Finance and Economics Discussion Series

Divisions of Research & Statistics and Monetary Affairs

Federal Reserve Board, Washington, D.C.

Monetary Policy 101: A Primer on the Fed’s Changing Approach

to Policy Implementation

Jane E. Ihrig, Ellen E. Meade, and Gretchen C. Weinbach

2015-047

Please cite this paper as:

Jane E. Ihrig, Ellen E. Meade, and Gretchen C. Weinbach (2015). “Monetary Policy 101: A
Primer on the Fed’s Changing Approach to Policy Implementation,” Finance and Economics
Discussion Series 2015-047. Washington: Board of Governors of the Federal Reserve System,
http://dx.doi.org/10.17016/FEDS.2015.047.

NOTE: Staff working papers in the Finance and Economics Discussion Series (FEDS) are preliminary
materials circulated to stimulate discussion and critical comment. The analysis and conclusions set forth
are those of the authors and do not indicate concurrence by other members of the research staff or the
Board of Governors. References in publications to the Finance and Economics Discussion Series (other than
acknowledgement) should be cleared with the author(s) to protect the tentative character of these papers.
 Monetary Policy 101:
A Primer on the Fed’s Changing Approach to Policy Implementation

Jane Ihrig, Ellen E. Meade, and Gretchen C. Weinbach*

June 30, 2015

Abstract
The Federal Reserve conducts monetary policy in order to achieve its statutory mandate of
maximum employment, stable prices, and moderate long-term interest rates as prescribed by
the Congress and laid out in the Federal Reserve Act. For many years prior to the financial
crisis, the FOMC set a target for the federal funds rate and achieved that target through small
purchases and sales of securities in the open market. In the aftermath of the financial crisis,
with a superabundant level of reserve balances in the banking system having been created as a
result of the Federal Reserve’s large scale asset purchase programs, this approach to
implementing monetary policy will no longer work. This paper provides a primer on the Fed’s
implementation of monetary policy. We use the standard textbook model to illustrate why the
approach used by the Federal Reserve before the financial crisis to keep the federal funds rate
near the FOMC’s target will not work in current circumstances, and explain the approach that
the Committee intends to use instead when it decides to begin raising short-term interest rates.

Keywords: Federal Reserve, FOMC, monetary policy implementation, monetary policy tools, monetary
policy normalization, liftoff

JEL Classifications: E58, E52, E43

* Economists in the Division of Monetary Affairs at the Board of Governors of the Federal Reserve
System. We thank Miguel Acosta, Jim Clouse, Melanie Josselyn, Joe Kachovec, Steve Meyer, Ben Miller,
and Bill Nelson for comments. Melanie Josselyn and Joe Kachovec provided excellent research
assistance. The views expressed here are those of the authors and do not necessarily reflect the views
of other members of the research staff, the Board of Governors of the Federal Reserve System, or the
Federal Reserve System.
1 Introduction

The Federal Reserve conducts monetary policy in order to achieve its statutory mandate of
maximum employment, stable prices, and moderate long-term interest rates as prescribed by
the Congress and laid out in the Federal Reserve Act. 1 For many years prior to the financial
crisis, the Federal Open Market Committee (FOMC or “Committee”) set a target for the federal
funds rate, an overnight interbank borrowing rate, and achieved that target through small
purchases and sales of securities in the open market, known as open market operations. In the
aftermath of the financial crisis, with a superabundant level of reserve balances in the banking
system having been created as a result of the Federal Reserve’s large scale asset purchase
programs, this approach to implementing the FOMC’s monetary policy will no longer work.

This paper provides a primer on how the Federal Reserve will implement monetary policy when
the FOMC decides it is time to raise interest rates. We begin with the standard textbook model
of reserve balances to illustrate the approach used by the Federal Reserve before the financial
crisis to keep the federal funds rate near the FOMC’s target. We then explain why that pre-
crisis approach will not work in the current environment. After that, we discuss the policy tools
available to the FOMC to implement monetary policy today, and explain the approach that the
Committee intends to take when it decides to begin raising short-term interest rates.

2 How did the Fed implement monetary policy prior to the financial crisis?

Money and Banking textbooks have long sketched out the framework under which open market
operations (OMOs) were effective in implementing monetary policy, a framework based
importantly on two key features—reserve requirements and reserve scarcity. Reserve
requirements are set by the Federal Reserve and banks hold reserve balances to meet these
requirements (see Box 1 for more information about how reserve requirements affect reserve
balances). At the same time, before the financial crisis, banks generally kept their balances to a
minimum, in part because the balances earned no return. The combination of Federal Reserve-
created demand for reserve balances and banks’ desire to limit such balances drove an active
interbank market—known as the federal funds market—for exchanging these funds, one in
which banks borrowed from and lent funds to each other on a daily basis at an interest rate
known as the federal funds rate. With reserve balances generally scarce, the Federal Reserve
could meaningfully affect the market-determined level of the federal funds rate and keep it
close to the FOMC’s target level by announcing the target level of the federal funds rate and
making small changes in the supply of aggregate reserves, as needed.

1
The Federal Reserve’s statutory mandate is often referred to as a “dual mandate” of maximum employment and
price stability because moderate long-term interest rates will result if prices are expected to be stable.

Page 2 of 29
 Box 1: How do reserve requirements affect reserve balances?

The Federal Reserve Act (as amended by the Monetary Control Act of 1980) and the International
Banking Act of 1978 impose reserve requirements on most deposit-taking institutions in the United
States, requiring that they hold a certain fraction of their deposits in reserve. In particular, all
depository institutions—commercial banks, savings banks, thrift institutions, and credit unions—as well
as most U.S. branches and agencies of foreign banks (hereafter “banks,” for simplicity) are assessed
reserve requirements against certain deposit liabilities. Banks are required to satisfy their reserve
requirements in the form of vault cash, which they hold primarily to meet the liquidity needs of their
customers and, if the quantity of vault cash held is insufficient, also in the form of a balance maintained
at the Federal Reserve. The balances banks maintain at the Federal Reserve that are necessary for
meeting reserve requirements are called required reserve balances; any reserve balances held in excess
of what is necessary to meet reserve requirements are termed excess balances. 2

Prior to the financial crisis, many banks in the United States satisfied their entire reserve requirement
with vault cash, though about 900 banks did not and so also needed to maintain reserve balances at the
Fed to satisfy their requirements. The total amount of reserve balances in the banking system hovered
around $15 billion, with excess balances making up less than $2 billion of this total. However, as
discussed in greater detail in section 2 and shown in the figure below, reserve balances have grown
tremendously since the financial crisis. In late December 2014, reserve balances stood at more than
$2.6 trillion and have remained in that neighborhood since then, with excess balances making up all but
about $90 billion of this total.

Total reserve balances

2
In practice, banks meet their required reserve balances (also referred to as “reserve balance requirements”) with
some leeway. A penalty-free band is used to create a range on both sides of the required reserve balance within
which a bank needs to maintain its average balance over a given period. For more information on the calculation
and maintenance of reserve requirements, see the Federal Reserve Board’s Reserve Maintenance Manual and web
page on Reserve Requirements. Data on reserve balances are published weekly on the H.3 Statistical Release.

Page 3 of 29
Figure 1 presents the standard demand and supply framework for reserve balances shown in
many Money and Banking textbooks. Demand is downward sloping because the higher is the
opportunity cost of holding reserve balances, the lower is the demand for them. Conversely, as
the price of overnight borrowing falls, banks are generally inclined to borrow more in order to
satisfy their reserve requirements and also possibly to leave themselves with a bit of extra
balances to protect against unexpected outflows that can cause reserve balance deficiencies,
for which banks are charged a penalty. In theory, the price, or rate of interest, that an
institution is willing to pay to borrow funds overnight—that is, the federal funds rate—should
be capped at the Federal Reserve’s primary credit rate. The primary credit rate is the interest
rate that the Fed charges banks to borrow overnight from its primary credit program (part of
the Fed’s discount window), which provides banks with a source of back-up funding at an
interest rate that is well above the Fed’s target federal funds rate. Thus, a bank would be
unlikely to borrow in the federal funds market at a rate above the primary credit rate because it
could instead obtain the funding it needs more cheaply by borrowing directly from the Fed.3

Figure 1
Banks’ demand for and the Fed’s supply of reserve balances

The supply curve for reserve balances is vertical because the Fed is a monopolistic supplier of
reserves; the supply curve shifts to the right or left when the Fed adds or subtracts reserves
from the banking system using OMOs. 4 The intersection of the demand and the supply curves
occurs at the market federal funds rate.

3
Although, in theory, banks should be unwilling to pay more than the primary credit rate for overnight funding,
they sometimes do. Transactions costs as well as reputation effects (termed “stigma”) associated with borrowing
from the Fed lead some banks to borrow from other institutions in the federal funds market at interest rates that
exceed the primary credit rate. Data on banks’ aggregate borrowings from the Fed are published weekly on the
H.3 Statistical Release. For more information on the Fed’s discount window programs, see Purposes and Functions
(2005).
4
Note that when banks trade existing reserve balances among themselves in the federal funds market, that
trading leaves the aggregate amount of reserve balances unchanged.

Page 4 of 29
A key feature of the supply and demand framework for reserves prior to the financial crisis is
that the supply curve intersected the demand curve on the downward-sloping portion of the
demand curve. This meant that the Fed was able to achieve the FOMC’s desired target level for
the federal funds rate by using OMOs to adjust the supply of reserve balances. In particular, if
the market federal funds rate was above (below) the FOMC’s target federal funds rate, then the
Fed would execute purchases (sales) of securities that would add (drain) reserve balances to
(from) the banking system and shift the supply curve to the right (left). 5 Each business day the
Federal Reserve examined demand and supply conditions and, informed by staff models,
determined whether an adjustment to reserve supply was needed, which kind of OMO was
suitable, and what approximate size would be appropriate; it then executed one or more small
operations to produce the desired movement in supply necessary to achieve the Committee’s
target federal funds rate (see Box 2 for more information about the kinds of OMOs the Fed
conducted). 6

Box 2: What kinds of OMOs did the Fed traditionally conduct?

The kind of OMO the Fed would use prior to the financial crisis depended importantly on its assessment
of conditions in the market for reserves. For example, suppose banks were trading federal funds at
interest rates above the FOMC’s target federal funds rate. In this situation, the Federal Reserve—more
specifically, the Open Market Trading Desk at the Federal Reserve Bank of New York (the Desk)—would
purchase a security from the private sector, a transaction that cleared through banks and resulted in
reserve balances being added to the banking system. In Figure 1, this transaction would shift the supply
curve to the right for as long as the Fed owned the security, and thereby put downward pressure on the
market federal funds rate. (The Appendix describes the mechanism by which increases in the Fed’s
security holdings result in a commensurate increase in the amount of reserve balances in the banking
system.)

If the Desk thought that the demand for reserves was elevated only temporarily, perhaps around long,
holiday weekends when banks often hold extra reserves, it could instead use an OMO to purchase the
security for only a short time, thereby increasing reserve balances temporarily, with the transaction
including a step to sell the security back to the original holder at an agreed-upon time in the future.
Such a transaction is known as a “repurchase agreement,” also referred to as “repo” or “RP.” 7 All else

5
More specifically, the Fed kept the “effective” federal funds rate near the FOMC’s target level. The effective
federal funds rate is a volume-weighted average of federal funds transactions. The Federal Reserve Bank of New
York publishes data on the effective federal funds rate each day. For simplicity, we refer to this rate as the market
federal funds rate, or just the federal funds rate (to distinguish from the FOMC’s target federal funds rate).
6
For a discussion of how OMOs were forecast, see Judson and Klee (2010).
7
The market for repos is complex. There are two basic types of repo transactions: “bilateral” and “tri-party,”
referring to the number of participants involved in the transaction. Within the tri-party market there is a large
segment called the GCF repo market, referred to in Figure 2. For more information on the structure of repo
markets, see Copeland et al. (2012).

Page 5 of 29
equal, once the repurchase agreement was concluded, with the security having made its round-trip back
to the private sector, the supply curve for reserve balances would also be back where it started.

The Fed could also drain reserve balances temporarily if it needed to do so using a reverse repurchase
agreement, also known as a “reverse repo” or “RRP.” In this type of OMO, illustrated in the figure
below, the Desk would sell a security to the private sector, a transaction that would initially result in a
decline in the quantity of reserve balances in the banking system, shifting the supply curve to the left.
As with a repo transaction, this transaction would include a second step in which the transaction is
unwound—the Desk would repurchase the security at a specified price at an agreed-upon time in the
future and return the funds it had been holding, leaving reserve balances back where they started. In
either a repo or reverse repo transaction, the difference between the sale price and the repurchase
price of the security, together with the length of time between the sale and purchase steps of the
transaction, implies a rate of interest earned by the party that purchased the security and loaned the
funds. 8

Illustration of an RRP transaction

As shown in Figure 2, the federal funds rate and other short-term market interest rates
continue to move together. This reflects, in part, the fact that many of the same financial
institutions are active participants in the market for various money market instruments,
including federal funds, Eurodollar, and repo markets. 9 In particular, as shown in Table 1, banks

8
Technically, when the Fed buys a security in a repo transaction or sells a security in a reverse repo transaction the
size of its securities holdings is unchanged, in accordance with generally accepted accounting principles. However,
these transactions do temporarily change the composition of the Fed’s balance sheet while the trades are
outstanding. For example, a reverse repo transaction shifts funds out of reserve balances and into reverse repos,
resulting in a compositional change in the Fed’s liabilities and no change to its assets. For more information, see
the Appendix and the FAQ on the Federal Reserve Bank of New York’s website.
9
A Eurodollar transaction is very similar to a federal funds transaction—both are unsecured trades. While federal
funds are U.S.-dollar deposits at U.S. banks, Eurodollars are U.S.-dollar deposits at a bank outside of the United
States. A repo transaction, on the other hand, is a secured trade (details are provided in Box 2).

Page 6 of 29
are active borrowers in all three of these money markets, and while the lenders vary a bit
across the markets, there is also notable overlap. All in all, arbitrage generally works well to
keep money market rates highly correlated.

Figure 2
Overnight market interest rates

Source: For Eurodollar, Bloomberg; for Treasury GCF Repo, Depository Trust & Clearing
Corporation (DTCC); for Federal Funds, Federal Reserve Bank of New York; for Interest on Excess
Reserves, Federal Reserve Board.

Table 1
Major borrowers and lenders of cash in money markets

Repo and
Federal funds Eurodollar
Reverse Repo

Banks
Borrowers of cash Banks Banks
Securities dealers

Money market funds

Banks Money market funds
Securities dealers
Lenders of cash Securities dealers Other financial firms
GSEs*
GSEs* Nonfinancial firms
Hedge funds
* GSEs are government-sponsored enterprises such as Fannie Mae, Freddie Mac, and the Federal Home
Loan Banks (FHLBs).

Moreover, changes in the level of short-term interest rates over time have generally been
transmitted to other, longer-term interest rates as well, including those commonly faced by
businesses and households, thereby ultimately influencing the pace of economic activity. As
conditions in the economy changed over time, the Committee adjusted monetary policy
Page 7 of 29
accordingly, typically by raising or lowering its target for the federal funds rate, so as to foster
economic conditions it judged to be consistent with achieving its statutory goals of maximum
employment, stable prices, and moderate long-term interest rates. An increase in the target
federal funds rate is generally considered a “tightening” of monetary policy because higher
interest rates tend, all else equal, to discourage economic activity by making borrowing costs
more expensive for businesses and households; conversely, a reduction in the target federal
funds rate is considered an “easing” of monetary policy. This basic transmission channel of
monetary policy is depicted in Figure 3.

Figure 3
Transmission channel of monetary policy

3 How did the global financial crisis affect the Fed’s operational framework?
The first event commonly associated with the financial crisis took place in early August of
2007.10 As shown in the left panel of Figure 4, on the eve of that event, the Federal Reserve’s
assets were comprised principally of Treasury securities holdings of $791 billion; its liabilities
were mainly currency, with banks holding about $15 billion in reserve balances at the Federal
Reserve. At the onset of the financial crisis, the FOMC began reducing its target for the federal
funds rate.11 Over a short period of time, the target rate was moved down substantially from
its peak level just prior to the crisis of 5¼ percent, where it stood in August 2007, to its effective
lower bound of 0 to 25 basis points in December 2008, where it still stands today. 12
In addition, to counteract the devastating effects on the U.S. economy of the financial crisis and
subsequent Great Recession, the FOMC carried out a series of large-scale asset purchase
programs (LSAPs) between November 2008 and October 2014 in which the Fed purchased in
the secondary market about $1,690 billion in Treasury securities, $2,070 billion in agency
mortgage-backed securities (MBS), and $170 billion in debt issued or guaranteed by

10
On August 9, 2007, BNP Paribas suspended withdrawals from three of its investment funds due to problems in
the subprime mortgage market.
11
The Federal Reserve also responded to the financial crisis with a number of credit and liquidity programs
designed to support the liquidity of financial institutions and foster improved conditions in financial markets.
Although these programs led to significant increases in the Federal Reserve's balance sheet, the programs have
expired or were concluded, and they are not boosting the Fed’s balance sheet today. Details of these liquidity
programs are available on the Federal Reserve Board’s website.
12
The target or “intended” federal funds rate is published on the Federal Reserve Board’s website.

Page 8 of 29
 Figure 4
A simplified Federal Reserve balance sheet: Before and after the financial crisis
(billions of dollars)

Before: After:
August 8, 2007 December 24, 2014

Assets Liabilities Assets Liabilities

Reserve Reserve
Securities 791 14 Securities 4,247 2,610
balances balances
Other Other
78 Currency 777 262 Currency 1,294
assets assets
Other 45 Other 548
Capital 34 Capital 57
Total 869 Total 869 Total 4,509 Total 4,509

government agencies.13 With short-term interest rates already near zero, the purpose of these
operations was to put downward pressure on longer-term interest rates in the economy (see
Figure 3)—the purchases reduced the available supply of securities in the market, leading to an
increase in the prices of these securities and a reduction in their yields. 14
In order to maintain these effects on yields, the Fed has been keeping its securities holdings
steady by reinvesting principal payments on agency debt and agency MBS and by rolling over
Treasury securities as they mature. As a result of these LSAPs and the FOMC’s securities
reinvestment policy, as illustrated in the right panel of Figure 4, the Fed’s securities holdings
rose to nearly 5½ times their pre-crisis level, with agency MBS representing about 40 percent of
the securities portfolio. 15 In addition, reserve balances became the largest liability, amounting
to $2.6 trillion. As we will explain shortly, in an environment with a superabundant level of
reserves, the Federal Reserve can no longer rely on small changes in the supply of aggregate
reserves to adjust the level of the federal funds rate.
Another important factor affecting the federal funds market and thus the implementation of
monetary policy going forward is the introduction of interest payments on reserve balances.
Since October 2008, the Federal Reserve has paid interest on the balances that banks maintain

13
In addition, the Fed conducted a maturity extension program where it sold $667 billion in shorter-dated Treasury
securities and purchased the same amount of longer-dated Treasury securities. More details regarding all the
purchase programs may be found on the Board’s website.
14
The LSAPs also helped to support mortgage markets. For more information on LSAPs, see the Board’s website.
15
The increase in securities holdings on the Fed’s balance sheet between August 2007 and December 2014 is less
than total securities purchased because the FOMC did not implement its reinvestment policy until August 2010.

Page 9 of 29
to satisfy their reserve requirements and on banks’ excess balances. 16 Since the Fed began
paying interest on reserves, the market federal funds rate has generally been below the IOER
rate. This situation has arisen because, in addition to banks not needing to borrow actively
from each other because of the high quantity of reserves in the banking system, there are also
nonbank lenders in the federal funds market who, by law, are not eligible to earn the IOER rate
on the balances they keep at the Fed (see the first column of table 1). Thus, those nonbanks
have an incentive to lend reserves at any rate above zero while banks have an incentive to
purchase federal funds at rates below the IOER rate in order to earn the spread between the
market rate at which they purchased the funds and the IOER rate they earn from the Fed on
those funds. This is an example of an arbitrage transaction, and we discuss the role of arbitrage
in more detail below.
Figure 5 compares the market for reserve balances before the financial crisis—repeating Figure
1 in panel (a), on the left—with the current environment, depicted by panel (b) on the right.
There are two key differences between the panels. The first is that with the Fed paying interest
on reserves, the lower portion of banks’ demand curve flattens out near the IOER rate,
reflecting the arbitrage activity just described. The second difference is that the supply curve in
panel (b) is shown far to the right on the x-axis, representing the superabundant level of
reserves in the banking system. As a consequence, the supply curve now intersects the
demand curve on the flat portion of the demand curve. The FOMC has said that when
economic conditions and the economic outlook warrant a less accommodative monetary policy,
it will raise its target range for the federal funds rate. 17 But with the supply curve in its current
position, the steps that the FOMC traditionally took to put upward pressure on market interest
rates—announcing a higher target level for the federal funds rate and being prepared to
conduct an OMO of the sort that the Fed has traditionally done to ensure the federal funds rate
hit the FOMC’s new target level (that is, possibly selling a small amount of securities into the
market and draining an equally small amount of reserves)—will no longer suffice. The key
question is, when the FOMC decides it is time to begin removing monetary policy
accommodation by raising its target range for the federal funds rate, how will the desired
increase in short-term interest rates be accomplished?

16
The Financial Services Regulatory Relief Act of 2006 authorized interest payments on reserve balances beginning
in 2011, and during the financial crisis, the Emergency Economic Stabilization Act of 2008 advanced the effective
date of this authority to October 2008. The Federal Reserve pays interest on reserve balances that banks hold to
meet reserve requirements and on their excess balances. These rates are currently the same, although they could
be set at different levels.
17
The FOMC outlined its plans for policy normalization in its Policy Normalization Principles and Plans that it issued
to the public after its September 2014 meeting. The FOMC provided additional detail regarding its plans in the
minutes of the March 2015 meeting. These FOMC communications are discussed in section 5.

Page 10 of 29
 Figure 5
Banks’ demand for and the Fed’s supply of reserve balances

4 What tools could the Fed use to raise interest rates?

The Federal Reserve has many policy tools it can rely upon to help move the federal funds rate
up. Some of these tools have been used in some form for decades and others are relatively
new. In this section, we begin by reviewing in general terms the ways in which the tools are
expected to influence the federal funds rate, and then provide more specific information about
how each of the policy tools works in terms of these channels.
We focus on three key channels of influence and the role that each will likely play in raising
short-term interest rates when the time comes:

• Encourage arbitrage: A policy tool can encourage arbitrage in money markets when it
offers an interest rate that acts as a reservation rate, the lowest rate of return that a
financial institution would be willing to accept for investing its funds when assessing
available investment opportunities. 18 Generally speaking, financial institutions with
access to a given policy tool have an incentive to borrow funds in money markets at
rates that are below the interest rate that the Federal Reserve offers on the policy tool
and invest the funds in the policy tool, putting upward pressure on money market rates.
• Increased scope of influence: If a policy tool provides a reservation rate to a broader set
of financial institutions than banks, we say it has an increased scope of influence in
money markets. Access to this tool will narrow the set of institutions that might lend

18
Note that the Fed is a risk-free counterparty. In general, when two institutions engage in a financial transaction,
there is some risk that one of the parties will not live up to its contractual obligations. Because of this risk—
commonly known as default risk—institutions seek compensation in the form of additional return on their
investment (known as a risk premium). Because there is no risk that the Federal Reserve will be unable to return
money, banks do not require additional compensation in the rates they earn from the Fed.

Page 11 of 29
 money below the rate earned on the policy tool and put upward pressure on the lowest
interest rates in money markets.
• Increase reserve scarcity: Use of the policy tool can increase reserve scarcity by draining
reserve balances and moving the level of aggregate reserves closer to its traditional
position. If the aggregate level of reserve balances were reduced sufficiently, banks
would need to resume borrowing federal funds to meet their demand for reserve
balances, leading them to put upward pressure on the market federal funds rate.
Table 2 summarizes the channels through which each of the policy tools, described in turn
below, aims to increase short-term interest rates. Overall, each of the policy tools is generally
designed to do one or both of the following things: (1) Provide an investment option that acts
as a reservation rate that market participants will factor into their investment decisions; (2)
reduce the quantity of reserve balances in the banking system.
Table 2
The channels through which the policy tools put upward pressure on rates

Increased Increase
Policy tools Encourage
scope of reserve
(described below) arbitrage
influence scarcity
Increase IOER rate
Offer RRPs (overnight or term)
Offer term deposits
Reduce Fed’s securities holdings
Increase reserve requirements

4.1 Rate of interest on excess reserve balances

The FOMC has indicated that the Federal Reserve intends to move the federal funds rate into
the target range set by the FOMC primarily by adjusting the rate of interest on excess reserve
balances or the IOER rate. 19 The IOER rate encourages arbitrage—it acts as a reservation rate
for banks as they make their money market investment decisions. As described above, all else
equal, an increase in the IOER rate would be expected to put upward pressure on the federal

19
Note that the Federal Reserve has designated two rates of interest on reserve balances, one rate for required
reserve balances (the IORR rate) and a separate rate for excess reserve balances (the IOER rate); for simplicity and
given the predominance of excess balances, we refer to the IOER rate throughout this piece. For a time in 2008,
both the IORR and IOER rates were determined by a formula linked to the federal funds rate and set at different
levels; in December 2008, the FOMC reduced the federal funds rate target to a range of 0 to 25 basis points and
set the IORR and IOER rates equal to 25 basis points.

Page 12 of 29
funds rate because banks would have an incentive to borrow in the federal funds market at
rates below the IOER rate and place those balances at the Fed to earn the IOER rate. 20
Similarly, other money market rates should increase as banks arbitrage between holding excess
reserve balances and these alternative money market instruments.

Of course, banks need to be willing and able to actively perform this arbitrage for these effects
to be realized. To date, banks have been willing to arbitrage the IOER rate so that, as shown in
Figure 2, the federal funds rate has been highly correlated with other money market rates. But,
as also illustrated in the figure, these money market rates have remained below the IOER rate
because, as noted above and in Table 1, banks are not the only participants in the federal funds
market. Government-sponsored enterprises (GSEs), institutions that also hold reserve accounts
at the Fed but are not eligible to earn interest on those balances, also participate in the federal
funds market. These institutions are willing to lend out federal funds at rates that are relatively
low because their return on balances held at the Fed is zero. 21 Moreover, because reserve
balances are superabundant among banks, nearly all of the federal funds trading reflects
borrowing by banks from non-IOER-earning institutions, at relatively low rates, to engage in
such arbitrage activity. Next we discuss a tool that was designed to support short-term interest
rates from below.

4.2 Overnight RRPs

The FOMC has indicated that, when the time comes to begin to raise the target federal funds
rate, it intends to use an overnight reverse repurchase agreement facility as needed to help
keep the federal funds rate within its target range. Box 2 provided a description of how the Fed
conducted RRP transactions prior to the financial crisis, and noted that such transactions were
conducted for the purposes of causing a temporary decline in reserve balances in order to put
upward pressure on the federal funds rate. At that time, the Fed occasionally conducted
relatively small-dollar amounts of overnight RRPs or “ON RRPs” with a group of institutions
known as “primary dealers.” 22 Today, the Fed is routinely conducting ON RRPs in the form of
test exercises; these ON RRPs have three key operational differences relative to their use in
monetary policy operations before the financial crisis.
The first difference is that ON RRPs have been offered on a daily basis with an offering rate that
is pre-announced and can act as a reservation rate, encouraging arbitrage in money markets. In
particular, the offering rate is the maximum interest rate the Fed is willing to pay in the

20
In conducting this arbitrage activity, reserve balances are shuffled among banks; this activity will not bring about
a change in the aggregate quantity of reserve balances in the banking system (see the Appendix for more detail).
21
See Goodfriend (2015).
22
A primary dealer makes markets in Treasury securities; that is, it buys (sells) Treasury securities directly from (to)
the government with the intention of acting as the “middleman” between the government and market participants
in the private sector.

Page 13 of 29
operation. As was the case with the payment of interest on excess reserves, the Fed’s ON RRPs
is an investment option that potential participants take into account when deciding which of
various money market instruments to invest in. Counterparties will compare the Fed’s ON RRP
offering rate to other money market rates and determine whether to invest in those other
money market instruments or instead to bid in the Fed’s ON RRP operation (see Box 3 for a
more detailed description). Because the Fed sets the ON RRP rate, it can influence the extent
to which money market participants consider ON RRPs an attractive investment option. If the
Fed’s offering rate is relatively low, demand for ON RRPs could be small. Alternatively, if the
Fed’s ON RRP rate was greater than comparable alternative interest rates in money markets,
counterparties could bid relatively large amounts. In testing ON RRPs, the Fed at times has
varied the offering rate, and this has generally demonstrated that demand for ON RRPs is
indeed sensitive to the pattern of interest rates.

The second difference is that the set of counterparties that are eligible to participate in the
Fed’s ON RRP operations is much broader than it was in the past, increasing the sphere of
influence that the ON RRP rate has in money markets. The Fed conducted traditional OMOs
with a set of primary dealers; today, the institutions that are eligible to participate in the Fed’s
ON RRP operations include about two dozen banks as well as a lengthy list of nonbanks (money
market funds, primary dealers, and GSEs). 23 As a result, more institutions—and, importantly,
more types of institutions—are able to consider the Fed’s ON RRPs a direct investment option.
In particular, the eligible nonbank institutions, which are unable to earn interest on reserves,
may be encouraged to engage in arbitrage activity relative to the ON RRP rate because they
have little incentive to lend funds in money markets at interest rates below the one they can
receive directly from the Fed. Such activity has the important effect of providing a floor under
the level of money market interest rates and supporting them from below.

The third operational difference relates to the way in which ON RRPs could be used today to
increase reserve scarcity. When the Fed announces an ON RRP test operation, it also
announces an aggregate offering amount—the total amount of dollars the Fed is willing to
accept at the operation. During the initial portion of the testing period for ON RRPs, which
began in September 2013, the aggregate offering amount on these exercises was first
uncapped, so that the Fed accepted all bids, subject to a limit on each institution’s bid, that
eligible participants wanted to place at the Fed. In September 2014, the aggregate offering size
was capped at $300 billion, where it stands today. With an aggregate cap on ON RRP
operations, the Fed needs to use a procedure, described in Box 3, to award the aggregate
offering amount when the total amount bid at an operation exceeds the total amount offered.

23
The Federal Reserve Bank of New York publishes on its website a list of the Fed’s primary dealers and eligible ON
RRP counterparties.

Page 14 of 29
The Fed’s testing of ON RRP operations has demonstrated that these operations can set a soft
floor under the level of the federal funds rate and other short-term market interest rates, as
long as market participants are confident that the aggregate cap on ON RRPs is large enough to
meet demand.

If the Fed wanted to increase the scarcity of reserves in the banking system, it could set the
offering amount on its ON RRP operation relatively high, and possibly also adjust the offering
rate, to encourage demand for these operations. However, the FOMC has discussed concerns
associated with having a persistently large ON RRP program, a topic to which we will return in
section 5. Thus, the role that ON RRPs may play in increasing reserve scarcity over time is likely
to be limited. Instead, the Fed could use other tools, such as term RRPs or term deposits to
increase reserve scarcity, and we discuss these next.

Box 3: How does the Fed determine take-up at each ON RRP operation?
Financial institutions that are eligible to engage in ON RRP operations with the Fed compare the Fed’s
offering rate on ON RRPs to market interest rates on similar assets and choose how many securities to
bid for—and thus how many dollars to place with the Fed—at each operation. During the testing phase
of the ON RRP exercises, the Fed has typically allowed eligible institutions to bid up to $30 billion at each
operation. How does the Fed decide which bids to accept? The answer depends on whether the Fed
has capped the ON RRP operation in aggregate size or not, and if so, whether the total amount bid is
above or below the cap. If an operation is not capped in aggregate or if the total amount bid is less than
the cap, all bids are accepted at the Fed’s offering rate. But if the operation is capped and the total
amount bid exceeds the cap, an auction process is used to allocate the available amount of the
operation. Next we walk through these two cases.

First, suppose that the Fed offers ON RRPs in an unlimited aggregate amount. Further, let’s suppose, as
has typically been the case during the testing phase of the ON RRP exercises, that the Fed’s offering rate
is 5 basis points. Also assume, as illustrated in the table below, that five institutions submit bids that
range between $5 and $15 billion for a total of $58 billion. With no capacity constraints on the
operation, all counterparties would be awarded their bid amount at the operation’s offering rate of 5
basis points. What this means is that these institutions would place $58 billion at the Fed overnight, and
the Fed would return the funds with interest the next day. (See the Appendix for a full analysis of an ON
RRP transaction, including its effects on reserve balances and the Fed’s securities holdings.)

Now suppose instead that the Fed wanted to limit the size of this operation to $40 billion in aggregate—
that is, each institution could still bid up to $30 billion, but the Fed wanted to award only $40 billion in
total. The Fed could do this by announcing ahead of time that there will be an aggregate cap of $40

Page 15 of 29
 Sample bids for an uncapped ON RRP operation
Bid amount
Institution
($ billions)
1 10
2 5
3 20
4 8
5 15

billion on the operation. In this case, the Fed would not only ask institutions to submit bid amounts, but
also bid interest rates (third column in the table below). The FOMC would sort the bids from the lowest
bid interest rate to the highest bid interest rate, and fill each requested bid amount until the $40 billion
in take-up had been reached, prorating any bids at the highest bid rate needed to achieve $40 billion. In
this case, institutions 1, 2, and 3 would receive their full bid amounts of $10, $5, and $20 billion, for a
total of $35 billion. To award the remaining $5 billion, institution 4’s bid amount would be reduced from
$8 billion to $5 billion. (If more than one institution had bid 4 basis points in this example, the
remaining $5 billion would be prorated among those institutions according to the share of the total
amount bid at 4 basis points that is accounted for by each institution). Because the maximum operation
size of $40 billion was met by the Fed paying 4 basis points, all amounts would be awarded at 4 basis
points, referred to as the “stop-out rate.”

Sample bids for a capped ON RRP operation

Bid amount Bid rate
Institution
($ billions) (basis points)
1 10 2
2 5 3
3 20 3
4 8 4
5 15 5

Both types of operation limits—an institution-level cap and an aggregate cap—have been used during
the testing phase of ON RRP operations, with the institution-level cap typically set at $30 billion and the
aggregate cap set at $300 billion. The Federal Reserve has been reporting the results of its daily ON RRP
test operations, including the bid amounts submitted and accepted, as well as the high, low, and
awarded bid rates. 24 On September 30, 2014, for example, demand for Fed ON RRPs was more than
$400 billion. The operation’s aggregate cap was $300 billion, resulting in some counterparties having
their bids prorated or declined, and the resulting stop-out rate was 0 basis points.

The results of the most recent ON RRP operation may be found on the Federal Reserve Bank of New York’s
24

website.

Page 16 of 29
4.3 Term RRPs
In addition to conducting RRPs on an overnight basis, the Fed can conduct the same type of
transaction with the same set of eligible financial institutions but with the RRP outstanding over
a longer time period; such transactions are known as “term RRPs.” In general, term RRPs work
through the same three channels as the overnight operations, but drain reserves from the
banking system for a longer period of time.
The Fed has been testing term RRPs for some time, although less regularly than the ON
operations, as shown in Figure 6. During the testing period, the Fed has offered term RRPs at
varying rates and maturities, with most operations timed to cover quarter-end dates, a time
when other investment options in money markets tend to dwindle. 25 About $200 billion of
term RRPs were outstanding at year-end 2014 and March-end 2015 on average. 26 Testing has
also showed that term RRPs serve in part as a substitute for ON RRPs—that is, take-up of term
RRPs has tended to reduce the demand for ON RRPs to some extent.
Figure 6
Total ON RRP and term RRP in test operations

4.4 Term Deposit Facility (TDF)

The Fed may also choose to offer interest-bearing deposits to banks through its Term Deposit
Facility or “TDF.” When a bank elects to place funds in the TDF, the funds are moved out of
reserve balances for the life of the term deposit. Thus, the TDF acts to increase reserve

25
Investment options for major cash lenders tend to dwindle at quarter-ends because some large banking
institutions reduce the size of their balance sheets—that is, they tend to borrow less and accommodate less
investment activity of other institutions—at that time in light of regulatory reporting requirements.
26
The Federal Reserve also regularly conducts RRPs for international organizations, and the amount of outstanding
RRPs reported on the Fed’s balance sheet includes about $100 billion of these transactions. The Federal Reserve’s
balance sheet is published weekly on the H.4.1 statistical release.

Page 17 of 29
scarcity. The TDF also encourages arbitrage as banks compare the yield the Fed offers on a
term deposit with other investments of a similar term.

The Fed has been testing the functionality of the TDF since June 2010. During these operational
tests, two types of term deposit operations have been conducted. In the first type, the Fed
offers a given dollar amount of term deposits; banks then bid for the size of the deposit they
want and specify the interest rate on the deposit. The Fed accepts bids beginning with the
lowest bid rate and proceeding to higher bid rates until the total offered amount is exhausted.
When this type of term deposit operation is conducted, all banks receive an interest rate that is
identical to the rate paid to the last bank whose bid was accepted—that is, all banks receive the
highest bid rate. In the second type of term deposit operation, the Fed sets an offering rate
and allows banks to deposit the amount of funds they desire, up to a predetermined maximum.

In its test operations, the Fed has varied some features, including the length of the term, the
offering rate, and whether banks are permitted to withdraw their deposits prior to the end of
the term, subject to a penalty. 27 The latter feature has proven to be particularly attractive to
banks in making their cash management decisions. During testing in February 2015, term
deposits outstanding grew to about $400 billion on the Fed’s balance sheet.

4.5 Reduce the Fed’s securities holdings

As noted above, the Federal Reserve purchased large quantities of securities during its large-
scale asset purchase programs and it has been reinvesting any maturing or prepaying securities
in order to maintain its holdings of longer-term securities at sizable levels and help maintain
accommodative financial conditions. The FOMC has said that, at some point after it begins to
increase the target range for the federal funds rate, it will reduce the Federal Reserve's
securities holdings in a gradual and predictable manner, primarily by ceasing to reinvest
repayments of principal on securities held by the Federal Reserve.
Ending reinvestments would cause the Fed’s holdings of securities to decline when either a
Treasury security matures or an agency MBS prepays. The pace at which this would occur is
driven in part by the maturity dates of the Fed’s holdings of Treasury and agency securities,
which are known for certain, and also by the pace at which agency MBS might prepay, which
can only be estimated. Prepayment of agency MBS occurs, for example, when households pay
off some or all of a mortgage balance early because they refinance their original mortgage with
a lower available mortgage rate, pay off their mortgage in full when they sell a house to move,
or pay down a portion of their mortgage to reduce the level of their debt.
As discussed below in Box 4, the Fed’s securities holdings could decline noticeably once the
FOMC decides to end reinvestments. Doing so would naturally shift the reserve supply curve

27
Results of all term deposit operations can be found on the Federal Reserve Board’s website.

Page 18 of 29
gradually to the left, increasing reserve scarcity. Of course, it would take a number of years for
the quantity of securities to decline sufficiently to create a meaningful inward shift in the supply
curve—that is, one that would be effective in helping to put upward pressure on the federal
funds rate.

Box 4: What would stopping reinvestments imply for the Fed’s securities holdings?
As of late December 2014, the Federal Reserve held a total of $4.2 trillion of securities, of which about
$2.5 trillion were Treasury securities, $1.7 trillion were agency MBS, and about $39 billion were agency
debt. The Federal Reserve Bank of New York regularly reports details regarding the Fed’s securities
portfolio, including the Committee on Uniform Securities Identification Procedures (CUSIP) number of
each Treasury security and agency MBS the Fed holds in its portfolio. 28

If the FOMC were to end its policy of reinvesting its securities holdings at some point over the near
term, the Fed’s holdings of securities would decline noticeably. As shown in the table below, nearly
$700 billion of securities would mature or roll off of the Fed’s portfolio in 2016 and 2017 taken together,
comprised of about $410 billion of Treasury securities and an estimated $285 billion of agency MBS.

Absent reinvestment, how many securities would mature or roll off in 2016-17?
(billions of dollars)

Date Treasury securities Agency MBS* Total

2016: H1 129 85 214

2016: H2 86 75 161

2017: H1 103 65 168

2017: H2 91 60 151

Cumulative 409 285 694

* Using actual holdings of securities as of December 24, 2014, and projected agency MBS prepayments
based on the model in Carpenter et al. (2015) along with Blue Chip interest rate projections at that time.

The Fed could also sell securities in order to reduce its holdings, which would cause reserve
balances to decline commensurately. However, the FOMC has indicated that it will raise
interest rates by taking actions to move the federal funds rate into the new target range rather
than through actions to adjust the size or composition of the Fed’s balance sheet. In particular,
in section 5 we discuss why securities sales are not a feature of the FOMC’s chosen approach.

28
These details may be found on the Federal Reserve Bank of New York’s website.

Page 19 of 29
4.6 Reserve requirements
In the past, the Federal Reserve has adjusted reserve requirements infrequently and not used
them as an active tool in monetary policy implementation. 29 Although the FOMC has not
indicated that it is considering adjusting reserve requirements when it raises short-term
interest rates, we nonetheless include this tool here for completeness. As was discussed above
in Box 1, reserve requirements are calculated as fractions of certain deposit account levels, and
have traditionally helped to create demand for reserve balances. These reserve ratios are re-
evaluated annually, although they have not been adjusted since 1992. In theory at least, the
Fed could increase existing reserve ratios in order to require that additional reserve balances be
held by banks, thereby contributing to the scarcity of excess reserve balances. 30

5 What is the FOMC’s preferred approach for policy “normalization”?

Recall that before the financial crisis, the Fed’s approach to implementing monetary policy
relied on the reserve scarcity channel: Given banks’ demand for reserves balances and the
relatively low quantity of reserves in the banking system, the Fed could make small adjustments
to the aggregate supply of reserves using OMOs in order to cause the supply of reserves to
intersect banks’ demand curve at the FOMC’s target federal funds rate. And recall that with the
superabundant level of reserve balances in the banking system that exists today, the Fed
cannot rely on the reserve scarcity channel to influence the level of short-term interest rates
when it begins to raise the target range for the federal funds rate. So, how will the Fed tighten
monetary policy going forward?
Returning to panel (b) of Figure 5, in theory the Fed could take steps to shift the supply of
reserve balances left until it intersects the downward-sloping portion of the demand curve
again—that is, return the banking system to a position of sufficient reserve scarcity. But that
would require the Fed either to rely on sizable operations through its ON RRP facility for an
extended period, run very significant term RRP and TDF operations, sell a substantial quantity
of its securities holdings, or undertake all of these activities in some combination. Instead, the
FOMC has chosen an approach to implementing monetary policy that relies primarily on the
other two key channels of influence on interest rates that we discussed—that is, it plans to use
its policy tools both to encourage arbitrage and to increase its scope of influence in short-term
money markets.
Many readers might ask, why doesn’t the Fed reverse course? That is, why doesn’t the Fed sell
securities in order to shrink the supply of reserve balances substantially so as to permit the Fed

29
Like the Federal Reserve, the European Central Bank and the Bank of Japan also impose reserve requirements
but do not use them as an active tool of monetary policy.
30
The Fed may also, after consulting Congress, impose so-called emergency reserve requirements if it finds that
extraordinary circumstances require such action, a step that the Fed has never before taken. Authorization for
emergency reserve requirements is given in Section 204.5 of the Federal Reserve Board’s Regulation D.

Page 20 of 29
to implement monetary policy as it did before the financial crisis? As noted above, the FOMC
has indicated that when it begins to remove policy accommodation, it prefers to target the
federal funds rate, as it did prior to the financial crisis, instead of affecting longer-term interest
rates by actively managing the size and composition of its balance sheet. This judgment reflects
a couple of considerations. First, it would take some time to sell a large fraction of the Fed’s
securities holdings and so such a tactic is not sufficiently nimble. For example, if the chosen
pace of the Committee’s securities purchases under its LSAPs programs is any guide, it would
take a number of years to sell a sufficient quantity of securities so as to cause a meaningful
inward shift in the reserves supply curve. Moreover, such sales might bring about unwanted
effects in financial markets—for instance, the upward pressure on market interest rates that
would accompany sizable sales of the Fed’s holdings of securities could be hard to gauge and
control. And second, monetary policy around the globe has generally been implemented in
short-term financial markets. For example, the FOMC has many years of experience affecting
conditions in the federal funds market, an overnight market, and prefers to continue to
implement its policy through this market even while its balance sheet is large.
The FOMC has formulated and issued plans regarding the approach it intends to take when it
decides that the time has come to begin raising short-term interest rates, one part of a process
that it refers to as policy “normalization.” 31 When the FOMC decides that economic conditions
warrant the commencement of the policy normalization process, it will tighten the stance of
monetary policy by adjusting the target range for the federal funds rate, a step that some have
nicknamed interest rate “liftoff.” Framing the stance of policy in terms of the federal funds rate
focuses the Committee’s monetary policy communications on an interest rate that is familiar to
the public. The FOMC has also indicated that it plans to continue to set a target range for the
federal funds rate that is 25 basis points wide, something it has been doing since December
2008. Continuing to maintain a target range for the federal funds rate will be an effective way
for the Fed to communicate its policy stance and is consistent with the FOMC’s approach in
recent years. Targeting a particular level of the federal funds rate, something the FOMC did
before the financial crisis, would likely not convey very much more information about the
Committee’s intentions for the general level of short-term interest rates so long as the quantity
of reserve balances remains highly elevated.
The primary policy tool that the FOMC will use to move the federal funds rate into its new
target range is the IOER rate; at liftoff, the IOER rate will be set to the top of the new federal
funds target range. As described above, increases in the IOER rate will help to pull the federal
funds rate and other short-term market interest rates into the target range via arbitrage.
The FOMC will also use an ON RRP facility as a supplementary tool to help to push money
market interest rates up from below both by encouraging arbitrage and by having an increased

31
See the Policy Normalization Principles and Plans that were issued following the September 2014 FOMC meeting
and augmented in the minutes of the March 2015 FOMC meeting. Potter (2015) provides an explanation for how
policy normalization will be implemented.

Page 21 of 29
scope of influence in money markets (remember that the Fed can undertake its ON RRP
operations with a different set of financial institutions than are eligible to earn the IOER rate).
The Fed will offer the ON RRP rate on its daily operations; at liftoff, the ON RRP rate will be set
to the bottom of the new target range.
The Fed will also need to set an offering amount of ON RRPs. The appropriate size of the ON
RRP facility has been an important consideration of the FOMC. The FOMC has discussed the
concern that a large and persistent ON RRP program could permanently alter patterns of
borrowing and lending in those markets—a concern the FOMC has referred to as increasing the
Federal Reserve’s role or size of its “footprint” in money markets. In addition, in times of stress
in financial markets, demand for a safe and liquid central bank asset might increase sharply,
potentially causing or exacerbating disruptions in the availability of funds to other participants
in money markets. To mitigate these two concerns, the FOMC has agreed that it will use an ON
RRP facility only to the extent necessary and will phase the facility out when it is no longer
needed.32 Nonetheless, balancing the need for keeping control over short-term interest rates
against the risks associated with a large ON RRP facility, the FOMC has determined that when
policy normalization commences, the aggregate amount offered through its ON RRP facility will
be temporarily elevated in order to help move the federal funds rate into its new target range.
The Committee has also said that it expects that it will be appropriate to reduce the capacity of
the facility fairly soon after it begins raising interest rates. 33
With a target range of 25 basis points and the IOER rate set to the top of that range and the ON
RRP rate to the bottom, there will be a 25-basis-point spread between these two administered
overnight rates at liftoff. A spread of this size is expected to be narrow enough to allow
sufficient control over short-term market interest rates but wide enough to keep the ON RRP
facility from becoming so attractive that its potential size invokes concerns about the size of the
Fed’s footprint in money markets or poses risks to financial stability.
Figure 7 illustrates how this will work. The region on the left represents the position of the
target range for the federal funds rate and the Fed’s two overnight administered rates as they
are set today, prior to liftoff, with a target range of 0 to 25 basis points (the shaded region), the
IOER rate at 25 basis points (the green line), and the ON RRP rate at 5 basis points (the red line).
Assuming, as depicted in the region to the right, that the new target range that is likely to
prevail when the FOMC decides to commence policy normalization is 25 to 50 basis points, the
IOER and ON RRP rates would be increased to 50 basis points and 25 basis points, respectively.
The increases in these two administered rates will act to move the market federal funds rate
and other short-term market interest rates up. So when it comes time to raise interest rates for

32
Details of the FOMC’s discussion of the footprint and financial stability issues associated with an ON RRP facility
are available in the minutes of the FOMC’s April 2014, June 2014, and July 2014 meetings. For further analysis of
these issues, see Frost et al. (2015).
33
See the minutes of the FOMC’s March 2015 meeting for a discussion of the options the FOMC considered for
setting the aggregate size of the ON RRP facility in the early stages of the policy normalization process.

Page 22 of 29
the first time, the Fed will simultaneously raise the target range for the federal funds rate and
increase both of these administered rates (the IOER and ON RRP rates) in order to move short-
term market interest rates higher. 34

Figure 7
The Fed’s administered rates and the federal funds (FF) target range

Don’t let the position of the IOER rate in Figure 7 confuse you. Recall that we explained earlier
that with a superabundant level of reserves in the banking system, the market federal funds
rate has been below the IOER rate since late 2008 because institutions that are not eligible to
earn interest on reserves have an incentive to lend reserves in the federal funds market at rates
well below the IOER rate. However, if reserves in the banking system were to become
sufficiently scarce again, banks would need to actively borrow reserve balances and that would
push up the federal funds rate relative to the IOER rate in Figure 7.
In section 4, we discussed several other policy tools that are available for the FOMC’s use. The
FOMC has said that it may use other supplementary tools if it needs to; this would possibly
include term RRP operations and term deposits. In addition, the Committee has indicated that
it plans to reduce its securities holdings in a gradual and predictable way by stopping the
reinvestment of maturing and prepaying securities at some point after the start of the policy
normalization process. As discussed in Box 4, securities holdings would decline noticeably over
time if reinvestments were stopped, as would the supply of reserve balances. The Committee
has also indicated that securities sales will not be part of the initial package of steps that it

34
The Fed will most likely also raise the primary credit rate when it begins raising short-term interest rates. Recall
that the primary credit rate is the interest rate at which banks can borrow reserves overnight from the Fed. Since
early 2010, the primary credit rate has been set at 75 basis points, 50 basis points above the top of the current
range for the target federal funds rate. Given that reserves are now superabundant and will remain so for some
time, banks generally will not need to borrow from the Fed and so are unlikely to be influenced by the level of the
primary credit rate. In addition, the reputational costs (stigma) associated with borrowing from the Fed are likely
much higher than was the case prior to the financial crisis in part because the Fed is now required to release
information about such borrowing to the public, albeit with a lag.

Page 23 of 29
intends to take to begin to raise interest rates, and that it does not anticipate selling agency
MBS as part of the policy normalization process although limited sales might be warranted in
the longer run to reduce or eliminate residual holdings. The Committee has emphasized that
the timing and pace of any securities sales would be communicated to the public in advance.
Of course, economic and financial developments will continue to evolve, both leading up to the
first increase in the target range for the federal funds rate and as the policy normalization
process progresses, and the Committee has said that it is prepared to adjust the details of its
approach in light of such developments.

6 Conclusion
Over the past seven years, the Federal Reserve has taken steps to counteract a financial crisis
and move the economy out of a protracted recession and back toward maximum employment
and price stability. The steps have included keeping the federal funds rate near zero since
December 2008 and purchasing securities in the open market to put downward pressure on
longer-term interest rates. These purchases resulted in a superabundant level of reserve
balances in the banking system. In this paper, we discussed why the traditional approach to
raising the federal funds rate, which was used for many years prior to the financial crisis, will
not work in the current environment. We described the toolkit available to Federal Reserve
policymakers and reviewed the Committee’s preferred approach to raising the federal funds
rate. Testing of the policy tools suggests that the proposed approach should work well. Of
course, after policymakers decide that the time has come to begin raising short-term interest
rates, they will be vigilant in using their available tools to adjust their approach, as needed, to
ensure appropriate control over the federal funds rate and other short-term interest rates.

Page 24 of 29
 Other References

Seth B. Carpenter, Jane E. Ihrig, Elizabeth C. Klee, Daniel W. Quinn, and Alexander H. Boote
(2015), “The Federal Reserve's Balance Sheet and Earnings: A Primer and Projections,”
International Journal of Central Banking, March, pp. 237-283.

Adam Copeland, Darrell Duffie, Antoine Martin, and Susan McLaughlin (2012), “Key Mechanics
of the U.S. Tri-Party Repo Market,” Economic Policy Review 18:3, Federal Reserve Bank of New
York, pp. 17-28.

Josh Frost, Lorie Logan, Antoine Martin, Patrick McCabe, Fabio Natalucci, and Julie Remache
(2015), “Overnight RRP Operations as a Monetary Policy Tool: Some Design Considerations,”
Finance and Economics Discussion Series 2015-010. Washington: Board of Governors of the
Federal Reserve System, http://dx.doi.org/10.17016/FEDS.2015.010
Marvin Goodfriend (2015), “The Fed Should Fix the Interest on Reserves Floor,” prepared for
the Shadow Open Market Committee Meeting, March, http://shadowfed.org/wp-
content/uploads/2015/03/GoodfriendSOMC-March2015.pdf.

Ruth A. Judson and Elizabeth Klee (2010), “Whither the liquidity effect: The impact of Federal
Reserve open market operations in recent years,” Journal of Macroeconomics 32:3, pp. 713-
731.
Simon Potter (2015), “Money Markets and Monetary Policy Normalization,” remarks to the
Money Marketeers, New York University, April 15.
http://www.ny.frb.org/newsevents/speeches/2015/pot150415.html

Purposes & Functions (2005), Board of Governors of the Federal Reserve System, Washington
DC, 9th edition, http://www.federalreserve.gov/pf/pf.htm.

Page 25 of 29
 Appendix
How do the Fed’s operations affect the balance sheets of the Fed, banks, and nonbanks?

A fundamental feature of U.S. monetary policy implementation is the way in which the Federal
Reserve’s operations—such as securities purchases, ON RRP transactions, and term deposits—
affect the balance sheets of not only the Fed, but also its key financial counterparties: banks
and nonbanks. The Fed’s balance sheet is reported to the public weekly on the H.4.1 statistical
release. 35 The balance sheets of banks and nonbanks tend to be reported less frequently,
typically at quarter ends. This appendix uses an accounting framework to walk through how
the Fed’s interactions with the private sector affect the balance sheets of the entities involved.

A.1 Federal Reserve securities purchases

The effects of Federal Reserve securities purchases are illustrated in Figures A1 and A2, which
display simplified balance sheets for a financial nonbank entity, a bank, and the Federal
Reserve. In the initial time period (Figure A1), a nonbank entity holds securities (SN) and
deposits (DN) on the asset side of its balance sheet, with liabilities (LN) and equity (EN)
counterbalancing. The bank holds securities (SB), loans (LNB), and reserves balances at the Fed
(R) as assets, counterbalanced with deposits (DB) and equity (EB). The Federal Reserve’s key
asset is securities (SF), main liabilities are reserves (R) and currency (C), and it holds capital (K).

Figure A1: Initial balance sheets

Nonbank entity Bank Federal Reserve

A L+E A L+E A L+K

SN LN SB DB SF R

DN EN LNB EB C
R K

A = assets; L = liabilities; E = equity; K = capital;

S = securities; D = deposits; LN = loans; R = reserves; C = currency

The shaded items in Figure A2 highlight the balance sheet effects of $1 worth of securities
purchases by the Federal Reserve from a nonbank entity. The nonbank entity sells a security
(SN ➜ SN – $1) and, in the course of clearing and settling that transaction through the banking
sector, receives deposits in return (DN ➜ DN + $1). As the bank is the intermediary for the

35
The H.4.1 statistical release may be found on the Federal Reserve Board’s website.

Page 26 of 29
transaction, it receives payment in its reserve account (R ➜ R + $1) from the Fed, and then
credits this payment to the nonbank’s deposit account (DB ➜ DB + $1). That increase in
deposits in the account that the nonbank holds at the bank is realized on the liability side of the
bank’s balance sheet. Meanwhile, the Federal Reserve ends up with the desired additional
securities on its balance sheet (SF ➜ SF + $1) and a larger reserves liability (R ➜ R + $1).

Figure A2: Balance sheets following Federal Reserve securities purchase

Nonbank entity Bank Federal Reserve

A L+E A L+E A L+K

SN - $1 LN SB DB + $1 SF + $1 R + $1

DN + $1 EN LNB EB C
R + $1 K

A = assets; L = liabilities; E = equity; K = capital;

S = securities; D = deposits; LN = loans; R = reserves; C = currency

Note that while an individual bank may take steps to reduce its reserve balances, the aggregate
quantity of reserves in the banking system generally cannot be reduced without further actions
by the Federal Reserve.36 For example, a bank can sell reserves to another bank in the federal
funds market and reduce its own holdings of such balances, but this activity leaves reserve
balances unchanged in total. Or a bank could reduce the rate of interest that it pays on
deposits with the aim of causing depositors to withdraw some of their deposits, a step that
would, all else equal, cause the bank’s reserve balances to decline to the same extent.
However, these depositors would most likely shift their funds to another bank (instead of
holding cash) and that bank’s reserve balances would then increase. Again, total reserve
balances in the banking system would be unchanged.

A.2 Federal Reserve ON RRP operations

The shaded items in Figure A3 highlight the effects of an ON RRP operation conducted by the
Fed on the same three balance sheets. Most ON RRPs are conducted with nonbank
counterparties, such as money market funds, and we illustrate this case. Figure A1 again
represents the initial time period, before the ON RRP operation is conducted. The shaded items
in Figure A3 illustrate the balance sheet effects of $1 worth of ON RRPs conducted by the Fed
with a nonbank entity. The nonbank counterparty increases its RRPs holdings (RRPN ➜ RRPN +

36
A bank could reduce its reserve balances by paying down loans from the Federal Reserve.

Page 27 of 29
$1), and in exchange lends funds to the Fed. In the course of settling the transaction through
the banking sector, the nonbank’s deposits are reduced (DN ➜ DN – $1). That reduction in
deposits is realized on the liability side of the bank’s balance sheet (DB ➜ DB – $1). Meanwhile,
the bank passes the funds of the nonbank on to the Fed, booking a reduction in reserve
balances in return (R ➜ R – $1). The Fed ends up with the desired additional RRPs on its
balance sheet (RRPF ➜ RRPF + $1), in effect a liability to repay the loan to the nonbank, and a
smaller stock of reserve balances outstanding (R ➜ R – $1). Recall that in repo transactions
(unlike a sale or purchase), the security does not actually change hands—that is, it is not
rebooked—and so the amount of securities on the Fed’s (and other institutions’) books is
unchanged.37

Figure A3: Balance sheets following Fed ON RRP operation with nonbank
Nonbank entity Bank Federal Reserve

A L+E A L+E A L+K

SN LN SB DB - $1 SF R - $1

DN - $1 EN LNB EB C
RRPN + $1 R - $1 RRPF + $1

A = assets; L = liabilities; E = equity; K = capital;

S = securities; D = deposits; LN = loans; R = reserves; C = currency;
RRP = reverse repurchase agreement

A.3 Federal Reserve term deposits

Figure A4 shows the accounting effects of awarding $1 of term deposits at the Fed’s Term
Deposit Facility, operations that are only available to banks. As before, Figure A1 represents
the initial time period, before the TDF operation is conducted. The shaded items in Figure A4
illustrate the balance sheet effects of $1 worth of term deposits conducted by the Fed with a
bank. The bank’s term deposits (“TDF” below) rise by $1 (TDF ➜ TDF + $1). The $1 in funds
placed by the bank at the TDF comes out of its reserve balances. As a result, the outstanding
stock of reserves in the banking system is smaller over the term of the deposit. The Fed
reduces the bank’s reserve balances and credits its term deposit account (R ➜ R – $1; TDF ➜
TDF + $1). Nonbank entities are not involved in the transaction.

37
See footnote 8.

Page 28 of 29
 Figure A4: Balance sheets following Federal Reserve TDF
Nonbank entity Bank Federal Reserve

A L+E A L+E A L+K

SN LN SB DB SF R - $1

DN EN LNB EB C
R - $1 TDF + $1

TDF + $1 K

A = assets; L = liabilities; E = equity; K = capital;

S = securities; D = deposits; LN = loans; R = reserves; C = currency;
TDF = Term Deposit Facility

Page 29 of 29