The minutes of the November 2018 meeting of the Federal Open Market Committee (FOMC or Fed) cover a lengthy discussion on possible changes to the policy tools and procedures the Fed uses to control money-market interest rates. The Fed appears to be moving toward the proposal that I made with Brian Sack in a 2014 PIIE Policy Brief. Our proposal would give the Fed firmer control of the most important short-term interest rates in the economy while saving taxpayers money and keeping the payments system safe and liquid.

Our 2014 proposal had four main components:

1. The Fed should operate in a “floor” system in which it provides abundant liquidity by maintaining a large balance sheet and controls market interest rates through the influence of administered interest rates on the liquidity that it creates. This system should be designed so that private overnight interest rates trade at levels close to, but slightly above, the Fed’s administered rates.
2. The Fed should augment its tool of paying interest on bank reserves (IOR) with an overnight reverse repurchase (RRP) facility offered on a fixed-rate basis to important nonbank institutions.
3. The Fed should set the IOR and RRP rates very close to, or even equal to, each other.
4. The Fed should drop the federal funds rate as its operating target, replacing it with its administered rates (IOR, RRP) or a private repo rate.

How has Fed policy evolved, and how is it likely to evolve going forward, in relation to these four components?

1. The minutes listed several advantages of a floor operating system, many of which were described in our Policy Brief. Most importantly, a floor system gives the Fed effective control of overnight interest rates in a system that is operationally simple. In addition, a floor system increases the safety and efficiency of the payments system, facilitates regulatory changes that require banks to hold more liquid assets, makes interest rates less volatile, and enables the Fed to use its balance sheet to achieve other objectives without sacrificing control of short-term interest rates.

The minutes also listed potential drawbacks of a floor system, but the drawbacks were either illogical or unpersuasive, and they do not appear to reflect a serious challenge to a floor system. The chief objection was that a floor system might require a long-run balance sheet that would be deemed excessive, though the judgment of what would be excessive appears to reflect potential political concerns rather than economic costs. The Fed should let the size of its balance sheet be determined by what is best for the economy and not accept an inferior operating regime out of concern for political pressures that may never materialize.

2. At the time we wrote our Policy Brief, the Fed was developing and testing an overnight RRP facility, but it was not yet clear how that facility would be used or how long it would exist. We felt that running the RRP facility at a fixed rate with unlimited quantity would be the most effective approach. The Fed has in fact run RRPs in this manner over the period since our proposal,[1] and it has proved to be an effective floor on market interest rates—one that has been much stronger than IOR would have been on its own.

3. The Fed for some time set the IOR rate 0.25 percentage point above the RRP rate. We had instead recommended that the two rates be placed closer together, arguing that this would make the financial system more efficient and equitable. The wide spread that the Fed implemented tilted the Fed’s liabilities towards bank reserves and away from RRP, and it left overnight market rates below the rate that the Fed was paying to banks, thus creating an implicit subsidy for banks (because nonbank institutions are not eligible to earn interest on reserves). The Fed could have saved taxpayers roughly $2 billion per year over this period by eliminating this spread (as I explained in a December 2015 blog post).[2]

Recent market developments, including heavy issuance of Treasury bills, have put upward pressure on private interest rates relative to the Fed’s administered rates, causing the federal funds rate to firm up to near the IOR rate. With that configuration of market rates, the Fed earlier this year reduced the level of the IOR rate relative to its target range for the federal funds rate, which also narrowed the spread of the IOR rate to the RRP rate to 0.20 percentage point. The minutes indicated that the Fed is considering another move in this direction, likely reducing the spread between the administered rates to 0.15 percentage point.

The narrowing of the spread between these administered rates is a welcome change that will establish a more effective floor system for the future.[3] Indeed, as argued in our Policy Brief, there are important benefits of narrowing the gap. Given the safety and liquidity of reserves, the interest rate on reserves should be the lowest of all rates. With Fed overnight RRPs equally safe and nearly as liquid, it makes sense for the IOR and RRP rates to be very close to each other or even equal.[4] The configuration of rates today, which has left RRP usage at minimal levels, gives the Fed the opportunity to make these changes with little disruption.

4. We recommended dropping the federal funds rate as a target because we thought that activity in the federal funds market would shrink markedly, and the rate would become more volatile and idiosyncratic, as overnight market rates approached the IOR rate. Although the federal funds rate has been less volatile than we expected, the minutes indicated that the Fed too has worries along these lines. Specifically, the minutes said that the Fed may consider replacing the federal funds rate target with a target for the overnight bank funding rate (OBFR), a broader measure that better reflects the overall cost of overnight unsecured bank funding.

We had proposed that the Fed should target either the administered rates or a secured private repo rate rather than an unsecured bank funding rate. The administered rates are set by the Fed and represent pure risk-free rates that should be propagated to a wide set of market rates in the financial system. If policymakers deem it important for the Fed to have a market rate as a target (for reasons we discussed in the Policy Brief), we see a repo rate as the preferred choice, given that the repo market is very large and liquid and has essentially no credit risk.[5] The minutes indicated that, in addition to considering a shift to the OBFR rate, “some participants saw it as desirable to explore the possibility of targeting a secured interest rate,” which would be in line with our proposal.

Altogether then, the Fed recently signaled some likely future decisions that will significantly improve the efficiency, stability, and effectiveness of its monetary policy framework.


1. A modest difference is that the Fed imposed a high, $30 billion, limit on RRP for any one counterparty.

2. Congress apparently had the taxpayer’s interest in mind when it wrote the legislation authorizing IOR, which states that the IOR rate is “not to exceed the general level of short-term interest rates.”

3. In effect, increased Treasury bill supply has achieved what we had wanted to accomplish through the RRP facility.  Bills and RRPs are both short-term risk-free instruments provided by the government, and hence it would not be surprising if the availability of bills at rates around the IOR rate were to have similar benefits to what we intended from offering RRPs around the IOR rate.  Of course, bills are not offered as a fixed-rate, full allotment instrument, but the rapid increase in the supply of bills pushed their yields up, achieving much of the same effect for now.  The reason to move the RRP rate closer to the IOR rate at this time is to ensure that the current configuration of rates will be sustained even after any temporary effects from bill supply dissipate.

4. Our Policy Brief argued that regulatory differences between banks and nonbanks, including fees collected by the Federal Deposit Insurance Corporation, should be decided on their merits and that it is not appropriate for the Fed to tilt the playing field to offset such differences.

5. Some argue that targeting an unsecured rate is more appealing if policymakers wish to automatically counteract spikes in risk aversion between market participants—that is, periods of financial stress.  However, we feel that it makes more sense to respond to such stress through explicit decisions about a range of instruments available, including direct lending facilities, rather than trying to hard code the response into the choice of the monetary policy instrument. Any such hard coding would be inadequate as a policy response during such episodes, which limits its advantage in our view.