STABILIZATION POLICY AT THE ZERO LOWER BOUND

• Author: Christopher J. Waller,Paola Boel
• DOI: http://doi.org/10.1111/iere.12396
• Date: 01 November 2019
• Published date: 01 November 2019

INTERNATIONAL ECONOMIC REVIEW

Vol. 60, No. 4, November 2019 DOI: 10.1111/iere.12396

STABILIZATION POLICY AT THE ZERO LOWER BOUND∗

BYPAOLA BOEL AND CHRISTOPHER J. WALLER 1

Sveriges Riksbank, Sweden; Federal Reserve Bank of St.Louis, U.S.A., Deakin University,

Australia

Weconstruct a monetary economy with aggregate liquidity shocks and heterogeneous idiosyncratic preference

shocks. In this environment, not all agents are satiated at the zero lower bound (ZLB) even when the Friedman

rule is the best interest-rate policy the central bank can implement. As a consequence, central bank stabilization

policy, which takes the form of repo arrangements in response to aggregate demand shocks, temporarily relaxes

the liquidity constraint of impatient agents at the ZLB. Due to a pecuniary externality, this policy may have

beneficial general equilibrium effects for patient agents even if they are unconstrained in their money balances.

1. INTRODUCTION

After a major shock such as the 2007–8 financial crisis, central banks around the world

drove the policy rate to zero. Upon doing so, many argued that the central bank was “out

of ammunition” to deal with subsequent, but less severe, shocks to the economy. In short, the

central bank had no ability to conduct stabilization policy once the zero lower bound (ZLB) was

hit. The basic argument of the ineffectiveness of monetary policy at the ZLB goes back to Keynes

and the idea of a liquidity trap. Once the ZLB is hit, the opportunity cost of holding money is

zero. So, agents can be patient in disposing of any excess money balances. Consequently, any

further injections of money will be held as idle balances. As a result, fiscal policy must play

a more important role to stabilize subsequent shocks or the central bank must rely on more

unconventional policies such as quantitative easing (QE) and forward guidance.2

An underlying assumption of the liquidity trap argument is that agents are homogeneous and

they all face the same constraints. Most importantly, agents have the same degree of patience

in terms of their willingness to hold liquid assets. However, if agents vary in their degree of

patience, then the opportunity cost of holding liquid assets is not the same across individuals in

society and impatient agents will make different portfolio decisions compared to patient agents.

This pattern of portfolio differences is borne out by microeconomic evidence—Kaplan et al.

(2014) show that up to one-third of U.S. households hold almost no liquid assets and face high

borrowing costs. In such a world, at the ZLB, some agents may be unconstrained in their liquid

asset holdings whereas others are still constrained. Furthermore, shocks to the economy can

worsen these constraints for some agents in society.

Does this heterogeneity in liquid asset holdings give the central bank an opportunity to engage

in stabilization policy in a beneficial way? The objective of this article is to address this question.

To do so, we construct a New Monetarist model to explore the stabilization response of the

∗Manuscript received March 2018; revised November 2018.

1We thank the editor (Jes´

us Fern´

andez-Villaverde) and two anonymous referees as well as Aleks Berentsen, Daria

Finocchiaro, Pedro Gomis-Porqueras, Per Krusell, and Cyril Monnet for comments that greatly improved the quality

of the article. The views expressed are those of the individual authors and do not necessarily reflect official positions

of Sveriges Riksbank, the Federal Reserve Bank of St. Louis, the Federal Reserve System, or the Board of Governors.

Please address correspondence to: Christopher Waller, Research Division, Federal Reserve Bank of St. Louis, St. Louis,

MO 63166-0442. Phone: +1-314-444-6237. Fax: +1-314-444-8731. E-mail: cwaller@stls.frb.org.

2Christiano et al. (2011) study the impact of fiscal policy at the ZLB, whereas Eggertsson and Woodford (2003)

examine the use of unconventional policies at the ZLB.

1539

C

(2019) by the Economics Department of the University of Pennsylvania and the Osaka University Institute of Social

and Economic Research Association

1540 BOEL AND WALLER

central bank to aggregate shocks at the ZLB. In New Monetarist models, the key friction is

limited commitment, which induces agents to hold liquid assets in order to engage in spot trade.

A key change to the standard model is to introduce idiosyncratic shocks to agents’ discount

factors. Consequently, some agents will be more patient than others in terms of their opportunity

costs of holding liquid assets across time. Thus, more patient agents will hold larger quantities

of liquid assets than impatient agents, which will affect their consumption in some trades.

Furthermore, differences in impatience imply that patient agents are willing to accept a lower

interest rate on assets relative to impatient ones. This has a dramatic impact on monetary

policy. The central bank can only drive the interest rate to zero for the most patient agents in

the economy—all the others are still constrained by their money holdings even at the ZLB. The

central bank would like to drive the nominal interest rate below zero to benefit the impatient

agents, but it is constrained by the patient agents’ portfolio decisions. It is in this sense that the

ZLB is a true constraint on monetary policy.

The severity of the liquidity constraints on impatient agents will vary with aggregate shocks

to the economy. But due to the ZLB, the central bank cannot use traditional interest rate policy

to respond to the shocks. We first examine how a passive monetary policy at the ZLB affects

the real allocation of consumption. Doing so mimics a central bank that is “out of ammunition”

and does not respond to shocks while operating at the ZLB. We show that for small “aggregate

demand” shocks, all agents hold enough liquid assets to consume the first best. For large

demand shocks, instead, impatient agents are constrained by their liquid asset holdings, so they

consume less than the first best. In addition, due to a general equilibrium price effect, patient

agents actually consume more than the first best. These inefficiencies in consumption open the

door for monetary policy to affect consumption and improve welfare.

We show how the central bank can use liquidity injections, in conjunction with forward

guidance, to respond to aggregate shocks. Forward guidance takes the form of committing

to undo the injections at a later date. As a result, the liquidity injections are temporary (i.e.,

repurchase agreements) much like open market operations in the federal funds market. If the

injections were permanent, then the only effect would be a jump in nominal prices and there

would be no real effects on the economy.

We also show that this policy involves no liquidity injection in the lowest aggregate demand

state and then monotonically increasing the size of the injections with the size of the shock. In

this sense, monetary policy is procyclical. These injections effectively redistribute consumption

from low aggregate demand states to high aggregate demand states, thereby improving welfare.

In addition, by exploiting the general equilibrium price effect, the central bank redistributes

consumption across agents—patient agents consume less in high demand states and impatient

agents consume more—in a way that improves welfare.

Stabilization policy works differently in our environment compared to a standard New Key-

nesian (NK) model because the frictions are of a different nature. In an NK model, the relevant

frictions are nominal rigidities. When there is a demand shock (e.g., a positive marginal utility

shock to consumption), then demand for goods increases. With sticky prices, absent monetary

policy action, agents end up consuming more than the efficient quantity since prices do not

rise. Thus, the role of monetary policy is to restrain demand and pull consumption back closer

to the efficient quantity. It does this by increasing the interest rate. With a negative demand

shock the opposite is true—agents decrease consumption too much, so the monetary authority

lowers the interest rate to stimulate demand. Thus, in an NK model, monetary policy tries to

restrain consumption in high demand states and encourage consumption in low demand states.

In this sense, monetary policy is countercyclical.

In our model, the relevant friction is a liquidity or borrowing constraint. When the constraint

binds, consumption is below the efficient quantity. If a positive demand shock occurs, the ef-

ficient quantity increases, but unless the liquidity constraint is relaxed consumption does not

change and it ends up being even further away from the efficient quantity. Thus, optimal stabi-

lization policy attempts to relax the liquidity constraint in order to increase consumption and

bring it closer to the efficient quantity. Therefore, when liquidity or borrowing constraints are