Lambda Group

The Science of Monetary Policy: A New Keynesian

Perspective
(Richard Clarida, Jordi Galí, and Mark Gertler-1999)

F. Martín Martínez Palomino

(Step-by-step Solution)
June 24, 2023

1 The Central Bank:

Objective function:
∞
( )
1 X
k 2
max V = − Et β (πt+k + αx2t+k ) (1)
2
k=0

s.t.
Curva de Phillips
πt = βEt πt+1 + λxt + ut (P hillips Curve) (2)
where:

xt = −φ(it − Et πt+1 ) + Et xt+1 + gt (Dynamic IS equation) (3)

1.1 The Policy Problem and Discretion versus Rules

The policy problem is to choose a time path for the instrument it to engi- neer time paths of the target variables
xt and πt that optimize the objective function.

1.1.1 Discretion:
Opt.
∞
( )
1 X
k 2
max V = − Et β (πt+k + αx2t+k ) (4)
2
k=0
s.a.
Curva de Phillips
πt = βEt πt+1 + λxt + ut (5)

1. Solution:
∞
( )
1 X
k
h
2
i
V = − Et β (βEt+k πt+k+1 + λxt+k + ut+k ) + αx2t+k
2
k=0

∂V
= 0 =⇒ 2(βEt+k πt+k+1 + λxt+k + ut+k )λ + 2αxt = 0
∂xt+k

1
2λπt+k + 2αxt+k = 0

λ
xt = − πt (6)
α
plugging eq.(6) into the Phillips curve eq.(5):

−λ
πt = βEt πt+1 + λ( πt ) + ut
α

λ2
πt (1 + ) = βEt πt+1 + ut
α

we get:

βEt πt+1 ut λ2
πt = + ; α0 = 1 + (7)
α0 α0 α
Solution through undetermined coefficients:
Conjecture:

πt = η1 ut (8)

iterating eq.(8) one period forward

πt+1 = η1 ut+1 ut+1 = ρut + ût+1

replacing it into the conjecture

βη1 ρut ut
πt = +
α0 α0

βη1 ρ + 1
πt = ( )ut = η1 ut (9)
α0
so, we have our tentative solution must be equal to our conjecture (eq.(8)=eq.(9))

βη1 ρ + 1
= η1
α0

so:

1
η1 =
α0 − βρ

now, as we know the solution of η we can replace its value into our initial conjecture

1 1
πt = ut =⇒ πt = λ 2 ut
α0 − βρ 1 + α − βρ

2
Solution of Inflation:

α
πt = ut (10)
λ2 + α(1 − βρ)

using eq.(6):

λ
xt = − πt
α

we can find the solution of the output gap

λ α
xt = − ut
α λ2 + α(1 − βρ)

Solution of the output gap:

−λ
xt = ut (11)
λ2 + α(1 − βρ)

2. The Interest Rate

To calculate the intersest, rearranging the dynamic IS (eq. (3)),

xt = −φ(it − Et πt+1 ) + Et xt+1 + gt

xt − Et xt+1 − gt = −φ(it − Et πt+1 )

1
Et πt+1 − (xt − Et xt+1 − gt ) = it (12)
φ
iteranting eq.(10) a period forward and taking expectations
α
πt = ut
λ2 + α(1 − βρ)
αρ
Et πt+1 = ut (13)
λ2 + α(1 − βρ)
plugging eq.(13) into eq.(12):
αρ 1
ut − (xt − Et xt+1 − gt ) = it
λ2 + α(1 − βρ) φ
αρ 1 1
ut + (Et xt+1 − xt ) + gt = it
λ2 + α(1 − βρ) φ φ
−λ(ρ−1)
taking into account that Et xt+1 − xt = λ2 +α(1−βρ) ut

αρ 1 −λ(ρ − 1) 1
2
ut + 2
ut + g = it
λ + α(1 − βρ) φ λ + α(1 − βρ) φ

αρ 1 λ(1 − ρ) 1
ut + ut + gt = it
λ2 + α(1 − βρ) φ λ2 + α(1 − βρ) φ
Let’s define for tractability:
1
q=
λ2 + α(1 − βρ)

3
so
1 1
αρqut + λ(1 − ρ)q ut + gt = it
φ φ

recall eq.(13): Et πt+1 = αρqut

λ(1 − ρ) 1
Et πt+1 + Et πt+1 + gt = it
φαρ φ
The optimal feedback policy for the interest rate is

λ(1 − ρ)
it = γπ Et πt+1 + σgt where γπ = 1 + >1 (14)
ραφ

1.2 Commitment
Let’s consider a rule for the target variable xt that is contingent on the fundamental shock ut

xct = −wut (15)

given the Phillips curve:

πtc = βEt πt+1

c
+ λxct + ut

1. Solution
replacing eq.(15) into the Phillips curve:

πtc = βEt πt+1

c
+ λ(−wut ) + ut

πtc = βEt πt+1

c
+ ut (1 − λw)

Solution through undetermined coefficients:

Conjecture:

πtc = η2 ut

iterating it a period forward

c
πt+1 = η2 ut+1 ut+1 = ρut + ût+1

replacing the conjecture

πtc = βη2 ρut + ut (1 − λw) = η2 ut

ut (βη2 ρ + 1 − λw) = η1 ut

1 − λw = η2 (1 − βρ)

1 − λw
η2 =
1 − βρ

so, we have:

4

1 − λw
πtc = ut (16)
1 − βρ

Now, the challenge is to determine what the optimum value of w is :

∞
( )
c 1 X
k 2
max V = − Et β (πt+k + αx2t+k )
2
k=0

∞
( 2 ) ( )
c 1 1 − λw X
k
V = − σu2 + αw 2
Et β
2 1 − βρ
k=0

∞
( X
)
∂V 1 2 1 − λw −λ k
= − σu 2 + 2αw Et β =0
∂w 2 1 − βρ 1 − βρ
k=0

1 − λw λ
= αw
1 − βρ 1 − βρ

λ − λ2 w = αw(1 − βρ)2

λ = w α(1 − βρ)2 + λ2

λ
w∗ = (17)
α(1 − βρ)2 + λ2

replacing eq.(17) into eq.(16):

1 − λw
πtc = ut
1 − βρ
Solution of the Inflation:

α(1 − βρ) αc
πtc = ut = ut , αc ≡ α(1 − βρ) < α
λ2 + α(1 − βρ)2 λ2 + αc (1 − βρ)

to get the solution of the output gap, we need to replace the value of w∗ into eq.(15):

xct = −wut

Solution of the output gap

−λ
xct = ut (18)
λ2 + αc (1 − βρ)

2. The Interest Rate

Following same steps as in the last case, the optimal feedback policy for the interest rate is

λ(1 − ρ)
it = γπc Et πt+1 + σgt where γπc = 1 + >1 (19)
ραc φ

5
1.3 Comparison

Solution under commitment Solution under discretion

2 2
α(1−βρ) α(1−βρ) α α
πtc = c
λ2 +α(1−βρ)2 ut ⇒var(πt ) = λ2 +α(1−βρ)2 σu2 πt = λ2 +α(1−βρ) ut ⇒var(πt ) = λ2 +α(1−βρ) σu2

2 2
−λ λ −λ λ
xct = λ2 +α(1−βρ)2 ut ⇒var(xct ) = λ2 +α(1−βρ)2 σu2 xt = λ2 +α(1−βρ) ut ⇒var(xt ) = λ2 +α(1−βρ) σu2

var(πtc ) var(π c )
so: var(xct ) = (1 − βρ)2 var(π t)
var(xt ) and taking into account that (1 − βρ) < 1; thus: var(xct ) <
t var(πt )
var(xt )
c 1 1
it = γπ Et πt+1 + φ gt it = γπ Et πt+1 + φ gt
c λ(1−ρ) λ(1−ρ)σ
γπ = 1 + ραc φ > γπ = 1 + ραφ
So, we can conclude that the interest rate is higher under commitment.