Content uploaded by Jonathan Benchimol

Author content

All content in this area was uploaded by Jonathan Benchimol on Aug 27, 2020

Content may be subject to copyright.

Optimal Monetary Policy Under

Bounded Rationality

European Economic Association

35th Annual Congress

Jonathan Benchimol1and Lahcen Bounader2

This pre senta tion do es not n ecess arily re‡ect the v iews of t he Bank of Israe l

or the Int ernat ional M onetar y Fund, or a ny othe r institu tion.

August, 2020

1Bank of Israel

2International Monetary Fund

1 / 48

Empirical Motivation

IEssential concepts for CBs:

IManaging gaps (e.g., in‡ation, output).

Do agents measure/model/understand these gaps accurately ?

IManaging expectations.

Do agents measure/model/understand these expectations

accurately ?

ICollecting prices:

IMore or less easy supermarket/internet.

ICollecting values of the output gap:

ILess easy.

IMyopia(s) to in‡ation and output

)Relative distortion between prices and quantities

)(Optimal) monetary policy?

2 / 48

Theoretical Motivation

IOptimal monetary policy: rational NK models (Clarida et al.,

1999 and Woodford, 2003).

IAgents’expectations are exaggerated in New Keynesian

models (Blanchard, 2009).

IThe economy is inconsistent with any model of rationality

(Stiglitz, 2011).

IWhat is the optimal monetary policy when relaxing the

rational expectations hypothesis?

3 / 48

Literature

IInformation stickiness

Flexible or strict PLT is optimal (Ball, Mankiw, and Reis, 2005).

IRational inattention

Small di¤erences in terms of welfare compared to the rational

case (Mackowiak and Wiederholt, 2015).

ILearning

A form of PLT could be an adequate proxy of the optimal

policy (Eusepi and Preston, 2018).

IBehavioral New Keynesian

PLT is not desirable when …rms are behavioral (Gabaix, 2020).

4 / 48

Intuition

IHow agents’perceptions are key to monetary policy conduct?

IBounded rationality types in‡uence monetary policy reactions

and households’welfare di¤erently.

IIn‡ation expectations are pivotal to most targeting policies.

IDoes the behavioral agent’s welfare is necessarily lower than

the rational agent’s one?

IIgnoring some aspects of reality might be welfare increasing.

IWhy should mechanical rules hold whatever the state of the

world ?

IBehavioral states.

5 / 48

Model

IBuilding on the workhorse NK model and its behavioral

version (Gabaix, 2020).

IAgents are myopic to future disturbances: interest rate,

output-gap, in‡ation, real income, general and full myopia.

IBoundedly rational households and …rms, and rational CB.

IEncompasses the rational model and its policy

recommendations as a particular case.

IContributions:

IConsistent term structure of attention

)Original Phillips curve.

IVariable-speci…c myopias

)Bounded rationality tractability.

IDecreasing return to scale

)Realistic pass-through between real and nominal variables.

IFlexible-price economy

)Microfounded output-gap and natural interest rate.

IWelfare-relevant model

)Consistent optimal monetary policy evaluation.

6 / 48

Findings

IOptimal monetary policy is myopia-dependent.

IIf myopia distorts agents’in‡ation expectations, the optimal

policy entails an IT.

IOtherwise, the optimal policy is a PLT.

IRational in‡ation expectations are a minimal condition for

PLT optimality in a behavioral world.

ITo the extent that bygones are not bygones under PLT.

INo feasible instrument rules for implementing the optimal

monetary policy: casting doubt on the ability of simple Taylor

rules to be useful in guiding monetary policy.

IBounded rationality may be associated with welfare gains.

7 / 48

Environment

IBoundedly rational households maximize their life-time utility

subject to their budget constraint and a non-Ponzi scheme

condition.

¯mGeneral myopia

mrInterest rate myopia

myReal income myopia

IBoundedly rational …rms maximize their perceived pro…t

subject to production technology.

¯mGeneral myopia

mf

πIn‡ation myopia

mf

xOutput-gap myopia

ISticky-price economy: Calvo pricing mechanism.

8 / 48

Households

In…nitely-lived household maximizes

E0

∞

∑

t=0

βtU(ct,Nt)(1)

subject to

kt+1=1+rt+brBR

tktct+yt+byBR

t(2)

St+1=¯mf (St,et+1)(3)

where the following behavioral term structure of expectations

(BTSE) is assumed

EBR

t[b

Xt+k] = mXmkEt[b

Xt+k](4)

mX: level (intercept) of attention (contemporaneous attention).

m: slope of attention (cognitive discounting) as a function of the horizon (k).

9 / 48

IS Curve

IBehavioral IS curve

eyt=MEt[eyt+1]σ(itEt[πt+1]rn

t)(5)

I˜ytis the output gap

IM=MG=m/(RmY¯r)where

IR=1+¯r=1/β.

ImY=(φmy+γ)/(φ+γ).

Iσ=σG=mr/(γR(RmY¯r)).

IRational IS curve (mr=my=¯m=1)

˜yt=Et[˜yt+1]σre (itEt[πt+1]rn

t)(6)

Iσre =1/(γR).

10 / 48

Firms

IContinuum of …rms produces di¤erentiated goods using the

technology

Yt=AtN1α

t(7)

IBehavioral …rm maximizes

∞

∑

k=0

θk

pEBR

tΛt,t+kP

tYt+kjtΨt+kYt+kjt (8)

subject to the sequence of demand constraints.

IΛt,t+k: stochastic discount factor in nominal terms.

IΨt+k(.): cost function.

IYt+kjt: output in t+kfor a …rm that last reset its price in t.

IFOC:

p

tpt1=(1βθ)Θ

∞

∑

k=0

(βθ)kEBR

t[cmct+k]

+

∞

∑

k=0

(βθ)kEBR

t[πt+k](9)

11 / 48

Phillips Curve

IBehavioral Phillips curve

πt=βMfEt[πt+1]+κ˜yt(10)

IMf=θm

1(1θ)mf

π

Iκ=mf

x(1θ)(1βθ)

1(1θ)mf

π

Θγ+φ+α

1αwhere Θ=1α

1α+αe 1.

IRational Phillips curve (mf

x=mf

π=m=1)

πt=βEt[πt+1]+κre ˜yt(11)

Iκre =(1θ)(1βθ )

θΘγ+φ+α

1α.

12 / 48

Phillips Curve: Comparison

IOur behavioral Phillips curve

IMf=θm

1(1θ)mf

π

Iκ=mf

x(1θ)(1βθ)

1(1θ)mf

π

Θγ+φ+α

1α

IGabaix (2020) behavioral Phillips curve

IMf

G=mθ+1βθ

1βθmmf

π(1θ)

IκG=mf

x(1θ)(1βθ)

θ(γ+φ)

ICompared to Gabaix (2020), our fully microfounded Phillips

curve re‡ects a stronger role of mand the importance of

both, decreasing return to scale and in‡ation myopia in κ.

13 / 48

Phillips Curve: Contributions (1)

IGabaix (2020) apply the BTSE to

p

tpt1=(1βθ)

∞

∑

k=0

(βθ)kEBR

tπt+1+... +πt+kµt+k

where µt+k=mct+kis the level of real marginal cost.

IHowever, the BTSE should be applied to the deviation from

the steady state of the variable (Lemma 2.4).

IWe apply the BTSE to

p

tpt1=(1βθ)Θ

∞

∑

k=0

(βθ)kEBR

t[cmct+k]

+

∞

∑

k=0

(βθ)kEBR

t[πt+k](12)

ICorrect transition from subjective to objective expectations

)Our Phillips Curve is not nested in Gabaix (2020).

14 / 48

Phillips Curve: Contributions (2)

Iκ6=κGis related to our assumption of decreasing returns to

scale in the production function.

IGabaix (2020): constant return to scale )κG.

IOur formulation: κis a function of α(∂κ

∂α <0)

)lengthens the feedback from real to nominal variables.

IDecreasing return to scale

IMore realistic (Basu an d Fernal d, 1997 ; Jerma nn and Q uadrin i, 2007 )

IMore realistic role for in‡ation myopia in κ.

Iκis decreasing with αin the general case (α6=0):

IIncomplete feedback from output to in‡ation.

ICentral bank gives less weight to the output gap objective

compared to the constant return to scale case.

IMonetary policy should be more aggressive in bringing down

in‡ation. Intuition con…rmed by the robustness checks (cf.

decreasing vs. constant return to scale calibrations).

15 / 48

Summary

IBehavioral IS curve

eyt=MEt[eyt+1]σ(itEt[πt+1]rn

t)(13)

IBehavioral Phillips curve

πt=βMfEt[πt+1]+κ˜yt(14)

IMand Mfaugment both equations by reducing the excessive

weight given to rational expectations (Blanchard, 2009).

16 / 48

Welfare-Relevant De…nitions

INominal rigidities alongside real imperfections

)Ine¢ cient ‡exible price equilibrium

IOptimal for the central bank to target e¢ cient allocation

)Welfare-relevant variables.

)Model in terms of deviations wrt. e¢ cient aggregates

IWelfare-relevant output: xt=ytye

t

Iye

tis the e¢ cient output

IWelfare-relevant output gap: ˜yt=xt+(ye

tyn

t).

Iyn

tis the natural output (‡exible-price output).

17 / 48

Welfare-Relevant Model

IWelfare-relevant behavioral IS curve

xt=MEtxt+1σ(itEt[πt+1]re

t)(15)

IE¢ cient interest rate perceived by households:

re

t=rn

t+(1/σ)MEtye

t+1yn

t+1(ye

tyn

t)

IWelfare-relevant behavioral Phillips curve

πt=βMfEt[πt+1]+κxt+ut(16)

Iut=κ(ye

tyn

t)is an AR (1)cost-push shock (Galí, 2015)

ut=ρuut1+εu

tand εu

tN(0;σu),i.i.d. over time.

18 / 48

Model Calibration

Parameter Calibration Description

β0.996 Static discount factor

γ2 Household’s relative risk aversion

ε9 Elasticity of substitution between goods

α1/3 Return to scale

φ5 Frisch elasticity of labor supply

θ0.75 Probability of …rms not adjusting prices

ρa0.75 Technology shock persistence

ρu0.75 Cost-push shock persistence

Table 1: Model parameters: Calibration.

Source: Galí (2015).

19 / 48

Myopia Calibration

Models

No myopia Myopia

Rational Interest rate Output gap In‡ation Revenue General Full

mr1 0.85 1 1 1 1 0.85

mf

x1 1 0.85 1 1 1 0.85

mf

π1 1 1 0.85 1 1 0.85

my1 1 1 1 0.85 1 0.85

m1 1 1 1 1 0.85 0.85

Table 2: Myopia parameters: Calibration.

Source: Gabaix (2020).

20 / 48

Optimal Policy

IMicrofounded welfare loss measure derived from the second

order approximation of the behavioral household’s utility

W=1

2E0

∞

∑

t=0

βtπ2

t+wx

wπ

x2

t(17)

Iwπ=e

Θθ

(1βθ)(1θ),

Iwx=γ+φ+α

1α.

21 / 48

Commitment: Analytical Solution

ICentral bank:

ICredible + Able to commit to a policy plan )stabilization.

IChooses a path for the output gap and in‡ation over the

in…nitely lived horizon to minimize the welfare loss.

ICB problem FOCs (Lagrange multiplier: ϕt)

πt+ϕtMfϕt1=0 (18)

wx

wπ

xtκϕt=0 (19)

ISolution:

pt=wx

κwπ xt+1Mft1

∑

j=0

xj!(20)

IA form of PLT is optimal when m=1 and mf

π=1.

IA form of IT is optimal when this condition is not satis…ed.

22 / 48

Commitment: Simulation

510 15 20

0

0.05

0.1

0.15 Inflation

510 15 20

-1.2

-1

-0.8

-0.6

-0.4

-0.2

Output

510 15 20

-0.2

0

0.2

0.4

0.6

Interest rate

510 15 20

0.05

0.1

0.15

0.2 Price level

Rational

Interest rate myopia

Output-gap myopia

Inflation myopia

Revenue myopia

General myopia

Full myopia

Figure 1: Commitment: Impulse response functions.

Note: responses to a 1% cost-push shock.

23 / 48

Commitment: Analysis

ISuboptimality of PLT under in‡ation, full and general

myopia.

IIT is optimal due to the welfare cost induced by CB’s

decisions to stabilize the price level in a world where people

are boundedly rational regarding in‡ation.

IOptimality of PLT under output gap, revenue, and interest

rate myopia.

ICB’s reactions: output gap, in‡ation, and revenue myopia are

very close to the rational case.

IStrong central bank reaction: interest rate,general and full

myopia.

IRemaining cases: optimal required action is smoother, and the

central bank improves the policy trade-o¤ in a way that allows

a de‡ation to operate and then the price level to be stationary.

24 / 48

Commitment: Welfare

No myopia Myopia

Rational Interest rate Output gap In‡ation Revenue General Full

0.174 0.174 0.227 0.190 0.174 0.176 0.248

Table 3: Commitment: Welfare losses.

IIntuitive: rational case generates the lowest welfare loss.

IInterest rate and revenue myopia: same welfare losses as the

rational benchmark.

IThe CB loss does not penalize deviations of interest rate or

revenue: agents are well-informed about output and in‡ation

in these two cases.

IGeneral myopia: close to these cases.

IBounded rationality: not necessarily welfare decreasing.

25 / 48

Discretion: Analytical Solution

ICB “not bound by previous actions or plans and thus is free to

make an independent decision every period” (Plosser, 2007)

IMakes whatever decision is optimal in each period without

committing itself to any future actions.

IMinimizes the welfare loss related to the decision period,

taking into account that expectations are given.

IOptimal discretionary CB should follow this targeting criterion:

πt=wx

κwπ

xt(21)

IAfter a cost-push shock, a discretionary central bank has to

keep this proposition satis…ed to minimize the welfare loss.

IWhen in‡ationary pressures arise, the policymaker has an

incentive to drive output below its e¢ cient level to

accommodate the cost-push shock.

IProposition silent about the in‡uence of bounded rationality.

26 / 48

Discretion: Myopia

ICombining and solving forward:

πt=

wx

wπ

wx

wπ+κ2wx

wπβMfρu

ut(22)

xt=κ

wx

wπ+κ2wx

wπβMfρu

ut(23)

ICB has to let the output gap and in‡ation deviate

proportionally to the cost-push shock (ut).

IBounded rationality in‡uences the magnitudes of these

deviations through κ(mf

x,mf

π) and Mf(m,mf

π).

IOptimal policy response )indeterminate price level but

determinate in‡ation )a form of IT is the preferred regime.

IBounded rationality under discretion in‡uences magnitudes of

the reactions to a shock but does not impact the choice of

the policy regime.

27 / 48

Discretion: Simulation

510 15 20

0.05

0.1

0.15

Inflation

510 15 20

-1.5

-1

-0.5

Output

510 15 20

0.2

0.4

0.6

0.8

1

1.2

Interest rate

510 15 20

0.2

0.4

0.6

Price level

Rational

Interest rate myopia

Output- gap myopia

Inflation myopia

Revenue myopia

General myopia

Full myopia

Figure 2: Discretion: Impulse response functions.

Note: responses to a 1% cost-push shock.

28 / 48

Discretion: Welfare

IIT regime is always the desirable monetary policy.

No myopia Myopia

Rational Interest rate Output gap In‡ation Revenue General Full

0.270 0.270 0.386 0.287 0.270 0.236 0.341

Table 4: Discretion: Welfare losses.

IAlthough this result could seem counterintuitive, general

myopia impacts the level of expectations of all

macroeconomic variables of the model. In this case, people’s

expectations are distorted, which is consistent with a

discretionary policymaker.

29 / 48

Optimal Simple Rules

Name Targeting regime Instrument-rule

F1 Flexible in‡ation it=φππt+φy˜yt

F2 Flexible price level it=φppt+φy˜yt

F3 Flexible NGDP growth it=φg(πt+∆˜yt)+φy˜yt

F4 Flexible NGDP level it=φn(pt+˜yt)+φy˜yt

S1 Strict in‡ation it=φππt

S2 Strict price level it=φppt

S3 Strict NGDP growth it=φg(πt+∆˜yt)

S4 Strict NGDP level it=φn(pt+˜yt)

Table 5: Optimal simple rules: Description

30 / 48

Optimal Simple Rules: Coe¢ cients

F1 F2 F3 F4 S1 S2 S3 S4

φπφyφpφyφgφyφnφyφπφpφgφn

No (rational) 1.96 0.25 0.33 0.0 2.62 0.5 0.17 0.0 2.37 0.34 3.90 0.17

Interest rate 2.44 0.20 0.39 0.0 3.32 0.5 0.20 0.0 3.11 0.40 4.00 0.20

Output gap 1.39 0.32 0.26 0.0 1.81 0.5 0.13 0.0 2.02 0.27 3.43 0.13

In‡ation 1.43 0.27 0.30 0.0 1.55 0.5 0.15 0.0 1.99 0.31 3.26 0.15

Revenue 2.03 0.21 0.33 0.0 2.63 0.5 0.17 0.0 2.37 0.34 3.91 0.17

General 2.05 0.14 0.56 0.0 1.61 0.5 0.25 0.0 2.38 0.58 3.34 0.25

Full 1.54 0.18 0.49 0.0 1.10 0.5 0.21 0.0 2.10 0.50 2.82 0.21

Table 6: Optimal simple rules: Coe¢ cients.

31 / 48

Optimal Simple Rules: Myopia

IThe CB reacts di¤erently under each regime for each

myopia.

IMyopia in‡uences the

Isensitivity of the policy instrument to the CB target.

Itransmission of monetary policy to the output gap.

Itransmission from the output gap to nominal variables.

Itransmission from expectations to in‡ation.

ICB behavior (to control its target).

IUnder general and full myopia, the CB should react

aggressively to:

Icurb expectations.

Iimpact the desired variables.

32 / 48

Optimal Simple Rules: Regimes

ICB more sensitive to its target when operating under strict

targeting compared to ‡exible targeting.

IIn line with Rudebusch (2002) and Benchimol and Fourçans

(2019):

IThe strict NGDP growth targeting coe¢ cients (S3) are higher

than for the ‡exible NGDP growth targeting coe¢ cients (F3)

across all myopia types.

IWhen the central bank targets the NGDP level (F4 and S4) or

the price level (F2 and S2), the coe¢ cients are positive but

lower than one.

IDivine coincidence between stabilizing the price level and the

output gap )a form of PLT leads to self-stabilizing

dynamics for the output gap.

33 / 48

Optimal Simple Rules: Welfare

F1 F2 F3 F4 S1 S2 S3 S4

Regimes

Rational

Interest rate

Output gap

Inflation

Revenue

General

Full

Myopia

0.2093

0.2093

0.2264

0.2093

0.1997

0.1766

0.1766

0.1923

0.1766

0.1773

0.2161

0.2162

0.2361

0.2161

0.2110

0.1855

0.1857

0.2016

0.1855

0.1840

0.2093

0.2094

0.2264

0.2093

0.1997

0.1762

0.1763

0.1919

0.1762

0.1772

0.2167

0.2168

0.2378

0.2167

0.2134

0.1852

0.1854

0.2013

0.1853

0.1838

0.2848

0.2849

0.2317

0.2518

0.2976

0.3091

0.2456

0.2612

0.2848

0.2849

0.2310

0.2517

0.2993

0.3205

0.2450

0.2609

Figure 3: Optimal simple rules: Welfare losses.

Note: the shading scheme is de…ned separately in relation to each column. The

lighter the shading is, the smaller the welfare loss.

34 / 48

Optimal Simple Rules: Welfare

IFlexible targeting rules do not necessarily induce welfare losses

compared to strict rules.

IMost ‡exible targeting rules generate similar welfare losses

compared to their corresponding strict targeting rules.

IStrict PLT delivers the lowest welfare among the considered

rules, similar to the ‡exible PLT rule through di¤erent myopia

cases (divine coincidence when PLT CB).

IRational case delivers similar welfare losses to interest rate and

revenue myopia cases as in commitment and discretion cases.

IThe best monetary policy rule is the strict PLT rule,

whatever types of myopia considered.

IInability of these simple rules to replicate the …rst best

solution (commitment) )optimal policy depends on the

type of myopia characterizing agents.

35 / 48

Robustness: Model Parameters

Calibration name β γ φ e α θ

Galí (2008) 0.99 1 1 6 1/3 0.66

Relative risk aversion 0.99 2 1 6 1/3 0.66

Frisch elasticity 0.99 1 5 6 1/3 0.66

Constant return to scale 0.99 1 1 6 0 0.66

Sticky prices 0.99 1 1 6 1/3 3/4

Time preferences 0.996 1 1 6 1/3 0.66

Demand elasticity 0.99 1 1 9 1/3 0.66

Galí (2015) 0.996 2 5 9 1/3 3/4

Table 7: Calibration of the model parameters used for the robustness

checks.

36 / 48

Robustness: Commitment

Gali (2008)

Relative risk aversion

Frisch elasticity

Constant return to scale

Sticky prices

Time preference

Demand elasticity

Gali (2015)

Rational

Interest rate

Output-gap

Inflation

Revenue

General

Full

Myopia

0.2809

0.2809

0.3171

0.2809

0.2834

0.2235

0.2235

0.2533

0.2235

0.2257

0.1126

0.1126

0.1286

0.1126

0.1136

0.1248

0.1248

0.1424

0.1248

0.1260

0.4667

0.4667

0.5039

0.4667

0.4672

0.2832

0.2832

0.3199

0.2832

0.2861

0.2624

0.2624

0.2966

0.2624

0.2648

0.1741

0.1741

0.1901

0.1741

0.1760

0.3599

0.3962

0.2892

0.3223

0.1492

0.1699

0.1649

0.1873

0.5830

0.6043

0.3627

0.4001

0.3372

0.3727

0.2274

0.2478

Table 8: Commitment: Welfare losses.

37 / 48

Robustness: Discretion

Gali (2008)

Relative risk aversion

Frisch elasticity

Constant return to scale

Sticky prices

Time preference

Demand elasticity

Gali (2015)

Rational

Interest rate

Output-gap

Inflation

Revenue

General

Full

Myopia

0.5102

0.5102

0.5324

0.5102

0.4149

0.5625

0.3740

0.3740

0.4005

0.3740

0.3165

0.1528

0.1528

0.1713

0.1528

0.1403

0.1740

0.1740

0.1942

0.1740

0.1583

1.0109

1.0109

0.9864

1.0109

0.7347

0.8907

0.5212

0.5212

0.5432

0.5212

0.4222

0.5721

0.4649

0.4649

0.4892

0.4649

0.3828

0.2697

0.2697

0.2868

0.2697

0.2362

0.7148 0.5308

0.4484

0.2189

0.2155

0.2494

0.2412

1.3426 0.7315 0.6543

0.5264

0.3862

0.3407

Table 9: Discretion: Welfare losses.

38 / 48

Robustness: Myopia Parameters

Models

No myopia Myopia

Rational Interest rate Output gap In‡ation Revenue General Full Extreme

mr1 0.2 1 1 1 1 0.2 0.01

mf

x1 1 0.2 1 1 1 0.2 0.01

mf

π1 1 1 0.2 1 1 0.2 0.01

my1 1 1 1 0.2 1 0.2 0.01

m1 1 1 1 1 0.2 0.2 0.01

Table 10: Calibration of the myopia parameters used for the robustness

checks.

39 / 48

Robustness: Commitment

510 15 20

0

0.2

0.4

Inflation

510 15 20

-1.5

-1

-0.5

Output

510 15 20

0

1

2

3

Interest rate

510 15 20

0.2

0.4

0.6

0.8

1

Price level

Rational

Interest rate myopia

Output-gap myopia

Inflation myopia

Revenue myopia

General myopia

Full myopia

Extreme myopia

Figure 4: Commitment: Robustness.

Note: Impulse response functions to a 1% cost-push shock.

40 / 48

Robustness: Discretion

510 15 20

0.2

0.4

0.6

0.8

Inflation

510 15 20

-1.4

-1.2

-1

-0.8

-0.6

-0.4

-0.2

Output

510 15 20

1

2

3

Interest rate

510 15 20

1

2

3

Price level

Rational

Interest rate myopia

Output-gap myopia

Inflation myopia

Revenue myopia

General myopia

Full myopia

Extreme myopia

Figure 5: Discretion: Robustness.

Note: Impulse response functions to a 1% cost-push shock.

41 / 48

Robustness: Welfare

Myopia

Interest rate Output gap In‡ation Revenue General Full Extreme

Commitment 0.174 1.446 0.257 0.174 0.143 0.372 0.302

Discretion 0.270 3.357 0.348 0.270 0.145 0.372 0.302

Table 11: Welfare losses: Robustness.

Note: Model calibration: Table 1. Myopia calibration: Table 10.

IWelfare losses under discretion are always higher than under

commitment, except under full and extreme myopia.

IGeneral myopia leads to the best welfare losses under

commitment and discretion )myopia may improve

welfare.

42 / 48

Discussion

IThe rational expectations literature results about optimal

monetary policy are nested in our model (Fuhrer and Moore,

1995; Galí and Gertler, 1999; Walsh, 2017; Woodford, 2003, 2010).

IOptimal policy is myopia-dependent.

IOptimality of a form of PLT (interest rate, output gap or

revenue myopia) and IT (remaining cases) 6=literature

IInformation stickiness (Ball, Mankiw and Reis, 2005),

IRational Inattention (Mackowiak and Wiederholt, 2009, 2015),

ILearning (Eusepi and Preston, 2018),

IBehavioral NK (Gabaix, 2020).

IInability of simple rules to replicate the …rst best solution:

casting doubt on their usefulness in the CB toolkit.

ISvensson (2003): the concept of targeting rules is more

appropriate to the forward-looking nature of monetary policy.

IBounded rationality is not necessarily associated with welfare

losses.

43 / 48

Bridging the gap

IRevision of the IT framework after the GFC (Bernanke, 2017;

Evans, 2018;Blanchard and Summers, 2019).

IPLT overcome the challenges brought by the Zero Lower

Bound (Bernanke, 2017).

ICurrent IT + some adjustments to its parameters: raising the

in‡ation target (Blanchard and Summers, 2019) or allowing

interest rates to be negative.

IBefore the crisis, debate between IT and PLT (Svensson, 1999).

IWe bridge the gap between these competing views:

IBoth forms of targeting (PLT and IT) could be optimal but in

di¤erent circumstances.

IAssessing bounded rationality is a crucial indicator for the CB

targeting policy.

44 / 48

Policy implications (1)

IAgents’expectations matter for monetary policy conduct.

IManaging expectations in a behavioral world

)deviate from a mechanical rule

)more room for adapting policies according to people’s

perceptions.

IPolicymakers’educate the public through intensive

communication

)increase public understanding and trust of their monetary

policies, among other objectives

)attenuate myopia

)may increase welfare.

45 / 48

Policy implications (2)

ICentral banks have to explore,monitor, and analyze agents’

myopia.

IAssessing the degree to which economic agents are myopic is

one of the areas that central banks should invest in more.

IBorrowing an analogy from Thaler (2016), the central bank

Ishould invest in studying the degree to which Homo sapiens

are myopic

Iact consistently rather than educate people and attempt to

transform humans into Homo economicus.

ICall for targeting rules, considering myopia, in the central

banking apparatus in setting monetary policy decisions.

46 / 48

Summary

INo de…nitive answer: neither IT nor PLT is consistently

optimal under all states of the world.

IBounded rationality matters for the conduct of monetary

policy.

IOptimal simple rules: strict PLT in all bounded rationality

cases )puzzling observation about replicating the …rst-best

solution.

IInability of simple rules to replicate the …rst best solution calls

for a reconsideration of their roles in the conduct of monetary

policy.

INew re‡ection about the instrument rules in an economy

with behavioral agents.

IBounded rationality is not necessarily associated with

decreased welfare.

IThe central bank has to identify,assess and monitor

di¤erent myopia types to conduct monetary policy optimally.

47 / 48

Thanks

Thank you for your attention

IComments

IUpdated paper: JonathanBenchimol.com

IEmail: jonathan@benchimol.name

ISocial

ITwitter: @Benchimolium

ILinkedIn: Linkedin.com/in/Benchimol/

48 / 48