit policy to be welfare improving. In particular, the appropriateness of active credit policy is likely to depend on conditions that are specific to the markets for particular financial instruments, and that therefore cannot be assessed on the basis of macroeconomic conditions alone. At the same time, when active credit policy is justified, our analysis implies that there is no need to balance the benefits of such policy for the efficiency of financial intermediation against any supposed inflationary threat inherent in the increased size of the central bank’s balance sheet. For our analysis shows that decisions about interest rate policy are not constrained in any direct way by decisions about either the size or composition of the central bank’s balance sheet, as long as the central bank is willing to adjust the interest rate paid on reserves appropriately. Payment of interest on reserves can make even a large quantity of excess reserves consistent with high short term interest rates, and hence with monetary and financial conditions consistent with a central bank’s inflation target and this will be true even when the economy is no longer characterized by any large degree of slack productive capacity.
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
78
id (level)
Y 0
0
3 2 1 0
0
8
16
24
32
40
0
8
16
24
32
40
0
8
16
24
32
40
Lcb
b 0 6
4
4 2
2 0
0 0
8
16
24
32
40
0
8
16
cs
24
32
40
0
8
16
24
32
40
0
8
16
24
32
40
cb
0
6
4 3
4
2
2
1
0
0 0
8
16
24
32
40
0
8
16
Optimal Credit Policy
24
32
40
No Credit Policy
Fig. 12. Impulse responses in the case of an ‘‘additive w’’ shock of the size considered in Fig. 5, under the same assumptions as in Fig. 10.
Thus in considering an appropriate strategy for ‘‘exit’’ from the current unconventional posture of the Federal Reserve, it is important to recognize that, according to our model, there is no reason that the timing of the Fed’s reduction in its holdings of assets other than short term Treasuries must be tied in some mechanical way to the timing of a decision to raise the federal funds rate above its current historically low level. These are independent dimensions of policy, not only in the sense that they can be varied independently in practice, but in that they have different effects as well, and are appropriately adjusted on the basis of considerations that are fairly different. Just as the primary justification for undertaking non traditional asset purchases should relate to conditions specific to the markets for those assets, rather than to the central bank’s assessment about whether the level of the policy rate is correct, so should decisions about the proper time at which to unwind such purchases.
Acknowledgments We thank Harris Dellas, Gauti Eggertsson, Marvin Goodfriend, Bob Hall, James McAndrews, Shigenori Shiratsuka, Oreste Tristani, Kazuo Ueda and Tsutomu Watanabe for helpful discussions, Ging Cee Ng for research assistance, and the NSF for research support of the second author. The views expressed in this paper are those of the authors and do not necessarily reflect positions of the Federal Reserve Bank of New York or the Federal Reserve System. Appendix A. Supplementary data Supplementary data associated with this article can be found in the online version at doi:10.1016/j.jmoneco.2010.09.011. References Ashcraft, A., Garleanu, N., Pedersen, L.H., 2010. Two monetary tools: interest rates and haircuts. Unpublished, Federal Reserve Bank of New York. Auerbach, A.J., Obstfeld, M., 2005. The case for open-market purchases in a liquidity trap. American Economic Review 95, 110–137. Baba, N., Nakashima, M., Shigemi, Y., Ueda, K., 2006. The bank of Japan’s monetary policy and bank risk premiums in the money market. International Journal of Central Banking 2, 105–135.
´ rdia, M. Woodford / Journal of Monetary Economics 58 (2011) 54–79 V. Cu
79
Benigno, P., Woodford, M., 2005. Inflation stabilization and welfare: the case of a distorted steady state. Journal of the European Economic Association 3, 1185–1236. Bernanke, B.S., 2009. The crisis and the policy response. Stamp Lecture, London School of Economics, January 13, 2009. [Text available at /http://www. federalreserve.gov/newsevents/speech/bernanke20090113a.htmS]. Cu´rdia, V., Woodford, M., 2009. Credit frictions and optimal monetary policy. Unpublished, Federal Reserve Bank of New York. Cu´rdia, V., Woodford, M., 2010a. Conventional and unconventional monetary policy. Federal Reserve Bank of St. Louis Review 92 (4), 229–264. Cu´rdia, V., Woodford, M., 2010b. The central-bank balance sheet as an instrument of monetary policy. NBER Working Paper No. 16208. Cu´rdia, V., Woodford, M., 2010c. Credit spreads and monetary policy. Journal of Money, Credit and Banking 42 (s1), 3–35. Eggertsson, G.B., Woodford, M., 2003. The zero bound on interest rates and optimal monetary policy. Brookings Papers on Economic Activity 2003 (1), 139–211. Friedman, M., 1969. The optimum quantity of money. In: The Optimum Quantity of Money and Other Essays, Aldine, Chicago. Gagnon, J., Raskin, M., Remache, J., Sack, B., 2010. Large-scale asset purchases by the federal reserve: did they work? Federal Reserve Bank of New York Staff Report no. 441. Gertler, M., Kiyotaki, N., 2010. Financial intermediation and credit policy in business cycle analysis. In: Friedman, B.M., Woodford, M. (Eds.), Handbook of Monetary Economics, vol. 3. Elsevier, Amsterdam. Goodfriend, M., 2002. Interest on reserves and monetary policy. Federal Reserve Bank of New York Economic Policy Review, 77–84. Goodfriend, M., 2009. Central banking in the credit turmoil: an assessment of federal reserve practice. Unpublished, Carnegie-Mellon University. Goodfriend, M., King, R.G., 1997. The new neoclassical synthesis and the role of monetary policy. NBER Macroeconomics Annual 12, 231–283. Mehra, R., Piguillem, F., Prescott, E.C., 2008. Intermediated quantities and returns. Federal Reserve Bank of Minneapolis Research Department Staff Report no. 405. Oda, N., Ueda, K., 2007. The effects of the bank of Japan’s zero interest rate commitment and quantitative monetary easing on the yield curve: a macrofinance approach. The Japanese Economic Review 58, 303–328. Okina, K., Shiratsuka, S., 2004. Policy commitment and expectation formation: Japan’s experience under zero interest rates. North American Journal of Economics and Finance 15, 75–100. Reis, R., 2009. Interpreting the unconventional U.S. monetary policy of 2007–09. Brookings Papers on Economic Activity 2009 (2), 119–165. Sarkar, A., 2009. Liquidity risk, credit risk, and the federal reserve’s responses to the crisis. Financial Markets and Portfolio Management 23, 335–348. Sarkar, A., Shrader, J., 2010. Financial amplification mechanisms and the federal reserve’s supply of liquidity during the crisis. Federal Reserve Bank of New York Staff Report no. 431. Shiratsuka, S., 2009. Size and composition of the central bank balance sheet: revisiting Japan’s experience of the quantitative easing policy. Bank of Japan IMES Discussion Paper no. 2009-E-25. Stroebel, J.C., Taylor, J.B., 2009. Estimated impact of the fed’s mortgage-backed securities program. Unpublished, Stanford University. Taylor, J.B., 2009. The financial crisis and the policy responses: an analysis of what went wrong. NBER Working Paper no. 14631. Ueda, K., 2009. Non-traditional monetary policies: G7 central banks during 2007–2009 and the bank of japan during 1998–2006. CARF Working Paper no. F-180, University of Tokyo. Ugai, H., 2007. Effects of quantitative easing policy: a survey of empirical analyses. Bank of Japan Monetary and Economic Studies 25, 1–47. Walsh, C.E., 2009. Using monetary policy to stabilize economic activity. Financial Stability and Macroeconomic Policy. Federal Reserve Bank of Kansas City, Kansas City. Wallace, N., 1981. A Modigliani–Miller theorem for open-market operations. American Economic Review 71, 267–274. Woodford, M., 2003. Interest and Prices: Foundations of a Theory of Monetary Policy. Princeton University Press, Princeton.
Journal of Monetary Economics 58 (2011) 80–82
Contents lists available at ScienceDirect
Journal of Monetary Economics journal homepage: www.elsevier.com/locate/jme
Discussion
Comment on: ‘‘The central-bank balance sheet as an instrument of monetary policy’’ Harris Dellas a,b, a b
Department of Economics, University of Bern, Switzerland CEPR, London, UK
a r t i c l e i n f o Article history: Received 22 September 2010 Received in revised form 27 September 2010 Accepted 28 September 2010 Available online 25 October 2010
One definition of an economist is somebody who sees something happen in practice and wonders if it will work in theory Ronald Reagan
1. Motivation The recent conduct of monetary policy in the US (as well as in some other industrial countries such as the UK) has deviated significantly from past behavior. In addition to setting the interest policy instrument to virtually zero for a long period, the FED has engineered massive increases in the size of its balance sheet. For instance, its balance sheet more than doubled in 2009 and shows little sign of retrenchment since. At the same time, the composition of the central banks’ assets has changed dramatically, away from ‘‘safe’’ government securities towards riskier private securities (mortgage backed securities, etc.). The FED has also started paying interest on reserves and the amount of reserves held by depository institutions at the FED has increased by a factor of 100. All these unusual measures have been taken in order to combat the consequences of the recent financial crises. It would be of great interest to study their possible macroeconomic implications within an established macroeconomic model, in particular with regard to their potential to mitigate the impact of financial shocks. And also to address the question of whether such policies are a good idea from a welfare point of view, and, if yes, what are the conditions that make them a good idea. 2. The framework In their interesting paper, Curdia and Woodford address these two questions in the context of a New Keynesian model that includes financial intermediation, financial frictions and non trivial agent heterogeneity. The model contains some DOI of original article: 10.1016/j.jmoneco.2010.09.011
Correspondence address: Department of Economics, University of Bern, Switzerland.
E-mail address:
[email protected] URL: http://www.harrisdellas.net 0304-3932/$ - see front matter & 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jmoneco.2010.09.010
H. Dellas / Journal of Monetary Economics 58 (2011) 80–82
81
commonly used features and also some novel elements. The novel element is to be found in the specification of non trivial heterogeneity as the source of borrowing and lending. Agents differ in their spending opportunities which leads to borrowing and lending in equilibrium. The more conventional part of the model is that the borrowing and lending is required to be done through financial intermediaries because of unspecified frictions. That the intermediaries fund themselves by issuing debt. And that financial intermediation is ‘‘privately’’ costly as an exogenous fraction of bank loans is not repaid. It is also ‘‘socially’’ costly as it requires resources in order to take place. The last two considerations imply a spread between bank borrowing and lending rates. Central bank liabilities (the monetary base in this model) have two uses. They can be used to purchase government bonds. And to fund directly that is, without going through the financial intermediaries loans to the households. The latter activity uses up resources, thus state alike private banking is costly. There are three independent dimensions of monetary policy in the model. The choice of the quantity of money created by the central bank. Its allocation between government bonds and loans. And the interest rate to be paid on money (reserves) held by the financial intermediaries. The total quantity of money created (the size of the balance sheet of the central bank) matters for real allocations here even in the absence of price stickiness because of the financial friction and also due to the fact that central bank liabilities carry non pecuniary benefits: Holding them as reserves allows banks to mitigate the resource cost associated with private banking activities. The former element also makes the composition of the central bank portfolio matter. The model is made tractable in spite of non trivial heterogeneity via some clever modelling choices. Its solution takes a simple and intuitive form that corresponds to the solution of the standard New Keynesian model augmented with terms that highlight the role of the two key elements: heterogeneity cum financial friction. And non pecuniary benefits for central bank liabilities. There are two main results. 3. The main results The first concerns the optimal supply of reserves (the size of the balance sheet of the central bank). Reserves reduce the cost of financial intermediation and also ameliorate the financial friction through their effect on credit spreads. At the same time they cost nothing to produce. Consequently, optimal reserve policy entails the satiation of banks with reserves. This requires driving the opportunity cost of holding reserves, that is, the interest rate differential between the bank borrowing rate and the rate on reserves, to zero. Of course, this is simply the Friedman rule on the optimum quantity of money with the caveat that it only requires a zero interest rate differential, not a zero level of nominal interest rates. An implication of the analysis is that the provision of reserves beyond the satiation point holding central bank lending to the households fixed affects neither output nor welfare. Consequently, quantitative easing (defined as an expansion of the supply of outside money beyond the satiation point that does not involve the purchase of non government securities) serves no useful purpose. This is an interesting and useful application of the optimum quantity of money that serves to clarify the debate on the concept and implications of quantitative easing. The second main result concerns the portfolio of the central bank (for a given size). The authors find that there may exist situations in which it is optimal for the central bank to extend directly loans to the private sector thus deviating from the standard treasuries only policy. And that this is more likely to be the case when the zero bound on interest rates has been reached and the shock is placing the economy far away from its steady state. However, no simple relationship seems to exist between movements in interest rate spreads (borrowing lending rates) and the desirability of ‘‘state banking’’. In particular, the properties of central bank lending to the private sector (its size and duration) vary much with the type of financial shock considered. 4. Evaluation This is an excellent, well written paper. It has an important objective: To use theory to make sense of the unconventional monetary policies being practiced nowadays. To this purpose, it proposes a minimalistic (in terms of departure from the standard New Keynesian model developed by Woodford, 2003), yet rigorous, clear and pedagogically useful framework. This framework proves quite useful for studying the macroeconomic implications of financial shocks and for characterizing informatively and succinctly the macroeconomic and welfare properties of observed monetary policy actions. In my view, it is likely to serve as the new canonical ‘‘New Keynesian’’ model and used extensively in particular in the context of the optimal design of policy in financially distorted economies. Having a tractable, well defined measure of ex ante welfare in the presence of non trivial heterogeneity (lenders borrowers) represents a major technical and substantive innovation. 4.1. Limitations 1. The paper provides examples in which state banking (active central bank credit policy) is not only optimal but it can also significantly ameliorate the macroeconomic effects of adverse financial shocks. It has thus, in principle, the potential to justify unconventional monetary policy. But can its findings be used to provide cover for the recent actions of the FED, as
82
H. Dellas / Journal of Monetary Economics 58 (2011) 80–82
well as those of similarly behaving central banks (note that the authors do not make any claim of this sort)? My feeling is that the answer to this question is twofold: We do know; and probably no. First, the ‘‘we do not know’’ part of the answer. The paper also provides examples in which active state banking is not optimal or effective. The problem is that these two groups of examples (in favor and against state banking) are differentiated on the basis of assumptions about curvatures of functions and shock processes for which we know as yet next to nothing. For instance, the differences depend on (a) the marginal cost of central bank lending; (b) whether the source of the increase in spreads comes from an additive or multiplicative shock to the cost of private lending and also which of the components of this cost is affected; (c) the response of equilibrium credit spreads to an increase in private lending. While one can view this sensitivity in a positive light (as a stimulus for empirical research on these issues), it does constitute a serious problem. First, because some of the required ingredients may not have empirically identifiable counterparts (for instance, the marginal cost of central bank lending). And second, independent of whether progress can be made in the future in identifying some of these parameters, the lack of even educated guesses about their values at present makes it difficult to evaluate the paper’s qualitative analysis (and much more so its quantitative analysis). Relatedly, other important properties of the model (for instance, the duration of activism) also depend critically on these features as well as on assumptions about how monetary policy is conducted along the other dimensions discussed above. Consequently, while the paper demonstrates the possibility of optimal unconventional policy it does not shed much light on whether such policies are probably optimal as they are practiced and under the circumstance they are practiced in. And also on whether the pursuit of these policies may have made a difference for recent macroeconomic performance. And second, with regard to the ‘‘probably no’’ part of the answer. The model abstracts from aspects of financial markets that seem important for the desirability and nature of central bank activism. The most important elements here are: (a) The assumption of symmetric information. This matters for the evaluation of the policies studied as anticipation of state banking may lead to moral hazard and excessive risk taking by the private sector (endogenous default). The expectation of the acquisition of bank or other toxic assets could bias private lending as well as investment choices in socially undesirable directions. (b) The assumption that state lending is riskless. While this is inconsequential in the model due to the exogeneity and the absence of social costless of default, it is likely to be an important factor in the real world in a variety of distorted environments (for instance, when taxes are distortionary). (c) The risk that unconventional monetary policy could be captured by politicians and special interest groups. Before offering concluding remarks let me also make two other smaller points. I wonder first whether similar implications regarding the optimality and effectiveness of unconventional policy would not also obtain in a representative agent model with costly banking and firms (rather than consumers) being the recipients of bank loans. And second, while the joint use of two monetary policy instruments (interest on reserves and reserve supply policy) is a very clever way to get around the optimal deflation implication associated with the Friedman rule I find this result quite fragile. It would not survive simple extensions such as the existence of cash outside banks and/or a transactions role for bank deposits. In conclusion: I find the paper by Curdia and Woodford to represent one of the best analyses of unconventional monetary policy that are available in the literature. It is innovative, tractable, clear and full of useful insights. I think that it will serve as the canonical version of New Keynesian model with financial frictions. Reference Woodford, M., 2003. Interest and Prices: Foundations of a Theory of Monetary Policy. Princeton University Press, Princeton.
