Solutions - Chapter 2

Macroeconomics II

Professor: Caio Machado ([email protected])

Teaching assistants: Wei Xiong ([email protected]) and Diego Cussen ([email protected])

Last updated: April 9, 2019

Exercise 2.1

Item 1

The agent chooses the number of withdrawals n ∈ R+ to minimize

 
Y
C (n) = i + Zn
2n

Note that for simplicity we are assuming that n is chosen from R+ instead of the being
chosen from the set of non-negative integers {0, 1, 2, ...} (as if the agent could do one
and a half withdrawals, for example).
Y
Notice that the term 2n represents the agent average money holdings.1 Hence, when
1 To see that, notice that an agent that goes n times to the bank will withdraw Y/n each time (so
that the total amount withdrawed after n withdrawals is equal to Y). When the exercise says that the
individual spends his income uniformly, we mean that dM t
dt is constant between each withdrawals (where
Mt denotes the agent money holdings at some date t, measured in years). Hence, in interval of time
[0, 1/n] (i.e., the interval between the first and the second withdrawal) the agent money holdings M̃t are
a linear function of time that: (i) is equal to Y/n when t = Y/n (since he withdrawals Y/n at t = 0); (ii)
approaches zero when t = 1/n (the moment of the second withdrawal), since she is about to make a new
withdrawal when t = 1/n, and she only withdrawals when there is no more money holdings. Therefore,
M̃t = Yn − Y × t, for t ∈ [0, 1/n]. Hence, the average money holdings in the interval t ∈ [0, 1/n] is:
R 1/n  Y 
− Y × t dt 1/n #
"
t2 
 
D 0 n Y Y 1 Y
M = =n t−Y  = n 2 −Y 2 =
1/n n 2 0 n 2n 2n
Since the average money holdings between any two consecutive withdrawals is the same, the average

Macroeconomía II, 2019/1 1

the agent increases the number of withdrawals it reduces the opportunity cost of hold-
 
Y
ing money, represented by i 2n , since she holds less cash on average, but needs to
pay the cost of withdrawing (represented by Zn).

Item 2

Taking the first order condition we find the n that minimizes the cost:

iY
=Z
2n2

And hence, the optimal n, denoted by n∗ , is:

r
iY
n∗ =
2Z

Therefore, the average money holdings (that is, the money demand) is given by

r r r
Y Y 2Z Y 2 2Z YZ
MD = q = = = (1)
iY
2 2Z 2 iY 4 iY 2 i

Taking logs:
1
ln M D = (ln Y + ln Z − ln 2 − ln i ) (2)
2

Therefore, we have two main conclusions: (i) the elasticity of the money demand with
respect to income is equal to 1/2; (ii) the elasticity of the money demand with respect
to the nominal interest rate is equal to −1/2.

Item 3

That would probably reduce the cost of going to the bank (Z). As equation (1) shows,
that should reduce the money demand.
Y
money is 2n .

Macroeconomía II, 2019/1 2

Item 4

The agent now has two decisions to make. She must choose how much of his total
yearly transactions (Y) she is going to conduct using currency (denoted by X C ) and
how much she is going to conduct using electronic money (denoted by X E ). (Note that
X E and X C must be such that X E + X C = Y.) Once she has fixed an amount X C , then
she chooses how many times to go to the bank to achieve that amount of transactions
using currency.
First, let’s suppose an agent has chosen a given level of X C . From our solution
q
∗ iX C
of the previous items, we know that the agent will go n = 2Z times to bank and
q
C
her average money holding is X2 Zi . Therefore, the total cost of spending X C using
currency is:

r r
XC Z iX C
CC ( X C ) = i +Z
2 i}
| {z 2Z}
| {z
average money holdings # of time she goes to the bank

The total cost incurred for a given X C and X E (after optimally choosing how much to
go to the bank to achieve a given X C ) is given by:

r r
XC Z iX C
C(XC , X E ) = i +Z + τX E
| 2 i {z 2Z} |{z}
cost of electronic transactions
CC ( X C )

Which simplifies to:

√
C(XC , X E ) = 2iZX C + τX E

The agent chooses X C and X E to minimize C ( X C , X E ) above subject to X E + X C = Y,

X E ≥ 0 and X C ≥ 0. Hence, replacing X E = Y − X C in C ( X C , X E ) we get:

  √  
C̃ ( X C ) ≡ C X C , Y − X C = 2iZX C + τ Y − X C (3)

Macroeconomía II, 2019/1 3

Therefore, we can simplify the problem: the agent chooses X C ∈ [0, Y ] to minimize
(3) (and then X E is given by Y − X C ). Notice that the function above is concave. This
implies that we can only have corner solutions: either the agent chooses X C = 0 or
X C = Y. Therefore, the agent chooses X C = Y whenever:

√
τY >
|{z} | 2iZY
{z }
C̃ (0) C̃ (Y )

Which yields:

r
2iZ
τ>
Y
q
2iZ
Hence, when τ > the agent chooses to make all transactions with currency (i.e.,
Y
q
she chooses X = Y). When τ < 2iZ
C she chooses to only use electronic money (i.e.,
qY
2iZ
she chooses X C = 0). When τ = C
Y the agent is indifferent between X = 0 or

X C = Y.

Item 5

Now the agent solves the same problem as in item 4, with the additional constraint
that X C ≥ λY (since at fraction λ of the income must be spent with currency). Hence,
instead of choosing X C ∈ [0, Y ] to minimize (3), the agent now chooses X C ∈ [λY, Y ] to
minimize (3). As before, we can only have corner solutions to this problem, since the
objective function is concave. Hence, the agent chooses X C = λY if:

√ √
2iZλY + τ (Y − λY ) < | 2iZY
{z }
| {z }
C̃ (λY ) C̃ (Y )

Which becomes:  √ r
1− λ 2iZ
τ<
(1 − λ ) Y

Macroeconomía II, 2019/1 4

 √
(1− λ)
q
2iZ
Hence, when τ < (1− λ )
the agent chooses to make as little transactions as
Y √ q
C (1− λ) 2iZ
possible with currency and chooses X = λY. Similarly, when τ > (1−λ) Y , the

agents chooses to make all her transactions with currency and then chooses X C = Y.
√ q
(1− λ) 2iZ
When τ = (1−λ) the agent is indifferent between X C = λY and X C = Y. When
Y √
(1− λ) 2iZ
q q
2iZ
λ → 0 we have that (1−λ) Y → Y , which is the same cutoff on τ we found in

item 4. Moreover, limλ→0 λY = 0. Hence, the money demand converges to that of item
4 (which is not surprising, of course).

Exercise 2.2

Item 1

False. Using equation (2) from our answer of Exercise 4 one can see that the elasticity
of the money demand with respect to the interest rate is constant and equal to −1/2.
Therefore, when i increases 10%, the money demand falls 5%.

Item 2

False. Using (2) one can see that the elasticity of the money demand with respect to
income is constant and equal to 1/2.

Item 3

False. Dividing (1) by Y we get:

r r r
MD 1 YZ 1 YZ Z
= = =
Y Y 2 i Y2 2 i 2Yi

Hence, the share of the income held as money is decreasing in income.

Macroeconomía II, 2019/1 5

Exercise 2.3

Item 1

Using the Fisher equation, (1) becomes:

it
um (ct , mt ) = uc (ct , mt )
(1 + r t ) (1 + π t +1 )

Using (2) we get:

it it
um (ct , mt ) = (1+r
β t ) u c ( c t +1 , m t +1 )

= βuc (ct+1 , mt+1 )
(
 1+r
t) (

1 + π t +1 ) 1 + π t +1

The intuition is the following. In equilibrium, the household must be indifferent be-
tween having more money today or consuming tomorrow. If she gives up holding
one unit of money at t, she gets it extra dollars at t + 1, which implies a real increase of
it
1 + π t +1 in the amount of goods she can buy. Thus, her utility increases by 1+πit t+1 βu0 (ct+1 , mt+1 ).
For her to be indifferent, this must be equal to the marginal utility of holding money
today.

Item 2

Since consumption is independent of the monetary side of the economy, we maximize

the household utility by maximizing his utility of holding money. Since u(·) is con-
cave in m, this happens when um (c, m) = 0. Thus, (1) implies that interest rate that
maximizes the household utility in the steady state is:

iss
0= ⇒ iss = 0
1 + iss

Macroeconomía II, 2019/1 6

The Fisher equation implies that

1 + iss = (1 + r ss ) (1 + π ss )

1
And thus, the steady state welfare maximizing inflation is 1 + π ss = 1+r ss ⇒ π ss =
ss
− 1+r rss . The money demand when iss = 0 is given by um (css , mss ) = 0, implying mss =
5.
The intuition for it = 0 being optimal is the following. The opportunity cost of
holding money is the nominal interest rate (it is the “price” of money). The marginal
cost of producing money is zero. Efficiency implies that price equals the marginal cost,
thus it = 0.

Exercise 2.4

Item 1

Given that now the money accumulated at date t only yields utility at date t + 1, it is
useful to introduce a small change in notation. We denote the money the agent decides
to carry from date t to date t + 1 by Mt+1 (and not by Mt as we did in class). The re-
maining notation is the same used in class: Bt denotes the amount invested in bonds at
date t; k t is the amount of capital carried from t to t + 1; ct is the amount of consumption
at date t and Tt are the cash transfers received from the central bank at the beginning of
date t; it is the nominal interest rate paid at t + 1 for each dollar invested in bonds at t.
The household budget constraint in nominal terms can then be written as:

ct Pt + k t Pt + Mt+1 + Bt = f (k t−1 ) Pt + k t−1 Pt (1 − δ) + (1 + it−1 ) Bt−1 + Mt + Tt

Macroeconomía II, 2019/1 7

Item 2

First, we define the real variables as we did in class: bt ≡ Bt /Pt , mt ≡ Mt /Pt and
τt ≡ Tt /Pt . Moreover, rt is the real interest rate.
To write the budget constraint in real terms we divide it by Pt as usual:

Mt+1 Bt B Mt Tt
ct + k t + + = f ( k t −1 ) + k t −1 (1 − δ ) + (1 + i t −1 ) t −1 + +
Pt Pt Pt Pt Pt

But since:
Mt + 1 M P
= t +1 t +1 = m t +1 (1 + π t +1 )
Pt Pt+1 Pt

and
Bt−1 B P b
= t −1 t −1 = t −1
Pt Pt−1 Pt 1 + πt

We can write the budget constraint as:

bt − 1
c t + k t + m t + 1 ( 1 + π t + 1 ) + bt = f ( k t − 1 ) + k t − 1 ( 1 − δ ) + ( 1 + i t − 1 ) + mt + τt
1 + πt

But by the Fisher equation 1 + rt−1 = (1 + it−1 ) / (1 + πt ). Therefore:

ct + k t + mt+1 (1 + πt+1 ) + bt = f (k t−1 ) + k t−1 (1 − δ) + (1 + rt−1 ) bt−1 + mt + τt

Item 3

First, we need to write the agents problem. Note that given our change in notation
for Mt , the agent utility can be written exactly as before: ∑∞
t=0 u ( ct , mt ) (if we had not

changed the notation, the budget constraint would look the same, but mt−1 would show

Macroeconomía II, 2019/1 8

up in the instantaneous utility at t).

∞
max
{ct ,k t ,mt+1 ,bt }t≥0
∑ βt u (ct , mt )
t =0

s.t. ct + k t + mt+1 (1 + πt+1 ) + bt = f (k t−1 ) + k t−1 (1 − δ) + (1 + rt−1 ) bt−1 + mt + τt , ∀t ≥ 0

ct , k t , mt+1 , bt ≥ 0, ∀t ≥ 0

m0 = m > 0, b−1 = b > 0, k −1 = k > 0

Here we will use the fact that the solution is interior, and hence we will ignore the
non-negativity constraints. The Lagrangian of this problem is:

∞
L= ∑ βt [u (ct , mt ) − λt (ct + kt + mt+1 (1 + πt+1 ) + bt − f (kt−1 ) − kt−1 (1 − δ) − (1 + rt−1 ) bt−1 − mt − τt )]
t =0

We will take the FOCs with respect to ct , mt+1 , bt :

βt uc (ct , mt ) − βt λt = 0 (FOC1)

− β t λ t (1 + π t +1 ) + β t +1 [ u m ( c t +1 , m t +1 ) + λ t +1 ] = 0 (FOC2)

− β t λ t + β t +1 λ t +1 (1 + r t ) = 0 (FOC3)

Which simplifying becomes:

uc (ct , mt ) = λt (FOC1’)

λ t (1 + π t +1 )
u m ( c t +1 , m t +1 ) = − λ t +1 (FOC2’)
β
λt
1 + rt = (FOC3’)
βλt+1

Macroeconomía II, 2019/1 9

Dividing (FOC2’) by (FOC1’) forward:

u m ( c t +1 , m t +1 ) λ t (1 + π t +1 )
= −1
u c ( c t +1 , m t +1 ) βλt+1

Using (FOC3’) it becomes:

u m ( c t +1 , m t +1 )
= (1 + r t ) (1 + π t +1 ) − 1
u c ( c t +1 , m t +1 )

Finally, by the Fisher equation (1 + rt ) (1 + πt+1 ) = 1 + it . Therefore, it becomes:

u m ( c t +1 , m t +1 )
= it (*)
u c ( c t +1 , m t +1 )

Remember from your micro class, that if an is choosing between two goods, say z and y,
then the marginal rate of substitution between z and y must be equal to the price of z in
units of y (that is, using the usual notation, Umgz /Umgy = Pz /Py ). Here, the LHS of (*)
is the marginal rate of substitution between mt+1 and ct+1 . To understand why we can
interpret it as the relative price of mt+1 in terms of ct+1 , consider an agent who at date
t decides to consume one extra unit of the consumption good at t + 1. To do that, she
must increase her wealth at the beginning of t + 1 in Pt+1 units. She can achieve that by
purchasing Pt+1 / (1 + it ) dollars in bonds at date t and then using all the bond payment
at t + 1 to buy the consumption good (Pt+1 / (1 + it ) is the “price” of consumption at
t + 1 in date t dollars). Now suppose the agent decides to enter date t + 1 with one
extra unit of real balances. To achieve that, she must hold Pt+1 units of cash at date t.
But since that cash remains with her at date t + 1, she can reduce her bond holdings at
t in Pt+1 / (1 + it ) dollars and still get the same wealth at the beginning of t + 1. Hence
Pt+1 − Pt+1 / (1 + it ) is the “price” of holding one unit of real money balances at t + 1

Macroeconomía II, 2019/1 10

(in date t dollars). Therefore, the “relative price of money and consumption” is:

"Price" of mt+1
z }|{
Pt+1
Pt+1 −
1 + it
= it
Pt+1
1+i
| {z }t
"Price" of ct+1

Exercise 2.5

Item 1

The budget constraint in nominal terms is similar to the standard MIU model without
labor, we only use nt = 1 − lt in the production function:

ct Pt + k t Pt + Mt + Bt = f (k t−1 , 1 − lt ) Pt + k t−1 Pt (1 − δ) + (1 + it−1 ) Bt−1 + Mt−1 + Tt

Item 2

Dividing the nominal budget constraint by Pt (the notation is the same as in Walsh’s
book)

Bt−1 M
c t + k t + m t + bt = f ( k t − 1 , 1 − l t ) + k t − 1 ( 1 − δ ) + ( 1 + i t − 1 ) + t−1 + τt
Pt Pt

But:
Mt − 1 M P m
= t −1 t −1 = t −1
Pt Pt−1 Pt 1 + πt
Bt−1 B P 1 + i t −1
(1 + i t −1 ) = (1 + i t −1 ) t −1 t −1 = bt − 1 = ( 1 + r t − 1 ) bt − 1
Pt Pt−1 Pt 1 + πt

Macroeconomía II, 2019/1 11

Note that we used the Fisher equation above. Hence, we can write:

m t −1
c t + k t + m t + bt = f ( k t − 1 , 1 − l t ) + k t − 1 ( 1 − δ ) + ( 1 + r t − 1 ) bt − 1 + + τt
1 + πt

Item 3

The problem is

∞
max
{ct ,k t ,mt ,bt ,lt }t≥0
∑ β t u ( c t , m t , lt )
t =0
m t −1
s.t. c t + k t + m t + bt = f ( k t − 1 , 1 − l t ) + k t − 1 ( 1 − δ ) + ( 1 + r t − 1 ) bt − 1 + + τt , ∀t ≥ 0
1 + πt
ct , k t , mt+1 , bt ≥ 0, lt ∈ [0, 1] ∀t ≥ 0

m0 = m > 0, b−1 = b > 0, k −1 = k > 0

Since we are assuming the solution is interior, the first order conditions must be satis-
fied at the optimal. The Lagrangian of this problem is:

∞   
m
L= ∑ βt u(ct , mt , lt ) − λt ct + k t + mt + bt − f (k t−1 , 1 − lt ) − k t−1 (1 − δ) − (1 + rt−1 ) bt−1 − t−1 − τt
1 + πt
t =0

The first order conditions are:

βt uc ( ct , mt , lt ) − βt λ t = 0 (FOC1)

− β t λ t + β t +1 λ t +1 [ f k ( k t , 1 − l t +1 ) + 1 − δ ] = 0 (FOC2)

− β t λ t + β t +1 λ t +1 (1 + r t ) = 0 (FOC3)

β t +1 λ t +1
βt u m ( ct , mt , lt ) − βt λt + =0 (FOC4)
1 + π t +1

β t u l ( c t , m t , l t ) − β t λ t f n ( k t −1 , 1 − l t ) = 0 (FOC5)

Macroeconomía II, 2019/1 12

Simplifying a bit:
uc (ct , mt , lt ) = λ0t (FOC1)

λt
f k ( k t , 1 − l t +1 ) + 1 − δ = (FOC2’)
βλt+1
λt
1 + rt = (FOC3’)
βλt+1
βλt+1
um ( ct , mt , lt ) = λt − (FOC4’)
1 + π t +1

u l ( ct , mt , lt )
= λt (FOC5’)
f n ( k t −1 , 1 − l t )

Combining (FOC1’) and (FOC5’) we get:

u l ( ct , mt , lt )
= f n ( k t −1 , 1 − l t ) (1)
uc ( ct , mt , lt )

Dividing (FOC1’) by (FOC1’) forward, multiplying both sides by 1/β and using (FOC2’)
we get:
u c ( ct , mt , lt ) λt
=
βuc (ct+1 , mt+1 , lt+1 ) βλt+1

u c ( c t , m t , l t ) = β (1 + r t ) u c ( c t +1 , m t +1 , l t +1 ) (2)

Combining (FOC2’) and (FOC3’):

f k ( k t , 1 − l t +1 ) = r t + δ (3)

Finally, dividing (FOC4’) by (FOC1’) we get:

um ( ct , mt , lt ) 1 βλt+1
= 1−
uc ( ct , mt , lt ) 1 + π t +1 λ t

Macroeconomía II, 2019/1 13

 βλt+1 1
Using (FOC2’) we have λt = 1+r t . Hence:

um ( c t , mt , lt ) 1 1
= 1− = 1− (4)
u c ( ct , mt , lt ) (1 + π t +1 ) (1 + r t ) 1 + it

Note the last equality follows from the Fisher equation.

Denote the steady state variables with superscripts “ss”. Equation (2) implies that:

1
1 + r ss =
β

Hence, (4) in the steady state becomes:

um (css , mss , l ss ) β
ss ss ss
= 1− (5)
uc (c , m , l ) 1 + π ss

Equation (3) in the steady state becomes:

1
f k (kss , 1 − l ss ) = − (1 − δ ) (6)
β

Using the budget constraint we get

mss
css = f (kss , 1 − l ss ) + τ ss + − δkss − mss (7)
1 + π ss

But since Mt − Mt−1 = Tt , we have that

Mt − Mt − 1
= τt
Pt

m t −1
τt = mt −
1 + πt

Macroeconomía II, 2019/1 14

 mss
Hence, τ ss = mss − 1+π ss . Therefore, (5) becomes:

css = f (kss , 1 − l ss ) − δkss (8)

Finally, using (1):

ul (css , mss , l ss )
= f n (kss , 1 − l ss ) (9)
uc (css , mss , l ss )

Equation (5), (6), (8) and (9) characterize the steady state in this economy. Note that in-
flation will affect the real money holdings in steady state by equation (5). Moreover, for
a given mss , (6), (8) and (9) characterize consumption, capital and leisure in the steady
ul (css ,mss ,l ss )
state, and mss only affect these variables through equation (9). Hence, if uc (css ,mss ,l ss )

does not depend on mss ,css , kss and l ss will not depend on mss , and therefore, will not
ul (css ,mss ,l ss )
depend on inflation π ss . In that case, we have superneutrality. If uc (css ,mss ,l ss )
depends on
mss then superneutrality fails: inflation affects mss , which affects the steady state level
of consumption, capital and leisure. Intuitively, inflation may distort agent decisions
of how much to work if indirectly their money holdings affect the marginal utility of
labor and consumption. In practice, it is hard to tell how money holdings affect this
marginal utility, and that is why a more micro-founded model of money (such as the
CIA) is desirable.

Exercise 2.6

Item 1

Note the notation in this exercise is a little different from the one we used in class: We
denote the money the agent decides to carry from date t to date t + 1 by Mt+1 (and
not by Mt as we did in class). Similar change of notation holds for capital (Kt+1 is the
capital decides to accumulate at date t). Also, note that we did not introduce bonds.
The budget constraint of the centralized economy is:

Macroeconomía II, 2019/1 15

 Pt ct + Pt It + Mt+1 = Pt Y (Kt , nt ) + Mt + Tt (4)

(You could replace It by Kt+1 − (1 − δ) Kt in the equation above). The budget constraint
of decentralized economy is:

Pt ct + Pt Kt+1 + Mt+1 = Pt (1 − δ)Kt + Pt wt nt + Pt rt Kt + Mt + Tt (5)

Item 2

Dividing by Pt the budget constraint of the centralized economy:

Mt + 1 Mt Tt
ct + It + = Yt + +
Pt Pt Pt

Defining mt ≡ Mt /Pt and τt ≡ Tt /Pt and using that

Mt + 1 M Pt m t +1
= t +1 =
Pt Pt+1 Pt+1 1 + π t +1

We can write it as:

m t +1
ct + It + = Yt + mt + τt
1 + π t +1

Similarly, for the decentralized economy we have:

Mt + 1 Mt Tt
c t + K t +1 + = (1 − δ ) K t + w t n t + r t K t + +
Pt Pt Pt

m t +1
c t + K t +1 + = (1 − δ)Kt + wt nt + rt Kt + mt + τt
1 + π t +1

Item 3

The homogeneity of degree one property of Y (K, N ) allows us to write:

Macroeconomía II, 2019/1 16

 Y (Kt , Nt ) = Yn (Kt , Nt )nt + Yk (Kt , Nt )Kt

In perfect competition, the maximizing firm would pay each factor according to its
marginal contribution to production:

wt = Yn (Kt , Nt ) rt = Yk (Kt , Nt ) (6)

So the nominal budget constraint (4) in the centralized economy becomes:

Pt ct + Pt It + Mt+1 = Pt (Yn (Kt , Nt )nt + Yk (Kt , Nt )Kt ) + Mt + Tt

| {z }
Y (Kt ,nt )

Replacing It = Kt+1 − (1 − δ) Kt and (6) above we get:

Pt ct + Pt Kt+1 + Mt+1 = Pt (1 − δ)Kt + Pt wt nt + Pt rt Kt + Mt + Tt

Which is identical to (5).

Exercise 2.9

Item 1

Dividing the budget constraint by Pt :

Bt−1 M
c t + k t + bt + m t = f ( k t − 1 ) + ( 1 + i t − 1 ) + t−1 + (1 − δ)k t−1 + τt
Pt Pt

Rearranging:

Bt−1 Pt−1 M P
c t + k t + bt + m t = f ( k t − 1 ) + ( 1 + i t − 1 ) + t−1 t−1 + (1 − δ)k t−1 + τt
Pt−1 Pt Pt−1 Pt

Macroeconomía II, 2019/1 17

 1 + i t −1 m
c t + k t + bt + m t = f ( k t − 1 ) + bt−1 + t−1 + (1 − δ)k t−1 + τt
1 + πt 1 + πt

And using the Fisher equation:

m t −1
c t + k t + bt + m t = f ( k t − 1 ) + ( 1 + r t − 1 ) bt − 1 + + (1 − δ)k t−1 + τt
1 + πt

Similarly, we divide the CIA constraint by Pt and rearrange:

Mt − 1
ψct ≤ + τt
Pt

Mt−1 Pt−1
ψct ≤ + τt
Pt−1 Pt
m t −1
ψct ≤ + τt
1 + πt

Item 2

The lagrangian is given by

∞   
m
L= ∑β t
u(ct ) − µt ψct − t−1 − τt
1 + πt
t =0
 
m
−λt ct + k t + bt + mt − f (k t−1 ) − (1 + rt−1 ) bt−1 − t−1 − (1 − δ)k t−1 − τt
1 + πt

The FOC with respect to ct is:

βt u0 (ct ) − λt − µt ψ = 0
 
(FOC1)

Macroeconomía II, 2019/1 18

The FOC with respect to k t is:

− β t λ t + β t +1 λ t +1 f 0 ( k t ) + 1 − δ = 0
 
(FOC2)

The FOC with respect to bt is:

− β t λ t + β t +1 λ t +1 (1 + r t ) = 0 (FOC3)

The FOC with respect to mt is:

t β t +1
− β λt + ( µ t +1 + λ t +1 ) = 0 (FOC4)
1 + π t +1

Item 3

Subtracting (FOC4) from (FOC3) and solving for µt+1 :

t +1 β t +1
β λ t +1 (1 + r t ) − ( µ t +1 + λ t +1 ) = 0
1 + π t +1

µt+1 = λt+1 [(1 + πt+1 ) (1 + rt ) − 1]

Thus, replacing µt = λt [(1 + πt ) (1 + rt−1 ) − 1] on (FOC1) we get:

u0 (ct ) − λt − ψλt [(1 + πt ) (1 + rt−1 ) − 1] = 0

Using the Fisher equation it becomes:

1−ψ
 
0
u (ct ) = ψλt 1 + i t −1 + (1)
ψ

Macroeconomía II, 2019/1 19

Iterating (1) forward and dividing by (1):

h i
1− ψ
u0 (c 1 + i t −1 + ψ
t) λt
=
u0 (c
h i
t +1 ) λ t +1 1− ψ
1 + it + ψ

λt
But (FOC3) implies λ t +1 = β (1 + rt ) and then:

 
1− ψ
u0 (c t)
1 + i t −1 + ψ
= β (1 + r t )   (2)
u0 (c t +1 ) 1 + it +
1− ψ
ψ

Thus,
1− ψ
1 + i t −1 + ψ
h ( i t −1 , i t , ψ ) = 1− ψ
.
1 + it + ψ

Item 4

When ψ = 1 we have:
u0 (ct ) β (1 + r t ) u 0 ( c t +1 )
= .
1 + i t −1 1 + it

The LHS is the marginal benefit of one unit of consumption at date t divided by a
function of the opportunity cost of consuming at t (which is the foregone interest it−1
that the agent has to incur by buying less bonds and carrying more money at t − 1).
The RHS is the marginal benefit of consuming at date t + 1 divided by a function of the
opportunity cost of consuming at t + 1 (which is the foregone interest it that the agent
has to incur by buying less bonds and carrying more money at t). For the agent to be
indifferent between consuming at both dates, this “cost-benefit” ratio must be equal.

Macroeconomía II, 2019/1 20

Exercise 2.10

Item 1

The problem is identical to the problem we had before, the only differencia is the CIA
constraint. The budget constraint in real terms remains the same, so I will skip the
derivation. Now we write the CIA constraint it as:

Pt ct ≤ Mt−1

Which rearranging becomes:

Mt − 1 M P m
ct ≤ = t −1 t −1 = t −1
Pt Pt−1 Pt 1 + πt

Hence, the household problem is:

∞
max
{ct ,k t ,Bt ,Mt }t≥0
∑ βt u (ct )
t =0
m t −1
s.t. c t + k t + bt + m t = f ( k t − 1 ) + ( 1 + r t − 1 ) bt − 1 + + (1 − δ)k t−1 + τt , ∀t ≥ 0
1 + πt
m t −1
ct ≤ , ∀t ≥ 0
1 + πt
b−1 = b ≡ B/P−1 > 0, M−1 = m = M/P−1 > 0

lim (bt ) = 0,
t→∞

ct , mt , k t ≥ 0, ∀t ≥ 0.

where u(c) = ln c.

Macroeconomía II, 2019/1 21

Item 2

The Lagrangian is given by

∞   
m
L= ∑β t
u(ct ) − µt c t − t −1
1 + πt
t =0
 
m
−λt ct + k t + bt + mt − f (k t−1 ) − (1 + rt−1 ) bt−1 − t−1 − (1 − δ)k t−1 − τt
1 + πt

The FOC with respect to ct is:

βt u0 (ct ) − λt − µt = 0
 
(FOC1)

The FOC with respect to k t is:

− β t λ t + β t +1 λ t +1 f 0 ( k t ) + 1 − δ = 0
 
(FOC2)

The FOC with respect to bt is:

− β t λ t + β t +1 λ t +1 (1 + r t ) = 0 (FOC3)

The FOC with respect to mt is:

β t +1
− βt λt + ( µ t +1 + λ t +1 ) = 0 (FOC4)
1 + π t +1

Item 3

We can rewrite (FOC1) and (FOC4) as (note that we used (FOC4) lagged):

u0 (ct ) = λt + µt

Macroeconomía II, 2019/1 22

 βλt−1
= λt + µt
1 + πt

Hence:
(1 + π t ) λ t −1
u0 (ct ) =
β

Dividing the equation above by the same equation forward and multiplying both sides
by 1/β:
1 u0 (ct ) 1 + π t λ t −1
=
β u 0 ( c t +1 ) 1 + π t +1 λ t β
λ t −1
But (FOC3) lagged immplies that λt β = 1 + rt−1 . Hence:

1 u0 (ct ) 1 + πt
0
= (1 + r t −1 )
β u ( c t +1 ) 1 + π t +1

Multiplying the RHS by (1 + rt )/(1 + rt ):

1 u0 (ct ) (1 + π t ) (1 + r t −1 )
= (1 + r t )
β u 0 ( c t +1 ) (1 + π t +1 ) (1 + r t )

Using the Fisher equation we get:

1 u0 (ct ) 1 + i t −1
0
= (1 + r t )
β u ( c t +1 ) 1 + it

Since u0 (c) = 1/c we can write:

1 + i t −1
c t +1 = β (1 + r t ) c t
1 + it

Macroeconomía II, 2019/1 23

Item 4

m t −1
Since the CIA constraint binds, we have that ct = 1+ π t for all t. Plugging that in the
Euler equation we have:

mt 1 + i t −1 m
=β (1 + r t ) t −1
1 + π t +1 1 + it 1 + πt

Mt 1 + i t −1 1 + π t +1 M
=β (1 + r t ) t −1
Pt 1 + it 1 + πt Pt−1
1 + i t −1 1 + π t +1
Mt = β (1 + rt ) Mt−1
(1+πt )

1 + it  1+ 
 πt


1 + i t −1
Mt = β ( 1 + π t + 1 ) ( 1 + r t ) Mt − 1
1 + it

Using the Fisher equation:

1 + i t −1
Mt = β (1+i
t ) Mt − 1
1+ it
 


Mt
= β (1 + i t −1 )
Mt − 1

Item 5

In scenario (a) we have that:

β (1 + i t −1 ) = 1 + µ

Hence, interest rates remain constant. In scenario (b) we have that:

β (1 + it−1 ) = 1 + µt

Hence, interest rates increase as time passes. Finally, in scenario (c) we have that:

β (1 + it−1 ) = 1 + 1/ (µt)

Macroeconomía II, 2019/1 24

Hence, interest rates decrease as time passes.
Note that in scenario (a) even though the money supply is increasing, the interest
rates remain constant. In scenario (b), the money supply is growing at a increasing rate,
and the interest are increasing. Note that this is the opposite one would expect from a
simple static model of money demand and money supply. This a common feature in
this kind of model. In a stochastic version of the CIA model, one can often find that
interest rates increase after a random monetary expansion (this is called the liquidity
puzzle).

Exercise 2.11

Item 1

Dividing the budget constraint by Pt :

Bt−1 M
c t + k t + bt + m t = R t k t − 1 + ( 1 + i t − 1 ) + t−1 + (1 − δ)k t−1 + τt + Dt
Pt Pt

Rearranging:

Bt−1 Pt−1 M P
c t + k t + bt + m t = R t k t −1 + ( 1 + i t − 1 ) + t−1 t−1 + (1 − δ)k t−1 + τt + dt
Pt−1 Pt Pt−1 Pt

1 + i t −1 m
c t + k t + bt + m t = R t k t − 1 + bt−1 + t−1 + (1 − δ)k t−1 + τt + dt
1 + πt 1 + πt

And using the Fisher equation:

m t −1
c t + k t + bt + m t = R t k t − 1 + ( 1 + r t − 1 ) bt − 1 + + (1 − δ)k t−1 + τt + dt
1 + πt

Macroeconomía II, 2019/1 25

Similarly, we divide the CIA constraint by Pt and rearrange:

Mt − 1
ct ≤ + τt
Pt

Mt−1 Pt−1
ct ≤ + τt
Pt−1 Pt
m t −1
ct ≤ + τt
1 + πt

Hence, the household solves:

∞
max
{ct ,k t ,Bt ,Mt }t≥0
∑ βt u (ct )
t =0
m t −1
s.t. c t + k t + bt + m t = R t k t − 1 + ( 1 + r t − 1 ) bt − 1 + + (1 − δ)k t−1 + τt + dt , ∀t ≥ 0
1 + πt
m t −1
ct ≤ + τt , ∀t ≥ 0
1 + πt
b−1 = b ≡ B/P−1 > 0, M−1 = m = M/P−1 > 0

lim (bt ) = 0,
t→∞

ct , mt , k t ≥ 0, ∀t ≥ 0.

Item 2

The Lagrangian is given by

∞   
m
L= ∑β t
u(ct ) − µt ct − t−1 − τt
1 + πt
t =0
 
m
−λt ct + k t + bt + mt − Rt k t−1 − (1 + rt−1 ) bt−1 − t−1 − (1 − δ)k t−1 − τt − dt
1 + πt

The FOC with respect to ct is:

βt u0 (ct ) − λt − µt = 0
 
(FOC1)

Macroeconomía II, 2019/1 26

The FOC with respect to k t is:

− β t λ t + β t +1 λ t +1 [ R t +1 + 1 − δ ] = 0 (FOC2)

The FOC with respect to bt is:

− β t λ t + β t +1 λ t +1 (1 + r t ) = 0 (FOC3)

The FOC with respect to mt is:

t β t +1
− β λt + ( µ t +1 + λ t +1 ) = 0 (FOC4)
1 + π t +1

Item 3

The first order condition of the firm is:

f 0 k̃ t = Rt


(FOC5)

Item 4

Market clearing implies that k̃ t = k t−1 . Hence, (FOC5) implies that:

f 0 ( k t −1 ) = R t

Combining (FOC2) and (FOC3) we get:

β t +1 λ t +1 [ R t +1 + 1 − δ ] = β t +1 λ t +1 (1 + r t )

r t = R t +1 − δ

Macroeconomía II, 2019/1 27

Item 5

All the FOCs are the same as in the model seen in class, except for (FOC2). Plugging
f 0 (k t−1 ) = Rt in (FOC2):

− β t λ t + β t +1 λ t +1 f 0 ( k t ) + 1 − δ = 0
 
(FOC2’)

Note that (FOC1), (FOC2’), (FOC3) and (FOC4) are identical to the first-order conditions
in the centralized model seen in class. Hence, we should arrive at the same steady state,
since we are departing from the exact same set of equilibrium conditions.

Macroeconomía II, 2019/1 28