Skip to Main Content

Paying Interest on Bank Reserves

Honoring Marvin Goodfriend
May 2022

Goodfriend, Marvin. 2002. "Interest on Reserves and Monetary Policy." Federal Reserve Bank of New York Economic Policy Review 8, no. 1 (May): 77-84.

Since the 2008 financial crisis, the Fed's main tool for exercising control over short-term interest rates has been the rate it pays on the now large reserve balances held by banks. This new approach marks a substantial change from the Fed's previous operating regime, in which the supply of reserves was tightly controlled to ensure that the equilibrium rate in the federal funds market was at, or near, the Fed's target. In 2002, well before this change and before the idea was under active consideration by the Fed, Marvin Goodfriend set out a proposal for such an approach to policy operations in the Federal Reserve Bank of New York's Economic Policy Review. This proposal served as a cornerstone for planning and discussions around the System leading up to implementation of an interest on reserves mechanism in October 2008. While interest on reserves, as implemented in the US, has yielded some surprises, Marvin's discussion remains an important benchmark.

In this essay, we review Marvin's 2002 proposals and provide some perspectives on the evolution and implementation of those ideas. In particular, we recount the 15-year history of interest on reserves at the Fed, showcasing the influence of Marvin's ideas. We document how those ideas were adapted as policymakers learned from the actual implementation of policy.

The issues identified by Marvin in the early 2000s remain central today to the assessment of an operating regime based on the management of interest on reserves. One particular set of issues — having to do with preserving independence of monetary policy from political motivations — has only grown in relevance.

Interest on Reserves

Bank reserves are balances held by depository institutions in their accounts at the corresponding Reserve Bank. Reserves are freely convertible, dollar-for-dollar, into currency and play a critical role in the settlement of financial transactions intermediated by banks. Banks trade reserves in the fed funds market.

The idea of paying interest on bank reserves first arose as a tool for reducing the distortions from the tax on money (as modeled by Lacker [1997]). Central bank money that is dominated in rate of return will still be held by economic agents because of its privileged role in the payment system. Eliminating the tax on currency is the motivation behind the Friedman rule.2 In his Program for Monetary Stability, Friedman (1959) also proposed paying interest on reserves held by banks.3

Evolution of an idea: implications for conduct of monetary policy

Goodfriend's (2002) paper goes beyond this public finance perspective to a concern for the operational conduct of monetary policy. At the time of his writing, major central banks around the world had adopted the approach of conducting monetary policy by manipulating short-term interest rates and reserves earned zero interest. The supply of reserves was one of the main levers used for affecting short-term interest rates. By contrast, Marvin's proposal was to peg the fed funds rate by paying a positive interest on reserves and assuring a plentiful reserve supply. An important argument in Goodfriend (2002) is that paying interest on reserves could allow the central bank to separately manage short-term interest rates and the supply of monetary liabilities.

Goodfriend (2002) traces his thinking on this topic to his earlier 2000 paper on monetary policy at the lower bound on nominal interest rates,4 which proposes supplementing a policy rate set at its effective floor with the expansion of reserves. With the short-term rate at its lower bound and reserves abundant, he argued, the convenience yield on money goes to zero — making reserves and fed funds lending perfect substitutes. With the demand for reserves thus satiated, the banking system will hold whatever quantity the Fed supplies.5

At the lower bound, Goodfriend characterized expanding the supply of reserves as an alternative means of policy accommodation when interest rate reductions are off the table. More generally, in his interest-on-reserves (IOR) paper,6 he argued that interest rate and reserves policy could operate independently, serving different objectives. By setting its interest rate on reserves, the Fed could conduct interest rate policy in much the same way central banks had grown accustomed to doing, systematically responding to economic conditions to maintain price stability.

Separation of interest rate and balance sheet policy

Importantly, once reserves were large enough to ensure that demand is satiated at the IOR rate, further adjustments to reserve supply could respond to liquidity conditions in broader money markets. Marvin dubbed this "managing the supply of broad liquidity."7 Here, he recognized the possibility of frictions affecting the efficiency with which market participants might exchange alternative short-term instruments. Importantly though, his view of the central bank's role in responding to such frictions was limited to managing the overall supply of liquidity by creating reserve balances through open market operations. Inherent in this view is a belief that markets will do a reasonably good job of distributing the liquidity that the central banks supply — money market problems are adequately addressed by managing that overall supply.8

The separation of interest rate and balance sheet policy was an important part of Marvin's vision for an IOR operating regime. Marvin, of course, had made significant contributions to our understanding of how modern central banks achieved and maintained low inflation by systematically adjusting short-term rates in response to macroeconomic conditions.9 So being able to continue reacting to the economy in the same manner was vitally important to him. Note that an implication of the separation of interest rate and balance sheet policy is that if the Fed expanded reserves in response to a perceived need for liquidity in money markets, there would be no expectation that interest rate moves — in particular, rate increases, if economic conditions warranted them — would need to wait for an unwinding of the reserves expansion. This observation is consistent with the Fed's approach to allowing a very gradual balance sheet run-off while at the same time raising interest rates.

Desirable features of an IOR regime

One particular virtue of an IOR operating regime, according to Goodfriend (2002), is that central bank interest rate control would be more robust to technological changes in the payment system that reduce the demand for central bank balances as a means of settlement. Marvin argued that the traditional Fed approach could become increasingly difficult in the face of such changes. This is because the traditional approach involved manipulating the supply of reserves so that the market-determined fed funds rate settled near the target. By contrast, in a regime in which the opportunity cost of holding reserves was eliminated, banks would be willing to hold the reserves supplied by the Fed, independently of the role of reserves in payment settlement.

Goodfriend (2002) also noted that the abundant reserves in such an operating regime could reduce the need for central bank credit that can arise when banks face unexpected payment flows. Historically, in the US interbank system, such credit extensions took the form of both intraday overdraft allowances and overnight discount window loans.10 Goodfriend had elsewhere expressed the view that modern money markets could function with less reliance on central bank credit.11 So, he saw a reduced reliance on Fed lending as a means to solidify the separation of monetary and credit policies. The IOR regime, then, could also serve as a means of facilitating that goal.

IOR and Fed income

One potential difficulty that Goodfriend (2002) anticipated involved the effects of reserve remuneration and larger Fed balance sheets on the Fed's income.12 However, he expected that under normal conditions, the Fed would continue to earn a positive spread on its balance sheet. The Fed's assets would have an average maturity well above the overnight maturity of reserves and so would usually bear a yield greater than IOR. Also, he took as given that (non-interest-bearing) currency would remain a nontrivial part of the Fed's liabilities for the foreseeable future. While the Fed's net interest margin would be reduced by paying a positive rate on reserve balances, that spread would be earned on an expanded balance sheet.

So Goodfriend did not see paying interest on reserves as a fundamental problem for the Fed's income, on average. That said, he did see the possibility of periods with negative Fed income when appropriate monetary policy called for relatively quick and significant increases in short-term interest rates.13 He was nervous that such episodes would mean that the Fed would have to go to Congress or the Treasury to obtain funding to cover operating costs. That is, such a situation could threaten the political independence of monetary policy. While acknowledging the risk, he saw it as manageable so long as the Fed carried an appropriately sized capital buffer or surplus account.14

Interest on reserves at the Fed

In 2006, Congress enacted a collection of regulatory relief measures for banks and other financial institutions. As part of this package — the Financial Services Regulatory Relief Act of 2006 — the Fed was granted authority to pay interest on the reserves held by banks.15 Part of the motivation for this provision was to reduce banks' incentives to engage in costly account management practices that served only to reduce reserve requirements. This practice involved automated procedures for shifting customer overnight balances out of reservable deposit accounts and into other instruments that did not carry reserve requirements — so-called sweep accounts. Changes in sweep account practices were a focus of the Congressional Budget Office's (CBO) analysis of the likely implications of this provision for the federal budget.16

Key aspects of the 2006 act

The language in the act (Section 201) is very brief, but there are a few specific conditions worth noting. First, the possibility of interest on reserves was not made available to all holders of reserve balances with the Fed — only to depository institutions. Importantly, this left out government sponsored enterprises (GSEs) such as Fannie Mae, Freddie Mac, and the Federal Home Loan Banks. These entities hold significant balances with the Fed and are typically lenders of overnight funds in the fed funds market and other segments of the short-term money market. The distinction between banks and GSEs would prove significant for the eventual workings of the IOR regime in ways not fully anticipated.

Second, the act specified that the rate paid by the Fed was not to exceed the general level of short-term market rates. At the time, this seemed like an innocuous requirement. In the model Goodfriend used to discuss his proposal,17 the rate paid by the central bank essentially becomes the market rate. This is because the demand for reserves is satiated, and interest-bearing central bank balances replace interbank loans for the most part. A market rate above the rate on reserves would tend to arise in other systems in which reserves remain scarce and there is still an active interbank market. So, standard analysis at the time did not foresee the possibility of markets rate below IOR. Indeed, the CBO's analysis assumed a rate on reserves that was on average 10 to 15 basis points below overnight money market rates. The experience since implementation of IOR has been quite different, in large part due to the behavior of the GSEs in the money markets.

Finally, the 2006 act gives the authority to set the rate paid on reserves to the Federal Reserve Board of Governors, as opposed to the monetary policymaking body — the Federal Open Market Committee (FOMC). Congress likely saw the provision as more of an operational modernization — to reduce adverse incentives created by the tax on reserves — than a fundamental change in the way monetary policy would be implemented.18 Yet, in the proposal put forward by Goodfriend (2002), the rate on reserves becomes the instrument for interest rate policy. The act's designation of decision-making authority creates the need for potentially delicate coordination among different bodies within the Federal Reserve System.

Consideration by a Fed workgroup

The 2006 act set 2011 as the start date for the Fed's new authority, giving the Fed time to work out the details of using this tool. Shortly after its enactment, the Fed created a System workgroup, consisting of staff from the Board of Governors and the Reserve Banks, to study the issue and prepare a number of alternatives to be considered by policymakers. A central question in designing an implementation regime was whether the Fed would continue to target an interbank rate by manipulating the supply of (relatively scarce) reserves. Doing so would result in a so-called corridor system, in which the target rate was above the rate paid by the Fed (and typically below the rate charged by the Fed for discount window loans).

The main alternative to this approach, as proposed by Goodfriend, is typically referred to as a floor system. In such a system, the Fed would provide enough reserves to essentially eliminate the need for interbank lending. Movements in short-term interest rates would be driven by the rate on reserves, and moderate fluctuations in the quantity of reserves would have little to no effect on those short-term rates.

The workgroup examined the relative merits of floor and corridor systems (as well as some hybrids). A technical appendix to the workgroup report provided an analytical framework for this comparison (and was later published as Ennis and Keister [2008]). The workgroup also devoted attention to an array of technical details, including the mechanics of monitoring reserve requirements and the treatment of different categories of reserves (required, clearing balances, and excess reserves).19

Accelerating the implementation of IOR

The Fed intended a long deliberative process before it implemented IOR. But by the time the workgroup presented its analysis to policymakers in early 2008, attention had shifted to dealing with the unfolding financial crisis. The Fed's crisis response led directly to an accelerated adoption of IOR. Beginning in late 2007, the Fed dramatically expanded the provision of central bank credit in various forms. In the interest rate targeting regime it maintained at the time, the Fed could not let credit expansions increase the size of its balance sheet and the supply of reserves. Accordingly, it offset its growing book of loans by selling Treasury securities from its portfolio (and in that way sterilizing that credit growth). When the crisis deepened in October 2008, around the time of the failure of Lehman Bros. and the bailout of AIG, the rapid rise in credit extensions overwhelmed the Fed's ability to sterilize — it risked running out of Treasuries to sell.

Partly to address this issue, the Fed asked for and received permission to accelerate its authority to pay interest on reserves — and in that way control interest rates without sterilizing the credit programs.20 In October 2008, then, the Fed started paying interest on reserves. The conditions under which this happened were not ideal for a test of using the rate paid on reserves to target and control market rates. With the extreme market volatility at the time, and with the Fed both rapidly expanding its balance sheet and taking its interest rate target to its effective lower bound, the choice between a corridor and a floor soon became trivial. Indeed, with the beginning of quantitative easing programs in 2009, it became clear that it would be some time before there was a relevant option to make reserves anything other than abundant.

A surprise: the overnight rate below the floor

As the Fed entered the regime of interest paid on abundant reserve balances, something else became clear. The theoretical floor on market overnight rates provided by the rate on reserves turned out not to be a firm floor. When the Fed moved its policy rate to the effective lower bound, it started targeting a range instead of a given number. From December 2008 until December 2015, the FOMC target range for the effective fed funds rate was 0 to 25 basis points and IOR was fixed at the top of the policy range. During this period, the effective fed funds rate consistently traded below IOR by, on average, 10 to 15 basis points. And, while theory would predict that with abundant reserves fed funds trading would vanish, activity did decline significantly but did not disappear.

These surprising facts are due primarily to the participation of the GSEs in the fed funds market. Since they could not earn interest on their reserve balances, there was an opportunity for them to lend those balances to banks that could earn interest from the Fed. While one would expect competition among borrowing banks to bid the rate on those loans up close to the rate on reserves, at least two market features are thought to have limited this arbitrage. First, a bank borrowing reserves from a GSE to earn overnight interest incurs costs from expanding its balance sheet. One direct source of costs is that banks pay FDIC premiums on the overall size of their balance sheet (not just on their insured deposits). This places a cap, below the yield on reserves, on what a bank would be willing to pay to a GSE. Second, competition may not even have driven fed funds loan rates as far as the ceiling implied by the cost of balance sheet expansion. One explanation for this is that the GSEs, holding market power in fed funds lending, were restrictive in which banks they would lend to and how much. The resulting small number of bidders, then, resulted in imperfect competition that may have further held down the effective fed funds rate.21

The prospects of liftoff and a new facility

Still, for the first several years of IOR, there really was no opportunity to assess the Fed's ability to conduct interest rate policy exclusively (or primarily) through manipulating the rate paid on reserves. Eventually, though, the likelihood of a rate increase being necessary became more apparent. In 2012, when the Fed began including the policy interest rate as one of the variables in the Summary of Economic Projections, a majority of participants anticipated that some rate increases would be appropriate by the end of 2015. As the FOMC discussed prospects for lifting its target rate off of its effective floor, some members and senior staff expressed uncertainty as to whether simply lifting the rate on reserves would reliably pull up the broader array of short-term rates that were seen as important for affecting economic activity. This lack of confidence ultimately led to the creation of the Overnight Reverse Repurchase Agreement facility, or ON RRP.22

The ON RRP facility allows an expanded set of counterparties to exchange excess cash for securities held by the Fed in an overnight transaction that pays a rate to the cash lender that is a bit below the rate paid on bank reserves. The expanded set of counterparties includes, importantly, the GSEs that are also active lenders in the fed funds market. Their access to the ON RRP was intended to place a more reliable floor under the fed funds rate — as a GSE would presumably not want to lend reserves to a bank at less than it could earn by doing a repo transaction with the Fed.

The liftoff and transition to a new policy

When the FOMC did eventually begin raising rates near the end of 2015, market rates generally followed its increases in the rate paid on reserves.23 As to whether the extra precaution of the ON RRP proved to be a necessary tool for interest rate control, the evidence is somewhat mixed. Most of the time, overnight market rates remained above the ON RRP floor. But at ends of months, it was common for market rates to fall to that floor. Correspondingly, those dates would see significant pick up in volume for the Fed's ON RRP facility. This pattern appears to have been due to incentives created by calendar-based reporting requirements for many of the financial institutions that participate in these markets.

Once rate increases were underway, the FOMC returned to the question of how it would implement its interest rate policy over the longer run. Would it maintain the abundant reserves regime that had emerged during the crisis response, in which the rate paid on reserves is the main tool for influencing market rates? Or would it move to a regime in which the supply of reserves was more restricted, so that an active interbank funds market would emerge, and market rates would depend both on the rate paid by the Fed and the supply of reserves? And what would be the implications of this choice for the size of the Fed's balance sheet?

These discussions led to the January 2019 "Statement Regarding Monetary Policy Implementation and Balance Sheet Normalization." There, the committee stated its intent to continue to operate with an "ample" supply of reserves, so that rate control would be achieved "primarily through setting the Fed's administered rates." It also said that it would continue to use the target range for the federal funds rate to express the stance of its interest rate policy.

Relative to the approach proposed by Goodfriend, this use of the funds rate range might seem like an unnecessary complication. Why not simply use the administered rate — the rate paid on reserves — as the policy rate? Here, the rate governance created by the legislation allowing interest on reserves might have come into play. Recall that the enabling act gave the Fed's Board of Governors the authority to set the rate on reserves. The January 2019 statement, by contrast, emphasized the rate decision made by the FOMC. This statement can be seen as a reassurance that adopting a regime in which administered rates are the main tool does not move authority for interest rate policy from the FOMC to the Board of Governors.

After describing its plan for how it would proceed to implement monetary policy, the committee made a series of announcements about the intended size of its balance sheet and the supply of reserves. The Fed intended to allow the balance sheet to run down through the redemption of securities without reinvestment of the proceeds. This process was made gradual and predictable by capping the amount of unreinvested redemptions. The Fed's announcements generally avoided projecting a specific number for the ultimate size of its balance sheet. They did indicate that the average level of reserves would likely be "somewhat above the level of reserves necessary to efficiently and effectively implement monetary policy."24 In the context of the January 2019 statement on implementation plans, an interpretation is that reserves supply would remain large enough so that the fed funds rate would remain at or below the rate paid on reserves.

Turbulence and renewed expansion of the balance sheet

Turbulence in money markets in September 2019, including spikes in repo rates and in the fed funds rate, led some to question whether the Fed had already taken the supply of reserves too low relative to its stated intentions. As a result, the Fed ended its runoff of the balance sheet and began a modest pace of growth in reserves. In its October 2019 announcement, the Fed included a plan to build back reserves at least to their level before the market stress of September. This plan was intended to extend through "at least the second quarter" of 2020. The Fed did not change the character of its long-run plans (as expressed earlier in 2019), but perceptions about how large the supply of reserves would need to be on an ongoing basis did change some.

Monetary policy implementation after COVID-19

The Fed's intent remained to operate with reserves no larger than necessary to effectively and efficiently conduct monetary policy.25 Yet, money markets events in September 2019 produced a reassessment of the desired long-run level of reserves. Further, with the onset of the COVID-19 pandemic in early 2020 and the economic disruption that followed, the normalization process for the Fed's balance sheet was again interrupted. The uncertainty caused by the public health crisis led many money market yields to rise as investors sought to hold only the safest, most liquid assets (in a so-called "dash for cash" episode). The Fed's response included an array of credit facilities as well as a substantial increase in the pace of asset purchases. At the same time, the Fed brought the interest rate paid on reserves to its lowest level to date.

The Fed's pandemic actions brought its balance sheet size to new highs. Recently, take-up in the ON RRP facility also rose to more persistently high levels.26 As of this writing (January 2022), the Fed has begun to taper its asset purchases, but there has been little discussion as yet about when the Fed will begin to let its net holdings of assets run down and whether the envisioned long-run level of reserves remains the same as it did pre-pandemic.

A new repo facility

In the wake of the pandemic, the Fed has added another standing facility to its toolbox for interest rate control. It created the Standing Repurchase Agreement (repo) Facility (SRF) that lends cash overnight to a broad set of counterparties against Treasury and agency securities. By lending at a rate somewhat above the rate on reserves, this facility is expected to limit the kinds of money market rate spikes experienced in September 2019. It is still an open question whether the SRF will become heavily used when the Fed normalizes the size of its balance sheet. Interestingly, earlier discussions of a standing repo facility included the argument that such a facility would permit the Fed to operate effectively even with a lower level of reserves.27 It seems unclear that this is the current intent of the SRF.

So, the current regime is somewhat more complicated than the one laid out by Goodfriend in 2002, which saw rate control as being adequately accomplished by managing the rate paid on reserves. Now, out of an abundance of caution, the Fed has added supports in the form of standing facilities in the repo market.28 While we of course cannot be sure, it seems likely to us that the supply of reserves will remain for some time well above levels that Marvin would have viewed as necessary for implementing an effective IOR regime.29


As on many topics related to modern central banking, Goodfriend's discussion of interest on reserves as a tool for monetary policy has proved to be prescient. The issues he identified as potential areas

of concern have all remained relevant, and his thinking has influenced the learning process of policymakers every step of the way. For instance, his consideration of the implications for the demand for reserves of payment innovations — written at a time when an array of payment tools now common were barely in use, if at all — identified legitimate concerns that are being voiced today by those studying the merits of a central bank digital currency.30 His conclusion — that an IOR regime diminishes the concern by making reserves an interest-bearing asset and hence rendering their payment settlement role less dominant — suggests that monetary control might survive the advent of digital currencies.

It is important to stress that Goodfriend's proposal was offered in a context where the credibility of the central bank was unquestioned. In such an environment, inflation expectations are firmly anchored. But when there is more uncertainty about the central bank's goals and its likely conduct of interest rate policy in pursuit of those goals — as some have argued coming out of the pandemic crisis — the possibility of drifting inflation expectations can be more of a concern. In this context, as Ennis and Wolman (2010) noted, the consequences of an interest rate policy that falls "behind the curve" can be magnified by the existence of large excess reserve balances. This provides another reason to be mindful of the overall size of reserve balances.

Among the most salient issues discussed by Goodfriend (2002) are those related to the possibility of negative Fed net income, especially as that might affect its independence to conduct appropriate monetary policy.31 The Fed's relationship to the legislative and executive branches of government was a central concern in much of Goodfriend's work.32 While in his 2002 article he saw the income and expense challenges created by interest on reserves as being manageable, his vision was certainly of a much smaller balance sheet. He also advocated a regime in which the Fed's intervention in money markets was simpler than one with multiple standing facilities. Again, we cannot be sure, but we think that on those issues our insightful and experienced colleague would be more concerned now than he was in 2002.

As a larger balance sheet and greater involvement in the money markets raise the risks to Fed independence, another common theme from Marvin's work is worth remembering. In many instances, Marvin advocated creating formal understandings between the Treasury and the Fed delineating the central bank's autonomy and accountability. He would often link these proposals to the 1951 Fed-Treasury Accord that ultimately allowed the Fed to conduct the independent interest rate policy to which we have become accustomed.33 As the Fed's operating regime expands in its reach, an accord that establishes reasonable and commonly understood limits to Fed activities could be a useful tool.34


Andolfatto, David, and Jane Irhig. 2019. "Why the Fed Should Create a Standing Repo Facility." On the Economy Blog, Federal Reserve Bank of St. Louis, March 6.

Bech, Morten L., and Elizabeth Klee. 2011. "The Mechanics of a Graceful Exit: Interest on Reserves and Segmentation in the Federal Funds Market." Journal of Monetary Economics 58, no. 5 (July): 415-431.

Benigno, Pierpaolo. 2019. "Monetary Policy in a World of Cryptocurrencies." Centre for Economic Policy Research Discussion Paper 13517, February.

Bowman, David, Michiel De Pooter, Etienne Gagnon, and Mike Leahy. 2013. "Update on Foreign Central Bank Operating Procedures and the Foreign Experience with Using Interest on Reserves as a Monetary Policy Instrument." April 19.

Carpenter, Seth, Jane Ihrig, Elizabeth Klee, Daniel Quinn, and Alexander Boote. 2015. "The Federal Reserve's Balance Sheet and Earnings: A Primer and Projections." International Journal of Central Banking 11, no. 2 (March): 237-283.

Ennis, Huberto M., and John A. Weinberg. 2007. "Interest on Reserves and Daylight Credit." Federal Reserve Bank of Richmond Economic Quarterly 93, no. 2 (Spring): 111-142.

Ennis, Huberto M., and Todd Keister. 2008. "Understanding Monetary Policy Implementation." Federal Reserve Bank of Richmond Economic Quarterly 94, no. 3 (Summer): 235-263.

Ennis, Huberto M., and Alexander L. Wolman. 2010. "Excess Reserves and the New Challenges for Monetary Policy." Federal Reserve Bank of Richmond Economic Brief 10-03, March.

Ennis, Huberto M., and Jeff W. Huther. 2021. "The Fed's Evolving Involvement in the Repo Markets." Federal Reserve Bank of Richmond Economic Brief 21-31, September.

Fessenden, Helen. 2015. "A Bridge Too Far?" Federal Reserve Bank of Richmond Econ Focus, Third Quarter.

Friedman, Milton. 1959. A Program for Monetary Stability. New York: Fordham University Press.

Friedman, Milton. 1969. The Optimum Quantity of Money. New York: Macmillan.

Frost, Joshua, Lorie Logan, Antoine Martin, Patrick E. McCabe, Fabio M. Natalucci, and Julie Remache. 2015. "Overnight RRP Operations as a Monetary Policy Tool: Some Design Considerations." Federal Reserve Bank of New York Staff Report 712, February.

George, Esther L. 2020. "Perspectives on Balance Sheet and Credit Policies: A Tribute to Marvin Goodfriend." Cato Journal 40, no. 3 (Fall): 589-594.

Goodfriend, Marvin. 1987. "Interest Rate Smoothing and Price Level Trend-Stationarity." Journal of Monetary Economics 19, no. 3 (May): 335-348.

Goodfriend, Marvin. 1991. "Interest Rates and the Conduct of Monetary Policy." Carnegie-Rochester Conference Series on Public Policy 34 (Spring): 7-30.

Goodfriend, Marvin. 2000. "Overcoming the Zero Bound on Interest Rate Policy." Journal of Money, Credit and Banking 32, no. 4, Part 2 (November): 1007-35.

Goodfriend, Marvin. 2002. "Interest on Reserves and Monetary Policy." Economic Policy Review 8, no. 1 (May): 77-84.

Goodfriend, Marvin. 2014. "Monetary Policy as a Carry Trade." Monetary and Economic Studies 32 (November): 29-44.

Goodfriend, Marvin, and Robert G. King. 1988. "Financial Deregulation, Monetary Policy, and Central Banking." Federal Reserve Bank of Richmond Economic Review 74 (May/June): 3-22.

Goodfriend, Marvin, and Jeffrey M. Lacker. 1999. "Limited Commitment and Central Bank Lending." Federal Reserve Bank of Richmond Economic Quarterly 85, no. 4 (Fall): 1-27.

Haltom, Renee Courtois, and Alexander L. Wolman. 2016. "How Did Short-Term Market Rates React to Liftoff?" Federal Reserve Bank of Richmond Economic Brief 16-09, September.

He, Dong. 2018. "Monetary Policy in the Digital Age." IMF Finance & Development 55, no 2 (June).

Ihrig, Jane, Zeynep Senyuz, and Gretchen C. Weinbach. 2020. "Implementing Monetary Policy in an 'Ample-Reserves' Regime." FEDS Notes, October 2.

Ireland, Peter N. 2019. "Interest on Reserves: History and Rationale, Complications and Risks." Cato Journal 39, no. 2 (Spring/Summer): 327-337.

Keister, Todd, Antoine Martin, and James McAndrews. 2008. "Divorcing Money from Monetary Policy." Federal Reserve Bank of New York Economic Policy Review 14, no. 2 (September): 41-56.

Keister, Todd, and James McAndrews. 2009. "Why are Banks Holding So Many Excess Reserves?" Current Issues in Economics and Finance 15, no. 8 (December): 1-10.

Lacker, Jeffrey M. 1997. "Clearing, Settlement and Monetary Policy." Journal of Monetary Economics 40, no. 2 (October): 347-381.

Lacker, Jeffrey M. 2001. "Introduction to Special Issue [on the Treasury-Fed Accord]." Federal Reserve Bank of Richmond Economic Quarterly 87, no. 1 (Winter): 1-6.

Poole, William. 1968. "Commercial Bank Reserve Management in a Stochastic Model: Implications for Monetary Policy." Journal of Finance 23, no. 5 (December): 769-791.

Reis, Ricardo. 2015. "Different Types of Central Bank Insolvency and the Central Role of Seignorage." Journal of Monetary Economics 73, no. 1 (June): 20-25.

Cite as: Ennis, Huberto M., and John A. Weinberg. 2022. "Paying Interest on Reserves." In Essays in Honor of Marvin Goodfriend: Economist and Central Banker, edited by Robert G. King and Alexander L. Wolman. Richmond, Va.: Federal Reserve Bank of Richmond.


We would like to thank Todd Keister, Bob King, Beth Klee, Jeff Lacker, and Alexander Wolman for comments. All errors are our own.


Friedman (1969).


See also, Ireland (2019).


Elsewhere in this volume, Ben Bernanke discusses Marvin's initial work on the zero lower bound, Goodfriend (2000) and a follow-up Jackson Hole paper, Goodfriend (2016).


Keister and McAndrews (2009).


Goodfriend (2002).


Keister et al. (2008) provides a detailed discussion of these ideas.


Goodfriend and King (1988). See the essay by Douglas Diamond elsewhere in this volume.


Elsewhere in this volume, John Taylor discusses Goodfriend (1991) and Michael Dotsey, Andreas Hornstein, and Alexander Wolman discuss Goodfriend (1987).


In Ennis and Weinberg (2007), we investigate formally the link between excess reserves and intraday credit. For the link between reserves and discount window lending, see Ennis and Klee (2021).


Goodfriend and King (1988) and Goodfriend and Lacker (1999).


For a recent thorough study of this issue see Carpenter et al. (2015).


Goodfriend (2014).


Reis (2015) provides a good overview of central bank solvency and its relationship with independence. In the years after Marvin wrote his piece, a number of political actions by Congress have called into question the ability of the Fed to independently manage its surplus account (Fessenden, 2015; and George, 2020). Indeed, in later writings, Marvin expressed greater concern about the independence consequences of cash flow volatility and about the true availability of future unencumbered seigniorage (Goodfriend, 2014).


This essay only discusses the experience with IOR in the US. For an international perspective, see Bowman et al. (2013) and the references therein.


Poole (1968).


Ireland (2019).


The document produced by the workgroup for the Board and the FOMC is available at


See Ireland (2019) for a critical assessment of the policy of paying interest on reserves. Somewhat ironically, the motivation for paying interest on reserves as a way to finance credit programs is in sharp contrast with Marvin’s thinking — he suggested that the ability to adjust the amount of excess reserves could actually reduce the urge of central banks to pursue other more objectionable credit policies.


Bech and Klee (2011).


Frost et al. (2015).


See, for example, Haltom and Wolman (2016).


See the March 2019 Fed press release, "Balance Sheet Normalization Principles and Plans."


See, for example, the Fed's online press release from January 30, 2019. For a detailed discussion of monetary policy implementation in a system with "ample" reserves, see Ihrig et al. (2020).


Initially, the ON RRP facility was envisioned to function mainly as a backstop in short-term secured credit markets (Frost et al., 2015). However, with the current configuration of interest rates and the very large Fed balance sheet, the ON RRP has become heavily used — and functions as a way for the Fed to pay interest directly on cash reserves held by money market funds and other financial institutions (such as the GSEs), which are eligible to use the ON RRP but are not eligible to receive IOR. This situation seems to be hardly something that Marvin (or we) would have recommended or even anticipated.


Andolfatto and Ihrig (2019).


See Ennis and Huther (2021) for a discussion.


It has become a real concern that the Fed will deem it necessary to go back to quantitative easing policies before the size of its balance sheet has had the time to fully normalize from its very high current level. This possibility of a “ratchet effect” suggests to us that normalization should become a priority as soon as conditions allow it.


See, for example He (2018) and Benigno (2019).

Phone Icon Contact Us

Lisa Davis (804) 697-8179