Some techniques

Question 5.2
Question 5.5

Tutorial 6: Rational expectations

Stochastic dynamic equilibrium

Ben J. Heijdra

Department of Economics, Econometrics & Finance

University of Groningen

March 15 & 16, 2023

Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 1 / 23

 Some techniques
Question 5.2
Question 5.5

Outline

1 Some techniques

2 Question 5.2

3 Question 5.5

Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 2 / 23

 Some techniques
Question 5.2
Question 5.5

Phase diagram analysis with one difference equation

[How to: Draw a phase diagram for a difference equation]

Suppose that we have one first-order difference equation:

zt+1 = G(zt , x0 )

(1) Draw a diagram with zt on the horizontal axis and zt+1 on the
vertical axis.
(2) Draw the function G(zt , x0 ).
(3) Draw a 45◦ line for which zt+1 = zt .
(4) Mark the steady state, which is such that zt+1 = zt = G(zt , x0 ).
Denote the steady-state value by z ∗ .
(5) For a given initial condition, draw the time path of the system.
Start with z0 and indicate this point along the horizontal axis.
Find z1 = G(z0 , x0 ) and indicate this point along the vertical axis.
Use the 45◦ line to also indicate z1 along the horizontal axis.
Find z2 = G(z1 , x0 ) and indicate this point along the vertical axis.
Use the 45◦ line to also indicate z2 along the horizontal axis.
Etcetera.
Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 3 / 23
 Some techniques
Question 5.2
Question 5.5

Phase diagram analysis with one difference equation

(6) Determine the backward-looking stability of the system.

By looking at the time path.
If the system moves towards the steady state then it is stable. This is
the case in Figures 1(a)-(b).
If the system moves away from the steady state then it is unstable.
This is the case in Figures 1(c)-(d).
By analyzing the sign of the derivative evaluated in the steady state.
If |Gz (zt , x0 )| < 1 for zt = z ∗ then it is stable.
If |Gz (zt , x0 )| > 1 for zt = z ∗ then it is unstable.

Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 4 / 23

 Some techniques
Question 5.2
Question 5.5

Figures 1(a)-(b)

(a) (b)
Stable and monotonic (0 < G∗z < 1) Stable and cyclical (−1 < G∗z < 0)

zt+1 = zt zt+1 = zt
zt+1 zt+1

zt+1 = G(zt, x0)

E0
z* ! z* !
E0

zt+1 = G(zt, x0)

z0 z* zt z0 z* zt

Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 5 / 23

 Some techniques
Question 5.2
Question 5.5

Figures 1(c)-(d)

(c) (d)
Unstable and monotonic (G∗z > 1) Unstable and cyclical (G∗z < −1)

zt+1 = G(zt, x0) zt+1 = zt zt+1 = zt

zt+1 zt+1 zt+1 = G(zt, x0)

z* ! z* !
E0 E0

z0 z* zt z0 z* zt

Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 6 / 23

 Some techniques
Question 5.2
Question 5.5

A variation on the Muth model

The (microeconomic) model of an isolated goods market is described by
the following equations:

QD
t = a 0 − a 1 Pt + V t , a1 > 0, Vt ∼ N (0, σ 2V ) (Q5.1)
QSt = b0 + b1 Pte + Ut , b1 > 0, Ut ∼ N (0, σ 2U ) (Q5.2)
QD
t= QSt
[≡ Qt ] (Q5.3)
Equilibrium price:
a 0 − b0 b1 1
Pt = − Pte + [Vt − Ut ]
a1 a1 a1
With stability we mean that in the absence of shocks (Vt = Ut = 0)
expectations are correct eventually (Pte = Pt ) and the price moves
towards the deterministic equilibrium price:
a 0 − b0
P̄ ≡
a 1 + b1

Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 7 / 23

 Some techniques
Question 5.2
Question 5.5

Part (b)

Consider the Adaptive Expectations Hypothesis (AEH):

Pte e
= Pt−1 e
+ λ[Pt−1 − Pt−1 ]
∆Pte ≡ Pte − e
Pt−1 e
= λ[Pt−1 − Pt−1 ]
Pte e
= Pt−1 + (1 − λ)[Pt−1 − Pt−1 ]
Pte e
= λPt−1 + (1 − λ)Pt−1

where λ > 0.
(b) Derive the stability condition for the model under AEH.
The model consists of two equations in Pt and Pte :

a 0 − b0 b1 1
Pt = − Pte + [Vt − Ut ]
a1 a1 a1
Pte = (1 − λ)Pt−1
e
+ λPt−1

Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 8 / 23

 Some techniques
Question 5.2
Question 5.5

Part (b)

(b) We can reduce the model to a difference equation for Pte :

a 0 − b0 λ
Pte = µPt−1
e
+λ + [Vt−1 − Ut−1 ]
a1 a1

with steady state value P̄ .

Stability condition:

a 1 + b1
|µ| = 1 − λ <1
a1

Monotonic adjustment: 0 < µ < 1 or 0 < λ < a1 /[a1 + b1 ]

Cyclical adjustment: −1 < µ < 0 or
a1 /[a1 + b1 ] < λ < 2a1 /[a1 + b1 ]
No adjustment: µ = 0 or λ = a1 /[a1 + b1 ]

Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 9 / 23

 Some techniques
Question 5.2
Question 5.5

Monotonic adjustment

Phase diagram Time path

Pte Pte
45◦

P̄
P̄

e
P̄ Pt−1 t

Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 10 / 23

 Some techniques
Question 5.2
Question 5.5

Cyclical adjustment

Phase diagram Time path

Pte Pte
45◦

P̄
P̄

e
P̄ Pt−1 t

Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 11 / 23

 Some techniques
Question 5.2
Question 5.5

Part (c)

Consider the Rational Expectations Hypothesis (REH):

Pte = E(Pt |Ωt−1 ) = Et−1 (Pt )

where Ωt−1 is the information set available at time t − 1:

Ωt−1 = Pt−1 , . . . ; Qt−1 , . . . ; a0 , b0 , a1 , b1 ; Ut ∼ N (0, σ 2U ), Vt ∼ N (0, σ 2V )



(c) Derive the equilibrium price level.

The model consists of two equations in Pt and Pte :

a 0 − b0 b1 1
Pt = − Pte + [Vt − Ut ]
a1 a1 a1
Pte = Et−1 (Pt )

Solve for Pte

Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 12 / 23

 Some techniques
Question 5.2
Question 5.5

Part (c)

The rational expectation of the price level is:

a 0 − b0
Pte = = P̄
a 1 + b1
The resulting equilibrium price is:
1
Pt = P̄ + [Vt − Ut ]
a1
The resulting equilibrium quantity traded is:

Qt = Q̄ + Ut

where Q̄ is the deterministic equilibrium quantity:

Q̄ ≡ a0 − a1 P̄ = b0 + b1 P̄

Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 13 / 23

 Some techniques
Question 5.2
Question 5.5

Demand and supply shocks under rational expectations

Pt

A
!
!D

E0
P !

C
B !
!

Q Qt

Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 14 / 23

 Some techniques
Question 5.2
Question 5.5

Expectational difference equation

Suppose that the stochastic process for Yt is given by:

Yt = α0 + α1 Et Yt+1 + Ut , 0 < α1 < 1, (Q5.9)

where α0 and α1 are constants, Et is the conditional expectation (based

on the period-t information set), and Ut is a stochastic shock term. We
assume that this shock term features first-order autocorrelation:

Ut = θUt−1 + Vt , 0 < θ < 1, (Q5.10)

where θ is a constant and Vt is a white noise error term (with E(Vt ) = 0

and E(Vt2 ) = σ 2 ).

Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 15 / 23

 Some techniques
Question 5.2
Question 5.5

Part (a)
Can you think of an economic example for which an expression like
(Q5.9) arises naturally?
Note that (Q5.9) is an expectational difference equation
Example: Arbitrage condition between a safe asset (paying a
constant interest rate, R) and shares (paying dividends, dt ) is
written as:
dt + Et pt+1 − pt
R= (A1)
pt

pt is the price of the share at the beginning of period t

LHS: net yield on the bank deposit
RHS: yield on shares
Rewrite (A1) to get:
1 h i
pt = dt + Et pt+1 (A2)
1+R
If the process for dt itself contains a stochastic term, (A2) has the
same form as (Q5.9) in the question
Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 16 / 23
 Some techniques
Question 5.2
Question 5.5

Part (b)
Compute the rational expectations solution for Yt . Hint: use the
method of undetermined coefficients by trying a candidate solution
of the form Yt = π 0 + π 1 Ut and computing those values for π 0 and
π 1 for which the candidate solution is the correct solution.
Trial solution:
Yt = π 0 + π 1 Ut (A3)
This implies:
Yt+1 = π 0 + π 1 Ut+1 (A4)
Take conditional expectations:

Et Yt+1 = π 0 + π 1 Et Ut+1 = π 0 + π 1 θUt (A5)

Substitute (A5) into (Q5.9):

Yt = α0 + α1 [π 0 + θπ 1 Ut ] + Ut
= (α0 + α1 π 0 ) + (1 + α1 θπ 1 ) Ut (A6)

Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 17 / 23

 Some techniques
Question 5.2
Question 5.5

Part (b)

To summarize:

Yt = π 0 + π 1 Ut (A3)
Yt = (α0 + α1 π 0 ) + (1 + α1 θπ 1 ) Ut (A6)

Trial solution and implied-by-trial solution have the same form

Make sure they are identical by choosing π 0 and π 1 :
α0
π 0 = α0 + α1 π 0 ⇔ π0 = (A7a)
1 − α1
1
π 1 = 1 + α1 θπ 1 ⇔ π1 = (A7b)
1 − α1 θ
Substitute π 0 and π 0 into (A3) to get the rational expectations
solution for Yt :
α0 1
Yt = + Ut (A8)
1 − α1 1 − α1 θ

Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 18 / 23

 Some techniques
Question 5.2
Question 5.5

Part (c)
[Not covered; not examinable] Compute the rational expectations
solution for Yt by the method of repeated substitution (forward
iteration). Verify that you get the same solution as in part (b).
Write (Q5.9) for period t + 1:
Yt+1 = α0 + α1 Yt+2 + Ut+1
Take expected value:
Et Yt+1 = α0 + α1 Et Et+1 Yt+2 + Et Ut+1
= α0 + α1 Et Yt+2 + θUt (S1)
where we have used the law of iterated expectations
(Et Et+1 Yt+2 = Et Yt+2 ) and knowledge of the shock process
(Et Ut+1 = θUt ) in going from the first to the second line.
Substitute (S1) into (Q5.9):
Yt = α0 + α1 [α0 + α1 Et Yt+2 + θUt ] + Ut
= α0 (1 + α1 ) + α21 Et Yt+2 + (1 + α1 θ)Ut (S2)
Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 19 / 23
 Some techniques
Question 5.2
Question 5.5

Part (c)

Compute the rational expectations solution for Yt by the method of

repeated substitution (forward iteration). Verify that you get the
same solution as in part (b).
If we keep on substituting like this we get:

Yt = α0 1 + α1 + α21 + · · · + α1N −1 + αN
 
1 Et Yt+N
h i
2
+ 1 + α1 θ + (α1 θ) + · · · + (α1 θ)N −1 Ut (S3)

By letting N → ∞ the terms in square brackets converge to,

respectively, 1/(1 − α1 ) (as 0 < α1 < 1) and 1/(1 − α1 θ) (as
0 < α1 θ < 1) and αN 1 Et Yt+N → 0. As a result, equation (S3)
converges to:
α0 1
Yt = + Ut (S4)
1 − α1 1 − α1 θ

Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 20 / 23

 Some techniques
Question 5.2
Question 5.5

Part (d)
[Not covered; not examinable] Compute the asymptotic variance
of Yt . Show that it depends positively on the autocorrelation
parameter θ. Hint: first compute Yt − θYt−1 and then write it as a
difference equation.
The asymptotic variance of Yt is defined as:
2
σ 2Y ≡ Et−∞ [Yt − Et−∞ Yt ]
Use the hint and write:
α0 1
Yt =
+ Ut (A9)
1 − α1 1 − α1 θ
α0 θ 1
θYt−1 = + θUt−1 (A10)
1 − α1 1 − α1 θ
Deduct (A10) from (A9):
α0 (1 − θ) 1
Yt − θYt−1 = + [Ut − θUt−1 ] ⇒
1 − α1 1 − α1 θ
α0 (1 − θ) 1
Yt = + θYt−1 + Vt (A11)
1 − α1 1 − α1 θ
Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 21 / 23
 Some techniques
Question 5.2
Question 5.5

Part (d)
Compute:
α0 (1 − θ)
Et−∞ Yt = + θEt−∞ Yt−1 (A12)
1 − α1
Deduct (A12) from (A11):
1
Yt − Et−∞ Yt = θ [Yt−1 − Et−∞ Yt−1 ] + Vt (A13)
1 − α1 θ
Square (A13) and take expectations:
2
σ 2Y ≡ Et−∞ [Yt − Et−∞ Yt ]
h i
2
= Et−∞ θ2 (Yt−1 − Et−∞ Yt−1 )
" 2 #
1 2 2θ(Yt−1 − Et−∞ Yt−1 )Vt
+ Et−∞ Vt +
1 − α1 θ 1 − α1 θ
 2
1
= θ2 σ 2Y + σ 2V (A14)
1 − α1 θ
Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 22 / 23
 Some techniques
Question 5.2
Question 5.5

Part (d)

We have used that for stationary processes:

2 2
Et−∞ [Yt − Et−∞ Yt ] = Et−∞ [Yt−1 − Et−∞ Yt−1 ]

We have used independence of Vt and Yt−1 :

Et−∞ [(Yt−1 − Et−∞ Yt−1 )Vt ] = 0

Solve (A14) for σ 2Y :
 2
1 1
σ 2Y = σ 2V (A15)
1 − θ2 1 − α1 θ

Intermediate Macroeconomics EBB842B05, 2022–2023 Tutorial 6: Rational expectations 23 / 23