Econ 202a Spring 2000

George Akerlof / Marc Muendler (TA)

Problem Set 3
1 Delegated control over monetary policy
(This question is based on K. Rogoff, QJE 1985; D. Romer, q. 9.12)
Suppose that output is given by the Lucas supply function
y = ȳ + b(π − π e). (1)
The social welfare function is
a
S = γ̃y − π 2. (2)
2
The coefficient γ̃ is a random variable with mean E[γ] = γ̃¯ and variance
Var(γ̃) = σ 2. This reflects the fact that political objectives can vary, de-
pending on the political process and a change of goals. It would be nice
to have a flexible central banker who could quickly adjust to the needs of
the situation, that is a central banker who would pursue a relatively more
expansionary policy whenever the realization of γ̃ is high, and a relatively
strict low-inflation policy whenever γ̃ is low. What is the best choice of a
central banker?
Let the central banker follow the objective function
a
S CB = cγ̃y − π 2, (3)
2
where c ∈ R.
The timing of the model is as follows. The private sector decides π e
without knowledge of γ̃. The central banker, however, chooses π after γ̃ is
known.
a) What is the central banker’s choice of π given π e , γ̃, and c?
b) What expectation of inflation, π e , does a rational private sector choose?
c) What is the expected value of the true social welfare function (2)?
d) What value of c maximizes expected social welfare? Is the ‘optimal’
central banker more or less conservative (inflation averse) than the
private sector? Interpret your result.

1
2 A simple model of overshooting
Suppose money demand takes the Cagan form

mdt − pt = φytd − ηit+1, (4)

where mdt denotes the logarithm of nominal money demand, pt the logarithm
of the price level, ytd the logarithm of aggregate demand and it+1 the nominal
interest rate (≈ ln(1 + it+1)). The coefficients φ and η are assumed to be
positive. Money supply is constant,

mst = m̄. (5)

In efficient international capital markets, uncovered interest parity (UIP)

holds:

it+1 = i∗t+1 + et+1 − et, (6)

where et denotes the logarithm of the nominal exchange rate. In this form, a
high value of et means a weak currency. The logarithm of the real exchange
rate is qt = et + p∗t − pt , so that a high value coincides with a depreciated
real exchange rate.
Suppose that aggregate demand depends on the real exchange rate. It is
high, whenever the real exchange depreciates (because exports benefit):

ytd = δ(et + p∗t − pt ). (7)

For the model to be consistent, we require δ ∈ (0, φ1 ). The full-employment

level of aggregate supply is

yts = ȳ. (8)

Finally, suppose in Keynesian style that prices are not immediately set to
the expected equilibrium level, but adjusted slowly. In particular, let prices
obey the response function

pt+1 − pt = π(ytd − yts ). (9)

In order to further simplify the model, let’s standardize all foreign vari-
ables to constants p∗t = i∗t+1 = 0, and let’s suppose that money markets

2
are always clearing immediately, mdt = mst = m̄. Then a simple Dornbusch
model of overshooting can be built as a system in three equations:
m̄ − pt = φytd − η(et+1 − et ), (10)

1
ytd = δ(et − pt ) δ ∈ (0, ), (11)
φ

pt+1 − pt = π(ytd − ȳ). (12)

a) Find the steady-state values of the exchange rate and the price level
(et+1 = et = ē, pt+1 = pt = p̄).
b) Express both (et+1 − et) and (pt+1 − pt ) as functions of et , pt and
exogenous variables. Find the two functional relationships between pt
and et that satisfy et+1 −et = 0 and pt+1 −pt = 0. Draw them in a phase
diagram with pt on the y-axis and et on the x-axis. Complete the phase
diagram indicating the motion of the system (using the conditions for
et+1 − et ≥ 0 and pt+1 − pt ≥ 0).
Finally, add a line to the diagram that obeys a ‘no-arbitrage condition’
as mandated by UIP: pt − p̄ = −θ̂ (et − ē) for some θ̂ > 0. (The θ̂ is
not quite the same as in the original Dornbusch model since we have
no uncertainty here.)
c) Is the steady-state stable? If not, what is the unique stable (“saddle”)
path given a steady-state of p̄ and ē?
d) Suppose all variables except for pt respond immediately to a monetary
shock.
What is the new steady-state? Draw a new ‘no-arbitrage’ line.
What happens to et right after a reduction in the monetary base from
m̄ to m̄0 ? How do et+s and pt+s evolve over time?

3 Dynamics of the simple model of overshoot-

ing
Consider the simple model of overshooting from question 2 again:
m̄ − pt = φytd − η(et+1 − et ), (13)

3
 1
ytd = δ(et − pt ) δ ∈ (0, ), (14)
φ

pt+1 − pt = π(ytd − ȳ). (15)

We are interested in the behavior of the system around the steady-state.

a) Using (14) and the steady-states of ē and p̄ that you found in question
2a), express ytd − ȳ as a function of et − ē and pt − p̄.

b) Using (13) along with the results in 2a) and 3a), express et+1 − ē as a
function of η1 (et − ē) and η1 (pt − p̄).

c) Using (14) and (15) along with the results in 2a), express pt+1 − p̄ as a
weighted sum of et − ē and pt − p̄.

1
From now on, assume that π = η
and φ = 3 for simplicity.

d) Write your findings from 3b) and 3c) into a system of two difference
equations that takes the form
   
et+1 − ē et − ē
= A· . (16)
pt+1 − p̄ pt − p̄

Find the eigenvalues and eigenvectors of the system (you may, of course,
simply plug your previous results into the formulas from section.)
2(δ+η)
[Hint: An intermediate result is tr(A) = η
and det(A) = 1 −
(1−2η)δ
η2
.]

e) Is the system stable? That is, do the exchange rate and price levels
converge to the steady-state?
If not, set the coefficient of the unstable root to zero. Then, using
the simplified system (16) and the results from 2a), express pt+1 as a
function of et+1 and exogenous variables. Draw this function into the
phase diagram from question 2 (Remember that δ < φ1 .)
What did you just find? What is the intuition for the fact that the
economy obeys this relationship?