In Which You Will Learn About the “Thanksgiving Turkey” and the “Lombard Trap”
The Bank of England, ECB, Bank of Canada and Reserve Bank of Australia have all recently been reducing their securities holdings and allowing emergency loans to roll off until the banking system is left with insufficient reserves in aggregate, at which point banks will need to meet their reserve demand by borrowing from the central bank’s standing lending facilities. The term being used for this approach is “demand-driven,” to distinguish it from a “supply-driven” approach in which the central bank owns enough securities that banks have all the reserves they want and then some, with no need to borrow. See here and here for more information.
The Fed operated a similar demand-driven approach from roughly 1982 to 1989 where the banking system was, by design, structurally short of reserves and banks needed to borrow the remainder from the Fed’s discount window. In this blog, we unpack how that approach – which was called “borrowed-reserves targeting” – worked and why it ended in order to shed light on the challenges today’s demand-driven approaches may face.
Background – the Fed’s Balance Sheet
The Fed’s assets consist of Treasury securities, repos of treasury securities (actually reverse repos, but the Fed labels them from the counterparties’ perspective) and discount window loans. In the 1980s, the Fed’s liabilities were mostly currency, as well as the Treasury’s general account and reserve balances, both of which were small. Because the amount of currency and the Treasury’s general account were exogenous, when the Fed adjusted the size of its assets, it adjusted the quantity of reserve balances. If the Fed chose a certain quantity of repo, it determined the quantity of reserve balances (the two weren’t equal, they just moved up and down one-for-one). If the banking system demanded a larger quantity of reserve balances than the Fed supplied through its repos and Treasury securities, banks would have to, in aggregate, borrow from the discount window. The amount of borrowing was referred to as “borrowed reserves” and total reserves minus borrowed reserves equaled “nonborrowed reserves.”
The Fed’s Demand-Driven Approach
Under borrowed reserves targeting, at each meeting, the FOMC set a target for the quantity of discount window borrowing, rather than a target for the federal funds rate. Which is not to say that the Committee did not have a specific federal funds rate in mind – they did. But because the amount of borrowing was increasing in the level of the fed funds rate, holding the discount rate fixed, it was possible to achieve a desired fed funds rate by targeting the amount of borrowing. It seems a roundabout way to implement policy from our current perspective but less so in historical context, as discussed below.
The supply curve used to implement policy under this approach is depicted in Exhibit 1, with the quantity of reserves on the x-axis and interest rate on the y-axis. The New York Fed and Board staff would first estimate the expected total reserve demand of the banking system at the desired range for the federal funds rate. From there, they would derive the necessary quantity of non borrowed reserves over each two-week reserve maintenance period by subtracting the borrowing target from reserve demand. The Desk would then conduct open market operations to deliver the necessary quantity of non-borrowed reserves. As Meulendyke (1989) described it, when non-borrowed reserves were insufficient to meet demand, the banking system as a whole had no practical alternative to borrowing from the discount window to meet their reserve demand. The purpose of the approach was to “force” the banking system to borrow the intended aggregate amount (Lindsey 1987).
As noted, the key to understanding the approach is the upward sloping part of the supply curve. That is, there was a relationship between the quantity of discount window borrowing and the spread of the federal funds rate over the discount rate. At that time, the discount rate was a below-market rate, so banks always had a financial incentive to borrow from the window. However, banks were reluctant to do so because frequent or prolonged borrowing brought administrative pressure. As the fed funds rate rose above the discount rate, the financial incentive to borrow from the window outweighed the cost of administrative pressure and banks began to borrow. The effective anchoring rate was thus not the discount rate itself but the discount rate plus an implicit reluctance premium.
As Goodfriend (1981) showed formally, this reluctance produced a smooth, upward-sloping curve: banks would borrow more as the spread widened, but gradually and continuously rather than jumping discontinuously when the funds rate crossed the discount rate. As the July 1987 Lindsey and Glassman memo to the FOMC put it, there is “a systematic, if somewhat loose, association between the spread of the funds rate over the discount rate and the willingness of institutions to draw on [discount window borrowing].” The memo estimated that on average a $100 million increase in borrowing was associated with a 24 to 25 basis point increase in the funds-discount spread.
One challenge in operationalizing the approach was that there was “considerable looseness in the relationship between intended borrowing and the federal funds rate, especially over the short-run” (Kohn, 1987). In other words, it was difficult to predict what the fed funds rate would be for a given discount window borrowing target. That uncertainty in the supply curve is depicted in Exhibit 1 by the shaded blue area. One factor which led to the instability was banks’ changing reluctance to borrow from the discount window. For example, in spring 1984, Continental Illinois’s faced funding difficulties and had borrowed significantly from the discount window. Larger banks “shied away from the discount window to avoid any risk of rumors about their financial condition”, pushing the funds rate higher for a given level of borrowing (Kohn, 1987). As Continental’s funding problems were brought under control, discount window borrowing and the fed funds rate spread normalized.
How the Fed Got There — and Why It Left
The demand-driven / borrowed reserves approach was the third in a sequence, each adopted in response to the perceived failings of the previous one.
Before October 1979, the Fed explicitly targeted a range for the federal funds rate, with the target range becoming progressively narrower through the 1970s. The intellectual critique of this approach had been made most forcefully by Karl Brunner and Allan Meltzer in a report to the House Committee on Banking and Currency in 1964 and elaborated through the Shadow Open Market Committee that they founded in 1973. Their argument was that interest rate targeting created a systematic inflationary bias: when inflation rises, a fixed nominal funds rate implies a falling real rate, which stimulates demand further, feeding more inflation in a self-reinforcing loop. A central bank targeting a nominal rate has no automatic stabilizer against an inflationary shock because the instrument moves in the wrong direction in real terms precisely when tightening is needed. The 1970s were the proof: the Fed was repeatedly unwilling to raise the funds rate fast enough with the result that monetary growth consistently exceeded target ranges (Lindsey, Orphanides and Rasche, 2004). One could apply the same argument to the post-COVID inflationary period, when the Fed was slow to raise rates as inflation surged.
The shift away from targeting the fed funds rate in October 1979 is often described as a direct response to the Brunner-Meltzer critique, but communications and Committee management considerations were equally as important. At the previous FOMC meeting in September 1979 there were dissents from a half-point discount rate increase, which markets saw as a signal that the Fed was unwilling to raise rates high enough to contain money growth. “The Reserve Board vote,” one municipal bond dealer said, “makes me think that this is as much of a push toward higher rates as we’re going to get for a while. I [don’t] think that 4-to-3 vote sat very well with a lot of traders today.” (Allen, 1979, p. D9). As former Fed officials who were closely involved in these events recall, the core problem was that the Committee was debating every quarter-point move when much more was needed, and the explicit rate-voting process was making it nearly impossible to move fast enough. As Volcker put it “When you have to make an explicit decision about interest rates all the time … people don’t like to do it. You’re always kind of playing catch-up.” (Treaster, 2004). Shifting to targeting the quantity of reserves meant that large increases in the fed funds rate could happen through the markets rather than through explicit Committee votes. Doing so provided political cover at a moment when the pain of disinflation was acute and political pressure to hold rates down was intense.
The quantity-based approach was known as the non-borrowed reserve targeting and lasted only three years. The resulting interest rate volatility was dramatic (the funds rate swung by hundreds of basis points within single maintenance periods) and by 1982 the FOMC had concluded that the cost was too high. In October 1982 the Fed shifted to the borrowed reserves framework.
Nonborrowed reserve targeting unambiguously entailed the Fed targeting a quantity rather than a price, as advocated by Melzer and Brunner. The borrowed reserve targeting approach had the appearance of the same feature, but once you recognize that the borrowed reserve target was chosen to deliver a desired fed funds rate and that nonborrowed reserves were adjusted to deliver the borrowed reserve target, the difference between the approach and targeting the funds rate becomes more rhetorical than real.
In addition to seeming to align with academic views on the appropriate way to conduct monetary policy, the demand-driven approach retained some of the Committee management benefits of the previous approach. Former officials who worked closely with these decisions recall that Volcker favored the approach in part because it meant the FOMC did not have to vote on every quarter-point change in the funds rate target — reducing the visibility of incremental tightening and concentrating significant operational discretion in the Chairman. When the upward sloping supply curve shifted, Volcker got involved personally in deciding whether to allow the shift to pass through into money market conditions. That discretion proved consequential. As mentioned above, in spring 1984, Continental Illinois’s difficulties caused a sharp reduction in banks’ willingness to borrow, pushing the funds rate higher. This occurred after Jim Baker had told Volcker, in President Reagan’s presence, that the president instructed him not to raise rates that year. Volcker allowed the tightening to pass through anyway — a decision that, as one former official who was present reflected, was courageous and consequential for continuing progress on disinflation, even if Volcker himself appeared not to remember it that way when the subject was raised years later.
The approach also partially addressed the Brunner-Meltzer critique since an increase in reserve demand caused by additional lending to the real economy automatically resulted in higher interest rates due to the upward sloping supply curve. As Don Kohn noted in his cover memorandum transmitting the July 1987 staff study to the Committee, “the capacity for such market forces to show through in the funds rate could be viewed as a desirable property of current procedures, to the extent it conveys information about market conditions and on occasion allows the funds rate to begin moving more quickly than otherwise to levels later judged to be appropriate.” Similarly, the Axilrod memo of December 1984 noted that the approach “has worked reasonably well over a difficult period” precisely because it retained “a very valuable degree of responsiveness of the funds rate… to actual or anticipated changes in credit or money demand.”
The Fed did not announce the borrowing target. Neither the Record of Policy Action (essentially the meeting minutes) nor the Directive (what is today called the “Implementation note”), both of which were released a few days after the subsequent meeting, included the target for non-borrowed reserves that the Committee had decided on and conveyed to the Desk. Instead of announcing an official target, the Directive included opaque language about “pressures on reserve positions”. The borrowing targets were released, eventually, in the “Annual Report of the Manager of the System Account,” which was published in the spring of the following year.
The approach ended not with a formal decision but with a mishap that became famous inside the Fed—the “Thanksgiving Turkey.” On the Wednesday before Thanksgiving 1989, the Desk needed to add reserves for routine technical reasons. Even though the funds rate had slipped only slightly below the Committee’s expected level, the Desk went ahead with a five-day system repurchase agreement – increasing the supply on non-borrowed reserves. Market participants interpreted the operation as signaling a policy easing when none was intended. On Friday morning the New York Times cited government officials claiming that the Fed had eased. The December 1989 FOMC policy record noted dryly that “conditions in reserve markets softened temporarily around Thanksgiving when operations to meet seasonal reserve needs were misread as signaling a further easing of monetary policy.” As Lindsey (2003) put it: “After the incident, the Desk put top priority on signaling the Committee’s intended funds rate. And that’s the story of how the FOMC lost its borrowing procedures.”
The underlying cause was not the mishap itself but the instability in the upward sloping part of the supply curve. Memos from Donald Kohn and Peter Sternlight in December 1987 and March 1988 document how the stock market crash of October 1987 had caused a sharp decrease in banks’ willingness to borrow from the discount window, making the fed funds rate much higher than predicted at the borrowing target. Borrowing ran persistently below the target as reluctance surged, and the Desk’s repeated downward revisions were “inadequate to capture behavior fully” through year-end 1987. By 1988 the Desk was already placing “slightly more weight than pre-October 19” on the funds rate; by late 1989 the fiction of borrowing-based targeting was exhausted, and the Thanksgiving Turkey merely solidified what had already become practice. Former officials recall that by this point it was generally understood in markets what the intended funds rate was from the Desk’s daily operations. The existing language about the borrowing assumption in the FOMC’s directive had become a roundabout way of communicating a rate target the Committee was not quite ready to acknowledge explicitly.
Lessons for Demand-Driven Approaches
The central lesson for demand-driven approaches today is about the instability of the effective anchoring rate. Under the discount window borrowing target, the anchor was not the discount rate but the discount rate plus an implicit reluctance premium — and that premium was not stable. The July 1987 Lindsey/Glassman memo showed that the minimum spread required to induce any borrowing shifted significantly across sub-periods. For example, the 1989 OMO Annual Report estimated that $243 million of adjustment credit required a spread of 228 basis points in 1989; in 1988, the same quantity had required only 137 basis points; in 1987, only 100 basis points. The upward sloping part of the supply curve had shifted, monetary conditions had been inadvertently tightened and the approach’s credibility had eroded.
The demand-driven approaches being adopted today differ from the Fed’s old demand-driven approach in one crucial respect: their lending rates are set above prevailing money market rates rather than below them. This means no stigma is required as a rationing device. While stigma is not a requirement, it is still a concern worth monitoring. If banks become more or less willing to borrow at the central bank’s lending facility — due changes in supervisory attitudes toward borrowing, or an event where banks avoid using the facility due to perceptions of financial health — the anchor of the lending facility rate will weaken, and so market rates could rise unexpectedly. The BoE, ECB and RBA are working hard to make their facilities routine and stigma-free precisely to avoid this. They are, in effect, trying to solve in advance the problem that eventually undid the Fed’s demand-driven approach. Whether they succeed will depend not just on the design of the facilities but on supervisory culture and the willingness of banks to be seen borrowing.
What modern central banks see as a feature, the Fed of the mid-1980s saw as a bug. Staff observed that if the discount rate were set as an above-market rate, and rules about use were eliminated, the Fed could fall into “the Lombard Trap.” A Lombard facility is a central bank lending facility that charges an above market rate and extends credit on a no-questions-asked basis, exactly how the discount window is designed today (sort of, see here). If the discount window during the borrowed-reserves regime were a Lombard facility, the approach would not work. If the Fed left the banking system short of reserves, the fed funds rate would rise immediately to the discount rate where they would stay no matter how small or large the shortfall. The Lombard Trap is to under-supplying reserves what the Liquidity Trap is to over-supplying reserves. To change the fed funds rate, the Fed would have to change the discount rate, precisely the outcome the Committee was trying to avoid.
We thank Jim Clouse, Don Kohn and Brian Madigan for helpful discussions and suggestions.
References
Allen, John H. “Bond Prices Score Strong Gains.” The New York Times, September 20, 1979, p. D9.
Axilrod, S.H. Memorandum to the Federal Open Market Committee, “Comments on Operating Procedures and Certain Monetary Targeting Issues,” December 11, 1984.
Brunner, Karl and Allan H. Meltzer. The Federal Reserve’s Attachment to the Free Reserves Concept. Report to the House Committee on Banking and Currency, 1964.
Esaki, H. and A. Meulendyke. Memorandum to Peter Sternlight, “The Impact of the Continental Bank Crisis on Reserve Management by Large Banks,” September 28, 1984.
Goodfriend, Marvin. “Discount Window Borrowing, Monetary Policy, and the Post-October 6, 1979 Federal Reserve Operating Procedure.” Federal Reserve Bank of Richmond Working Paper 81-2, 1981. Published as “Discount window borrowing, monetary policy, and the post-October 1979 Federal Reserve operating procedures,” Journal of Monetary Economics 12 (1983): 345–356.
Kohn, Donald L. and Peter D. Sternlight. Memorandum to the Federal Open Market Committee, “Strategies for Open Market Operations,” December 11, 1987.
Kohn, Donald L. and Peter D. Sternlight. Memorandum to the Federal Open Market Committee, “Issues in the Implementation of Open Market Operations,” March 25, 1988. (Attachment: Ann-Marie Meulendyke, “Federal Reserve Policy Targets and Operating Guides in Recent Decades: A Review.”)
Lindsey, David E. and James Glassman. Memorandum to Donald Kohn, “A Review of the Relation of the Funds Rate and Intended Discount Borrowings,” July 1, 1987. Transmitted to the FOMC by Donald L. Kohn.
Lindsey, David E. “A Modern History of FOMC Communication: 1975–2002.” Memorandum to the Federal Open Market Committee, June 24, 2003.
Lindsey, David E., Athanasios Orphanides, and Robert H. Rasche. “The Reform of October 1979: How It Happened and Why.” Finance and Economics Discussion Series 2005-02, Divisions of Research & Statistics and Monetary Affairs, Federal Reserve Board, Washington, D.C., December 2004.
Meulendyke, Ann-Marie. U.S. Monetary Policy and Financial Markets. Federal Reserve Bank of New York, 1989.
Federal Reserve Bank of New York. Monetary Policy and Open Market Operations During 1989. Open Market Group Annual Report, March 1990.
Strongin, Steven. “The Identification of Monetary Policy Disturbances: Explaining the Liquidity Puzzle.” Journal of Monetary Economics 35 (1995): 463–497.
Treaster, Joseph. “B. Paul Volcker: The Making of a Financial Legend. Hoboken, NJ: John Wiley & Sons”, 2004, pp. 61-62.
Wallich, Henry C. “Recent Techniques of Monetary Policy.” Federal Reserve Bank of Kansas City Economic Review, May 1984.
________
Disclaimer: The views expressed do not necessarily reflect those of the Bank Policy Institute’s member banks, and are not intended to be, and should not be construed as, legal advice of any kind.
