Monetary Policy,

Ination, and the Business Cycle

Chapter 2
A Classical Monetary Model
Jordi Gal
CREI and UPF
August 2007
In the present chapter we lay out a simple model of a classical mone-
tary economy, featuring perfect competition and fully exible prices in all
markets. As stressed below, many of the predictions of that classical econ-
omy are strongly at odds with the evidence reviewed in chapter 1. That
notwithstanding, we view the analysis of the classical economy as providing
a benchmark that will be useful in subsequent chapters when some of its
strong assumptions are relaxed. It also allows us to introduce some notation,
as well as the assumptions on preferences and technology that are used in
the remainder of the book.
Following much of the recent literature, our baseline classical model at-
taches a very limited role for money. Thus, in the rst four sections the
only explicit role played by money is to serve as a unit of account. In that
case, and as shown below, whenever monetary policy is specied in terms
of an interest rate rule, we do not need to make any reference whatsoever
to the quantity of money in circulation in order to determine the economys
equilibrium. When the specication of monetary policy involves the money
supply, we postulate a "conventional" money demand equation in order to
close the model, without taking a stand on its microfoundations. In section
5, we introduce an explicit role for monay, beyond that of serving as a unit of
account. In particular we analyze a model in which real balances are assumed
to generate utility to households, and explore the implications for monetary
policy of alternative assumptions on the properties of that utility function.
Independently of how money is introduced, the proposed framework as-
sumes a representative household solving a dynamic optimization problem.
That problem and the associated optimality conditions are described in sec-
tion 1. Section 2 introduces the representative rms technology and deter-
mines its optimal behavior, under the assumption of price and wage-taking.
Section 3 characterizes the equilibrium, and shows how real variables are
uniquely determined, independently of monetary policy. Section 4 discusses
the determination of the price level and other nominal variables under al-
ternative monetary policy rules. Finally, section 5 analyzes a version of the
model with money in the utility function, and discusses the extent to which
the conclusions drawn from our earlier analysis need to be modied under
that assumption.
1
1 Households
The representative household seeks to maximize the objective function:
1
0
1

t=0
,
t
l (C
t
. `
t
) (1)
where C
t
is the quantity consumed of the single good, and `
t
denotes hours
of work or employment.
1
The period utility l (C
t
. `
t
) is assumed to be con-
tinuous and twice dierentiable, with l
c,t
=
0l(CI,.I)
0CI
0, l
cc,t
=
0
2
l(CI,.I)
0C
2
I
_ 0, l
a,t
=
0l(CI,.I)
0.I
_ 0, and l
aa,t
=
0
2
l(CI,.I)
0.
2
I
_ 0. In words, the mar-
ginal utility of consumption l
c,t
is assumed to be positive and non-increasing,
while the marginal disutility of labor, l
a,t
, is positive and non-decreasing.
Maximization of (1) is subject to a sequence of ow budget constraints
given by
1
t
C
t
+Q
t
1
t
_ 1
t1
+\
t
`
t
1
t
(2)
for t = 0. 1. 2. ... 1
t
is the price of the consumption good,. \
t
denotes the
nominal wage, 1
t
represents the quantity of one-period nominally riskless
discount bonds purchased in period t, and maturing in period t + 1. Each
bond pays one unit of money at maturity, and its price is Q
t
. 1
t
repre-
sents lump-sum additions or subtractions to period income (e.g. lump-sum
taxes, dividends, etc.), expresed in nominal terms. When solving the prob-
lem above, the household is assumed to take as given the price of the good,
the wage and the price of bonds.
In addition to (2), we assume that the household is subject to a solvency
constraint that prevents it from engaging in Ponzi-type schemes. For our
purposes the following constraint is sucient:
lim
T!1
1
t
1
T
_ 0 (3)
foa all t.
1
Alternatively, N
t
can be interpreted as the number of household members employed,
assuming a large household and ignoring integer constraints.
2
1.0.1 Optimal Consumption and Labor Supply
The optimality conditions implied by the maximization of (1) subject to (2)
are given by:

l
a,t
l
c,t
=
\
t
1
t
(4)
Q
t
= , 1
t
_
l
c,t+1
l
c,t
1
t
1
t+1
_
(5)
for t = 0. 1. 2. ....
The previous optimality conditions can be derived using a simple vari-
ational argument. Let us rst consider the impact on utility of a small
departure, in period t, from the households optimal plan. That departure
consists of an increase in consumption dC
t
and and increase in hours d`
t
,
while keeping the remaining variables unchanged (including consumption and
hours in other periods). If the household was following an optimal plan to
begin with, it must be the case that
l
c,t
dC
t
+l
a,t
.d`
t
= 0
for any pair (dC
t
. d`
t
) satisfying the budget constraint, i.e.
1
t
dC
t
= \
t
.d`
t
for otherwise it would be possible to raise utility by increasing (or decreasing)
consumption and hours, thus contradicting the assumption that the house-
hold is on an optimal plan. Note that by combining both equations we obtain
the optimality condition (4).
Similarly, we can consider the impact on expected utility as of time t of
a reallocation of consumption between periods t and t + 1, while keeping
consumption in any period other than t and t + 1, and hours worked (in all
periods) unchanged. If the household is optimizing it must be the case that
l
c,t
dC
t
+, 1
t
l
c,t+1
.dC
t+1
= 0
for any pair (dC
t
. dC
t+1
) satisfying
1
t+1
dC
t+1
=
1
t
Q
t
dC
t
3
where the latter equation determines the increase in consumption expendi-
tures in period t+1 made possible by the additional savings 1
t
dC
t
allocated
into one-period bonds. Combining the two previous equations we obtain the
intertemporal optimality condition (5).
In much of what follows we assume that the period utility takes the form:
l(C
t
. `
t
) =
C
1o
t
1 o

`
1+,
t
1 +,
The consumers optimality conditions (4) and (5) thus become:
\
t
1
t
= C
o
t
`
,
t
(6)
Q
t
= , 1
t
_
_
C
t+1
C
t
_
o
1
t
1
t+1
_
(7)
Note, for future reference, that equation (6) can be re-written in log-linear
form as follows:
n
t
j
t
= o c
t
+, :
t
(8)
where lower case letters denote the natural logs of the corresponding variable
(i.e. r
t
= log A
t
). The previous condition can be interpreted as a compet-
itive labor supply schedule, determining the quantity of labor supplied as a
function of the real wage, given the marginal utility of consumption (which
under our assumptions is a function of consumption only).
As shown in Appendix 1, a log-linear approximation of (7) around a
steady state with constant rates of ination and consumption growth is given
by
c
t
= 1
t
c
t+1

1
o
(i
t
1
t
:
t+1
j) (9)
where we have dened i
t
= log Q
t
and j = log ,. Notice that i
t
corre-
sponds to the log of the gross yield on the one-period bond; we henceforth
refer to it as the nominal interest rate.
2
Similarly, j can be interpreted as
the households discount rate.
2
The yield on the one period bond is dened by Q
t
= (1 + yield)
1
. Note that
i
t
= log Q
t
= log(1 + yield
t
) yield
t
where the latter approximation will be accurate
as long as the nominal yield is "small."
4
While the previous framework does not explicitly introduce a motive for
holding money balances, in some cases it will be convenient to postulate a
demand for real balances with a log-linear form given by (up to an additive
constant):
:
t
j
t
=
t
j i
t
(10)
where j _ 0 denotes the interest semi-elasticity of money demand.
A money demand equation similar to (10) can be derived under a variety
of assumptions. For instance, in section 5 below we derive it as an optimality
condition for the household when money balances yield utility.
2 Firms
We assume a representative rm whose technology is described by a produc-
tion function given by
1
t
=
t
`
1c
t
(11)
where
t
represents the level of technology. We assume c
t
= log
t
evolves
exogenously according to some stochastic process.
Each period the rm maximizes prots
1
t
1
t
\
t
`
t
(12)
subject to (11), and taking the price and wage as given.
Maximization of (12) subject to (11) yields the optimality condition
\
t
1
t
= (1 c)
t
`
c
t
(13)
i.e. the rm hires labor up to the point where its marginal product equals
the real wage. Equivalently, the marginal cost,
WI
(1c)I .
o
I
, must be equated
to the price, 1
t
.
In log-linear terms, we have
n
t
j
t
= c
t
c :
t
+ log(1 c) (14)
which can be interpreted as labor demand schedule, mapping the real wage
into the quantity of labor demanded, given the level of technology.
5
3 Equilibrium
Our baseline model abstracts from aggregate demand components like invest-
ment, government purchases, or net exports. Accordingly, the goods market
clearing condition is given by

t
= c
t
(15)
i.e. all output must be consumed.
By combining the optimality conditions of households and rms with (15)
and the log-linear aggregate production relationship

t
= c
t
+ (1 c) :
t
(16)
we can determine the equilibrium levels of employment and output, as a
function of the level of technology:
:
t
=
ao
c
t
+0
a
(17)

t
=
jo
c
t
+0
j
(18)
where
ao
=
1o
o(1c)+,+c
, 0
a
=
log(1c)
o(1c)+,+c
,
jo
=
1+,
o(1c)+,+c
, and 0
j
=
(1 c)0
a
.
Furthermore, given the equilibrium process for output, we can use (9) to
determine the implied real interest rate, :
t
= i
t
1
t
:
t+1

:
t
= j +o 1
t

t+1

= j +o
jo
1
t
c
t+1
(19)
Finally, the equilibrium real wage, .
t
= n
t
j
t
. is given by
.
t
= c
t
c :
t
+ log(1 c) (20)
=
.o
c
t
+0
.
where
.o
=
o+,
o(1c)+,+c
and 0
.
=
(o(1c)+,) log(1c)
o(1c)+,+c
.
Notice that the equilibrium dynamics of employment, output, and the
real interest rate are determined independently of monetary policy. In other
words, monetary policy is neutral with respect to those real variables. In our
simple model output and employment uctuate in response to variations in
6
technology, which is assumed to be the only real driving force.
3
In particu-
lar, output always rises in the face of a productivity increase, with the size
of the increase being given by
jo
0. The same is true for the real wage.
On the other hand, the sign of the employment is ambiguous, depending on
whether o (which measured the strength of the wealth eect of labor sup-
ply) is larger or smaller than one. When o < 1, the substitution eect on
labor supply resulting from a higher real wage dominates the negative eect
caused by a smaller marginal utility of consumption, leading to an increase
in employment. The converse is true whenever o 1. When the utility
of consumption is logarithmic (o = 1) employment remains unchanged in
the face of technology variations, for substitution and wealth eects exactly
cancel one another. Finally, the response of the real interest rate depends
critically on the time series properties of technology. If the current improve-
ment in technology is transitory, so that 1
t
c
t+1
< c
t
, then the real rate
will go down. Otherwise, if technology is expected to keep improving, then
1
t
c
t+1
c
t
and the real rate will increase with a rise in c
t
.
What about nominal variables, like ination or the nominal interest rate?
Not surprisingly, and in contrast with real variables, their equilibrium behav-
ior cannot be determined uniquely by real forces. Instead, it requires that
we specify how monetary policy is conducted. Below we consider several
monetary policy rules and their implied outcomes.
4 Monetary Policy and Price Level Determi-
nation
We start by examining the implications of some interest rate rules. Later we
introduce rules that involve monetary aggregates. In all cases we make use
of the Fisherian equation:
i
t
= 1
t
:
t+1
+:
t
(21)
which implies that the nominal rate adjusts one-for-one with expected ina-
tion, given a real interest rate that is determined exclusively by real factors,
as in (19).
3
It would be straightforward to introduce other real driving forces like variations in
government purchages or exogenous shifts in preferences. In general, real variables will
aected by all those real shocks in equilibrium.
7
4.1 An Exogenous Path for the Nominal Interest Rate
Let us rs consider the case of the nominal interest rate following an exoge-
nous stationary process i
t
. Without loss of generality we assume that i
t
has mean j, which is consistent with a steady state with zero ination and
no secular growth. Notice that a particular case of this rule corresponds to
a constant interest rate i
t
= i = j, for all t.
Using (21) we can write,
1
t
:
t+1
= i
t
:
t
where, as discussed above, :
t
is determined independently of the monetary
policy rule.
Note that expected ination is pinned down by the previous equation.
But actual ination is not. Since there is no other condition that can be
used to determine ination, it follows that any path for the price level that
satises
j
t+1
= j
t
+i
t
:
t
+
t+1
is consistent with equilibrium, where
t+1
is a shock, possibly unrelated to
economic fundamentals, satisfying 1
t

t+1
= 0 for all t. Such shocks are
often referred to in the literature as sunspot shocks. We refer to an equilib-
rium in which such non-fundamental factors may cause uctuations in one
or more variables as an indeterminate equilibrium. In the example above,
we have thus shown how an exogenous nominal interest rate leads to price
level indeterminacy.
Notice that when (10) is operative the equilibrium path for the money
supply (which is endogenous under the present policy regime) is given by
:
t
= j
t
+
t
j i
t
Hence, the money supply will inherit the indeterminacy of j
t
. The same
will be true of the nominal wage (which, in logs, equals the real wage, which
determined by (20), plus the price leve, which is indeterminate).
4.2 A Simple Ination-Based Interest Rate Rule
Suppose that the central bank adjusts the nominal interest rate according to
the rule
i
t
= j +c

:
t
8
where c

_ 0.
Combining the previous rule with the Fisherian equation (21) we obtain
c

:
t
= 1
t
:
t+1
+ :
t
(22)
where :
t
= :
t
j. We distinguish between two cases, depending on whether
the coecient on ination in the above rule, c

, is larger or smaller than one.

If c

1, the previous dierence equation has only one stationary solu-

tion, i.e. a solution that remains in a neighborhood of the steady state. That
solution can be obtained by solving (22) forward, which yields
:
t
=
1

I=0
c
(I+1)

1
t
:
t+I
(23)
The previous equation fully determines ination (and, hence, the price
level) as a function of the path of the real interest rate, which in turn is a
function of fundamentals, as shown in (19). Consider, for the sake of illus-
tration, the case in which technology follows the stationary AR(1) process
c
t
= j
o
c
t1
+
o
t
where j
o
[0. 1). Then (19) implies :
t
= o
jo
(1 j
o
) c
t
, which combined
with (23) yields the following expression for equilibrium ination:
:
t
=
o
jo
(1 j
o
)
c

j
o
c
t
Note that a central bank following a rule of the form considered here can
inuence the degree of ination volatility by choosing the size of c

. The
larger is the latter parameter the smaller will be the impact of the real shock
on ination.
On the other hand, if c

< 1 the stationary solutions to (22) take the

form
:
t+1
= c

:
t
:
t
+
t+1
(24)
where
t
is, again, an arbitrary sequence of shock, possibly unrelated to
fundamentals, satisfying 1
t

t+1
= 0 all t.
Accordingly, any process :
t
satisfying (24) is consistent with equilib-
rium, while remaining in a neighborhood of the steady state. So, as in the
9
case of an exogenous nominal rate, the price level (and, hence, ination and
the nominal rate) are not determined uniquely when the interest rate rule
implies a weak response of the nominal rate to changes in ination. More
specically, the condition for a determinate price level, c

1 , requires that
the central bank adjust nominal interest rates more than one-for-one in re-
sponse to any change in ination, a property known as the Taylor principle.
The previous result can be viewed as a particular instance of the need to
satisfy the Taylor principle in order for an interest rate rule to bring about
a determinate equilibrium.
4.3 An Exogenous Path for the Money Supply
Suppose that the central bank sets an exogenous path for the money supply
:
t
. Using (10) to eliminate the nominal interest rate in (21), we can derive
the following dierence equation for the price level:
j
t
=
_
j
1 +j
_
1
t
j
t+1
+
_
1
1 +j
_
:
t
+n
t
where n
t
= (1 +j)
1
(j :
t

t
) evolves independently of :
t
.
Assuming j 0 and solving forward we obtain:
j
t
=
1
1 +j
1

I=0
_
j
1 +j
_
I
1
t
:
t+I
+n
0
t
where n
0
t
=

1
I=0
_
j
1+j
_
I
1
t
n
t+I
is, again, independent of monetary
policy.
Equivalently, we can rewrite the previous expression in terms of expected
future growth rate of money:
j
t
= :
t
+
1

I=1
_
j
1 +j
_
I
1
t
:
t+I
+n
0
t
(25)
Hence, we see how an arbitrary exogenous path for the money supply al-
ways determines the price level uniquely. Given the price level, as determined
above, we can then use (10) to solve for the nominal interest rate:
10
i
t
= j
1
[
t
(:
t
j
t
)]
= j
1
1

I=1
_
j
1 +j
_
I
1
t
:
t+I
+n
00
t
where n
00
t
= j
1
(n
0
t
+
t
) is independent of monetary policy.
As an example, consider the case in which money growth follows an AR(1)
process.
:
t
= j
n
:
t1
+
n
t
For simplicity let us assume the absence of real shocks, thus implying a
constant output and a constant real rate. Without loss of generality, we set
:
t
=
t
= 0 for all t. Then it follows from (25) that
j
t
= :
t
+
jj
n
1 +j(1 j
n
)
:
t
Hence, in response to an exogenous monetary policy shock, and as long as
j
n
0 (the empirically relevant case, given the observed positive autocorre-
lation of money growth), the price level should respond more than one-for-one
with the increase in the money supply, a prediction which contrasts starkly
with the sluggish response of the price level observed in empirical estimates
of the eects of monetary policy shocks, as discussed in chapter 1.
The nominal interest rate is in turn given by
i
t
=
j
n
1 +j(1 j
n
)
:
t
i.e. in response to an expansion of the money supply, an as long as j
n
0, the
nominal interest rate is predicted to go up. In other words, the model implies
the absence of a liquidity eect, in contrast with the evidence discussed in
chapter 1.
4.4 Optimal Monetary Policy
The analysis of the baseline classical economy above has shown that while
real variables are independent of monetary policy, the latter can have im-
portant implications for the behavior of nominal variables and, in particular,
11
of prices. Yet, and given that the households utility is a function of con-
sumption and hours onlytwo real variables that are invariant to the way
monetary policy is conducted it follows that there is no policy rule that
is better than any other. Thus, in the classical model above, a policy that
generates large uctuations in ination and other nominal variables (perhaps
as a consequence of following a policy rule that does not guarantee a unique
equilibrium for those variables) is no less desirable that one that succeeds in
stabilizing prices in the face of the same shocks.
The previous result, which is clearly extreme and empirically unappealing,
can be overcome once we consider versions of the classical monetary model in
which a motive to keep part of households wealth in the form of monetary
assets is introduced explicitly. Section 5 discusses one such model, in which
real balances are assumed to yield utility.
Our overall assessment of the classical monetary model as a framework
to understand the joint behavior of nominal and real variables and their con-
nection to monetary policy cannot be positive. The model cannot explain
the observed real eects of monetary policy on real variables. Its predictions
regarding the response of the price level, the nominal rate and the money
supply to exogenous monetary olicy shocks are also in conict with the em-
pirical evidence. Those empirical failures are the main motivation behind
the introduction of nominal frictions in othersise similar model, a task that
we undertake in chapter 3.
5 Money in the Utility Function
4
In the model developed in the previous sections, and in much of the recent
monetary literature, the only role played by money is to serve as a numraire,
i.e. unit of account in which prices, wages and securities payos are stated.
Economies with that characteristic are often referred to as cashless economies.
Whenever we have postulated a simple log linear money demand function,
we have done so in an ad-hoc manner, without an explicit justication for
why agents would want to hold an asset that is dominated in return by
bonds, while having identical risk properties. Even though in the analysis of
subsequent chapters we will stick to the assumption of a cashless economy,
4
The reader may skip this section and proceed to section 6 without any loss of conti-
nuity.
12
it is useful to understand how the basic framework can incorporate a role for
money other than that of a unit of account and, in particular, how it can
generate a demand for money. The discussion in the present section focuses
on models that achieve the previous objective by assuming that real balances
are an argument of the utility function.
The introduction of money in the utility function requires that we modify
the household s problem in two ways. First, preferences are now given by
1
0
1

t=0
,
t
l
_
C
t
.
`
t
1
t
. `
t
_
(26)
where `
t
denotes holdings of money in period t. We assume that period
utility is increasing and concave in real balances `
t
,1
t
. Secondly, the ow
budget constraint incorporates monetary holdings explicitly, taking the fol-
lowing form:
1
t
C
t
+Q
t
1
t
+`
t
_ 1
t1
+`
t1
+\
t
`
t
1
t
Letting /
t
= 1
t1
+`
t1
denote total nancial wealth at the beginning
of the period t (i.e. before consumption and portfolio decisions are made),
we can rewrite the previous ow budget constraint as:
1
t
C
t
+Q
t
/
t+1
+ (1 Q
t
) `
t
_ /
t
+\
t
`
t
1
t
(27)
with the solvency constraint taking now the form lim
T!1
1
t
/
T
_ 0.
The previous representation of the budget constraint can be thought of
as equivalent to that of an economy in which all nancial assets (represented
by /
t
) yield a gross nominal return Q
1
t
(= expi
t
), and where agents can
purchase the utility-yielding "services" of money balances at a unit price
(1 Q
t
) = 1 expi
t
i
t
. Thus, we see that the implicit price for money
services roughly corresponds to the nominal interest rate, which in turn is
the opportunity cost of holding ones nancial wealth in terms of monetary
assets, instead of interest-bearing bonds.
Consider next the households problem, which consists of maximizing (26)
subject to (27). Two of the implied optimality conditions are the same as
those obtained for the cashless model, i.e. (6) and (7), with the marginal util-
ity terms being now dened over (and evaluated at) the triplet (C
t
.
AI
1I
. `
t
).
In addition to (6) and (7), there is an additional optimality condition given
by
l
n,t
l
c,t
= 1 expi
t
(28)
13
where l
n,t
=
0l

CI,
L
I
T
I
,.I

0(AI1I)
0.
Again, in order to derive that optimality condition we can use a simple
variational argument. Suppose that the household is considering a deviating
from the optimal plan by adjusting consumption and money holdings in pe-
riod t by amounts dC
t
and d`
t
respectively, while keeping all other variables
unchanged at their optimal values. Optimality of the initial plan requires
that utility cannot be raised as a result of the deviation, i.e.
l
c,t
dC
t
+l
n,t
1
1
t
d`
t
= 0
for any pair (dC
t
. d`
t
) satisfying
1
t
dC
t
+ (1 Q
t
) d`
t
= 0
which guarantees that the budget constraint is met without the need to
adjust any other variable. Combining the previous two equations and using
the denition of the nominal rate i
t
= log Q
t
yields the optimality condition
(28).
In order to be able to make any statements about the consequences of
having money in the utility function we need to be more precise about the way
money balances interact with other variables in yielding utility. In particular,
whether the utility function is separable or not in real balances determines
the extent to which the neutrality properties derived above for the cashless
economy carry over to the economy with money in the utility function. We
illustrate that point by considering, in turn, two example economies with
separable and non-separable utility.
5.1 An Example with Separable Utility
We specify the households utility function to have the functional form
l
_
C
t
.
`
t
1
t
. `
t
_
=
C
1o
t
1 o
+
(`
t
,1
t
)
1i
1 i

`
1+,
t
1 +,
Note that, given the separability of real balances, neither l
c,t
nor l
a,t
depend on the level of real balances. As a result, (6) and (7) (as well as their
log-linear counterparts, (8) and (9)) continue to hold unchanged. It follows
that we can determine the equilibrium values for output, employment, the
14
real rate and the real wage following the same steps as above, and without
any reference to monetary policy.
The introduction of money in the utility function, allows us to derive a
money demand equation from the households optimal behavior. Using the
above specication of utility we can rewrite the optimality condition (28) as:
`
t
1
t
= C
oi
t
(1 expi
t
)
1i
(29)
which can be naturally interpreted as a demand for real balances. The latter
is increasing in consumption and inversely related to the nominal interest
rate, as in conventional specications.
Using the rst-order Taylor approximation log(1 expi
t
) co::t. +
1
expfig1
i
t
, we can rewrite (29) in approximate log-linear form (and up to an
uninteresting constant) as:
:
t
j
t
=
o
i
c
t
j i
t
(30)
where j =
1
i(expfig1)

1
ii
is the implied interest semi-elasticity of money
demand.
The particular case of i = o is an appealing one, since it implies a unit
elasticity with respect to consumption. Under that assumption, we obtain a
conventional linear demand for real balances
:
t
j
t
= c
t
j i
t
(31)
=
t
j i
t
where the second equality holds in our baseline model economy, in which all
output is consumed. The previous specication is often assumed in subse-
quent chapters, without the need to invoke its source explicitly.
As in the analysis of the cashless economy, the usefulness of (30) (or (31))
is conned to the determination of the equilibrium values for ination and
other nominal variables whenever the description of monetary policy involves
the quantity of money in circulation. Otherwise, the only use of the money
demand equation is to determine the quantity of money that the central bank
will need to supply in order to support, in equilibrium, the nominal interest
rate implied by the policy rule.
15
5.2 An Example with Non-Separable Utility
Let us consider next an economy in which period utility is given by
l
_
C
t
.
`
t
1
t
. `
t
_
=
A
1o
t
1 o

`
1+,
t
1 +,
where A
t
is a composite index of consumption and real balances dened as
follows
A
t
=
_
(1 0) C
1i
t
+0
_
`
t
1
t
_
1i
_ 1
1r
,o: i ,= 1
= C
1
t
_
`
t
1
t
_

,o: i = 1
with i represents the (inverse) elasticity of substitution between consumption
and real balances, and 0 the relative weight of real balances in utility.
Notice that the marginal utilities of consumption and real balances are
now given, respectively, by
l
c,t
= (1 0) A
t
io
C
i
t
l
n,t
= 0 A
t
io
_
`
t
1
t
_
i
whereas the marginal (dis)utility of labor is, as before, given by l
a,t
= `
,
t
.
The optimality conditions of the households problem, (4), (5) and (28), can
now be written as:
\
t
1
t
= `
,
t
A
t
oi
C
i
t
(1 0)
1
(32)
Q
t
= , 1
t
_
_
C
t+1
C
t
_
i
_
A
t+1
A
t
_
io
1
t
1
t+1
_
(33)
`
t
1
t
= C
t
(1 expi
t
)

_
0
1 0
_1

(34)
Notice that in the particular case in which the intertemporal and in-
tratemporal elasticities of substitution coincide (i.e. i = o), optimality con-
ditions (32) and (33) match exactly those obtained in the case of separable
16
utility, and thus lead to the same equilibrium implications derived for that
case and discussed in the previous subsection..
In the general case, however, both the labor supply equation (32) and
the Euler equation (33) are inuenced by the level of real balances, through
the dependence of the index A
t
on the latter. The level of real balances
depends, in turn, on the nominal interest rate (as implied by (34)). Those
features imply that monetary policy is no longer neutral in the case of non-
separable utiillity considered here. In particular, to the extent that dierent
monetary policy rules have dierent implications for the path of the nominal
rate (as will generally be the case), they will also have dierent eects on
real balances andthrough the latters inuence on the marginal utility of
consumptionon the position of the labor supply schedule and, hence, on
employment and output. This mechanism is analyzed formally below.
Notice that the implied money demand equation (34) can be rewritten in
log-linear form (and up to an additive constant) as in (31) above, i.e.
:
t
j
t
= c
t
j i
t
(35)
where, again, j =
1
i(expfig1)
. Thus, the implied interest semi-elasticity of
demand j is now proportional to the elasticity of substitution between real
balances and consumption, i
1
.
On the other hand, log-linearization of (32) around the zero ination
steady state yields
n
t
j
t
= oc
t
+,:
t
+ (i o)(c
t
r
t
)
Log-linearizing the expression dening A
t
around a zero ination steady
state, and combining the resulting expression with (34) we obtain
n
t
j
t
= oc
t
+,:
t
+(i o) [c
t
(:
t
j
t
)]
= oc
t
+,:
t
+j(i o) i
t
where =

1
(1o)
1
1

(1)
1
+
1
(1o)
1
1

[0. 1), and where the second equality makes

use of (35).
For future reference it is convenient to rewrite the previous optimality
conditions in terms of the steady state ratio /
n
=
A1
C
, i.e. the inverse
consumption velocity. Using the money demand equation, we have /
n
=
_

(1o)(1)
_1

. Noting that =
Ir(1o)
1+Ir(1o)
, and using the denition of j
17
evaluated at the zero ination steady state we can rewrite the optimality
condition above as
n
t
j
t
= oc
t
+,:
t
+. i
t
(36)
where . =
Iro(1

)
1+Ir(1o)
. Thus, we see that the sign of the eect of the nominal
interest rate on labor supply is determined by the sign of i o. When i o
(implying . 0) the reduction in real balances induced by an increase in
the nominal rate brings down the marginal utility of consumption (for any
given c
t
), lowering the quantity of labor supplied at any given real wage.
The opposite eect obtains when i < o. Note, however, that i
1
ij
is likely
to be larger than o for any plausible values of j and o. Thus, the case of
l
cn
0 (and hence . 0) appears as the most plausible one, conditional
on the specication of preferences analyzed here.
The corresponding log-linear approximation to (33) is given by
c
t
= 1
t
c
t+1

1
o
(i
t
1
t
:
t+1
(i o) 1
t
(c
t+1
r
t+1
) (c
t
r
t
) j)
= 1
t
c
t+1

1
o
(i
t
1
t
:
t+1
(i o) 1
t
c
t+1
(:
t+1
j
t+1
) j)
= 1
t
c
t+1

1
o
(i
t
1
t
:
t+1
. 1
t
i
t+1
j) (37)
where, again, the last equality makes use of (35). Thus, when i o (and,
hence, . 0) the anticipation of a nominal rate increase (and, hence, of
a decline in real balances), lowers the expected one period ahead marginal
utility of consumption (for any expected c
t+1
), which induces an increase in
current consumption (in order to smooth marginal utility over time).
In order to reect the changes implied by non-separable utility, we need to
modify the economys log-linearized equilibrium conditions. Thus, combin-
ing (36) with the labor demand schedule (14) we obtain the following labor
market clearing condition:
oc
t
+,:
t
+. i
t
=
t
:
t
+ log(1 c) (38)
which we can rewrite, using the goods market clearing condition (15) and the
log-linear production relationship (16) as (ignoring an uninteresting additive
constant):

t
=
jo
c
t

ji
i
t
(39)
where
ji
=
.(1c)
o+,+c(1o)
.
18
Condition (39) points to a key implication of the property of non-separability
(. ,= 0): equilibrium output is no longer invariant to monetary policy, at least
to the extent that the latter implies variations in the nominal interest rate.
In other words, monetary policy is not neutral. As a result, equilibrium con-
dition (39) does not suce to determine the equilibrium level of output, in
contrast with the economy with separable utility analyzed above. In order to
pin down the equilibrium path of output and other endogenous variables we
need to combine (39) with the remaining equilibrium conditions, including a
description of monetary policy.
One such additional condition can be obtained by imposing the goods
market clearing condition
t
= c
t
on Euler equation (37), which yields an
equation relating the nominal interest rate to the expected path of output
and expected ination:

t
= 1
t

t+1

1
o
(i
t
1
t
:
t+1
. 1
t
i
t+1
j) (40)
Finally, we need an equation which describes how monetary policy is
conducted. For the purposes of illustration we assume that the central bank
follows the simple ination-based interest rate rule
i
t
= j +c

:
t
+
t
(41)
where
t
now represents an exogenous policy disturbance, assumed to follow
the stationary AR(1) process

t
= j

t1
+

t
Similarly, and for concreteness, we assume that the technology parameter
follows the AR(1) process
c
t
= j
o
c
t1
+
o
t
Using (41) to eliminate the nominal rate in (39) and (40), and combining
the resulting two equations we can obtain (after some algebraic manipulation)
the following closed form expressions for the equilibrium level of ination, the
19
nominal rate, and output:
:
t
=
o(1 j
o
)
jo
c

(1 +.)(1 j
o
)
c
t

1 + (1 j

).
c

(1 +.)(1 j

)

t
i
t
=
o(1 j
o
)
jo
(1 +.)(1 j
o
)
c
t

(1 +.)(1 j

)

t

t
=
jo
_
1 +
o(1 j
o
)
ji
(1 +.)(1 j
o
)
_
c
t
+
j

ji
c

(1 +.)(1 j

)

t
where =
1+.
r
(1+.)
r
and =
c+,
o(1c)+c+,
.
A few remarks regarding the impact of monetary policy on the economys
equilibrium are in order. First, note that the interest rate multiplier of
output, conditional on an exogenous monetary policy shock is given by
ojI
oiI
=
ojIoI
oiIoI
=
ji
. In order to get a sense for the magnitude of that multiplier,
recall that
ji
=
.(1c)
o+,+c(1o)
. Let us assume parameter values o = , = 1
and c = 1,3, as in the baseline calibration that will be introduced in chapter
3. Using the denition of ., and the fact that i =
1
jj
is "large" for any
reasonable values of j, we have
ji

Ir
3
, and so the size of the inverse
velocity /
n
is a key determinant of the quantitative importance of monetary
non-neutralities in the model. Unfortunately, the magnitude of /
n
depends
crucially on the denition of money used. Thus, and focusing on postwar
U.S. data, /
n
0.3 if we take the monetary base as the relevant measure
of money.
5
In that case we have
ji
0.1, which implies a relative small
multiplier: a monetary policy shock that raised the nominal rate by one
percentage point (expressed at an annual rates) would generate a decrease
in output of about 0.025 percent. By way of contrast, if we use `2 as
the denition of money, we have /
n
3 and so the impact on output of
an analogous monetary policy shock is a 0.25 percent decline. The latter
value, while small, appears to be closer to the estimated output eects of a
monetary policy shock found in the literature. Yet, even in the latter case,
there are other aspects of the transmission of monetary policy shocks implied
by the model that are clearly at odds with the evidence, e.g. the response of
ination and the real interest rate. Thus, note that
d:
t
di
t
=
d:
t
,d
t
di
t
,d
t
= (1 + (1 j

).) j
1

0
5
This is the approach followed in Woodford (2003, chapter 2).
20
d:
t
di
t
= 1
d1
t
:
t+1
,d
t
di
t
,d
t
= (1 j

). < 0
i.e. in response to a monetary policy shock that raises the nominal interest
rate and lowers output, ination tends to increase, and the real rate to go
down (as a result of the dominant eect of higher expected ination). This
contrasts with the downward adjustment of ination and the rise in the real
rate observed as part of the economys response of the economy following a
contractionary monetary policy shock.
Finally, there is an additional argument that can be brought up and which
calls into question the relevance of the transmission mechanism underlying
the classical model with non-separable preferences and which has to do with
its implications regarding the long-run eects of monetary policy. To see this,
consider an exogenous monetary policy intervention that raises the nominal
rate permanently. The implied permanent change in output is determined by
(39), and given by
ji
. Thus, the long-run trade-o between output and the
nominal rate is identical to the short-run trade-o. How about the ination-
output trade-o? Equation (40), evaluated at the steady state, requires a
long-run increase in ination of the same size as the increase in the nominal
rate. Hence the long-run trade-o between ination and output is also given
by
ji
. But note that the same coecient describes the short-run output-
ination trade-o since, in the relevant case of a permanent policy change
(j

= 1), we have
ojIoI
oIoI
=
ji
.
As argued above, for a most plausible range of parameter values we have

ji
0. Thus, in the present model a permanent increase in ination will be
associated with a permanent decline in output. Given the determinants of

ji
, whether that long-run trade-o is large or small will largely depend on
the size of inverse velocity /
n
and, hence, on the relevant measure of money.
Thus, the lack of a signicant empirical relationship between long-run ina-
tion and economic activity (at least at low levels of ination), suggests a low
value for /
n
and
ji
, as implied by a narrow denition of money. Unfortu-
nately, in the present model, and as argued above, any calibration with the
desirable feature of a negligible long-run trade-o will also be associated with
negligible (and hence counterfactual) short run eects of monetary policy.
21
5.3 Optimal Monetary Policy in a Classical Economy
with Money in the Utility Function
In this section we derive the form of the optimal monetary policy in the
presence of money in the utility function. We start by laying out and solving
the problem facing a hypothetical social planner seeking to maximize the
utility of the representative household.
Note that, under our assumptions, there are no aggregate intertemporal
links in our simple model: even though each individual household can reallo-
cate its own consumption over time through nancial markets, there are no
mechanisms that make this possible for the economy as a whole. Thus, the
social planner would solve a sequence of static problems of the form
max l
_
C
t
.
`
t
1
t
. `
t
_
subject to the resource constraint
C
t
=
t
`
1c
t
The optimality conditions for that problem are given by

l
a,t
l
c,t
= (1 c)
t
`
c
t
(42)
l
n,t
= 0 (43)
Condition (42) requires that the marginal rate of substitution between
hours of work and consumption be equated to the marginal product of labor.
Condition (43) equates the marginal utility of real balances to the "social"
marginal cost of producing real balances, which is implicitly assumed to be
zero in our setting.
Under what conditions the equilibrium of the decentralized economy sat-
ises eciency conditions (42) and (43)? We rst note that condition (42)
is implied by the combined eect of prot maximization by rms (which
equates the real wage to the marginal product of labor; see equation (13))
and the optimal labor supply choice by the household (which equates the
real wage to the marginal rate of substitution between hours of work and
consumption; see equation (4)). Hence, (42) will be satised independently
22
of monetary policy. On the other hand, and as shown above, the households
optimal choice of money balances requires
l
n,t
l
c,t
= 1 expi
t

Accordingly, eciency condition (43) will be satised.if and only if i

t
= 0
for all t, a policy known as the Friedman rule. The rationale for that policy
is quite intuitive: while the social cost of producing real balances is zero, the
private (opportunity) cost is given by the nominal interest rate. As a result,
only when the nominal interest rate is zero are the private and social costs of
holding money equated. Note that such a policy implies an average (steady
state) rate of ination
: = j < 0
i.e. prices will decline on average at the rate of time preference. In other
words: under the Friedman rule the economy experiences a (moderate) de-
ation in the long-run.
Implementation of the Friedman rule requires some discussion. As shown
earlier a policy rule of the form i
t
= 0 for all t leaves the price level indetermi-
nate in our model. Even though that indeterminacy should not have any wel-
fare consequences (since (42) and (43) pin down consumption, employment
and real balances uniquely), a central bank could avoid that indeterminacy
by following a rule of the form
i
t
= c (:
t1
+:
t
)
for some c 1. Combined with (21) that rule implies the dierence equation
1
t
i
t+1
= c i
t
whose only stationary solution is i
t
= 0 for all t. Under that rule, equilibrium
ination is fully predictable and given by
:
t
= :
t1
More generally, any rule that makes the central bank adjust its policy
settings (e.g. the money supply) to guarantee that current ination moves
inversely, and one-for-one with the lagged real interest rate will imply a zero
nominal interest rate and, thus, an ecient amount of real balances.
23
6 Notes on the Literature
The modelling approach favored in much of the recent monetary literature,
and the one adopted in the present book (with the exception of section 5 of
this chapter), does not incorporate monetary assets ("money") explicitly in
the analysis. Under that approach the main role played by money is that of
a unit of account. Such model economies can be viewed as a limiting case
(the cashless limit) of an economy in which money is valued and held by
households. Woodford (2003) provides a detailed discussion and a forceful
defense of that approach.
Models that introduce monetary assets explicitly rely on one of two alter-
native formalisms in order to generate a demand for an asset thatas is the
case with moneyis dominated in its rate of return by alternative assets that
have identical risk characteristics: they either assume (i) that real balances
generate utility to households or, alternatively, (ii) that the presence of some
transaction costs in the purchases of goods can be reduced by households
holding of monetary assets.
The rst of those approaches money in the utility function traces
back to Sidrauski (1967), who introduced that assumption in an otherwise
standard neoclassical growth model (with inelastic labor supply). Woodford
(2003) oers a detailed analysis of the implications of alternative assump-
tions on the specication of utility and, in particular, of the likely degree of
monetary non-neutralities arising from the non-separability of real balances.
Walsh (2003, chapter 2) develops a real business cycle model with money in
the utility function, and analyzes the equilibrium properties of a calibrated
version of that model. Both analyses conclude, in a way consistent with the
discussion above, that even under a utility that is non-separable in real bal-
ances, the real eects of monetary policy are quantitatively very small for
plausible calibrations of the models.
A common approach to the modelling of a transactions motive for holding
money builds on the assumption, originally due to Clower (1967), that cash
mush be held in advance in order to purchase certain goods. Early examples
of classical monetary models in which a demand for money is generated
by postulating a cash-in-advance constraint can be found in the work of
Lucas (1982) and Svensson (1985). Cooley and Hansen (1989 ) analyze an
otherwise standard real business cycle models augmented with a cash-in-
advance constraint for consumption goods, showing that monetary policy is
near-neutral for plausible calibrations of that model. Walsh (2003, chapter
24
3) provides a detailed description of classical monetary models with cash-in-
advance constraints and their implications for the role of monetary policy.
The practice, followed in the present monograph, of appending a money
demand equation to a set of equilibrium conditions that have been derived
in the context of cashless economy is often found in the literature. King and
Watson (1995) is an example of that practice.
The analysis of the form of the optimal monetary policy in a classical
economy goes back to Friedman (1969), where a case is made for a policy
that keeps the nominal interest rate constant at a zero level. More recent
treatments of the conditions under which is optimal include Woodford (1990)
and Correia and Teles (1999).
Finally, the reader can nd two useful discussions of the notion of mon-
etary neutrality and its evolution in macroeconomic thinking in Patinkin
(1987) and Lucas (1996).
25
Appendix 1: Some Useful Log-Linear Approximations
Euler equation
We can rewrite the consumers Euler equation as
1 = 1
t
exp(i
t
oc
t+1
:
t+1
j) (44)
In a perfect foresight steady state with constant ination : and constant
growth we must have:
i = j +: +o
with the steady state real rate being given by
: = i :
= j +o
A rst-order Taylor expansion of exp(i
t
oc
t+1
:
t+1
j) around that
steady state yields:
exp(i
t
oc
t+1
:
t+1
j) 1 + (i
t
i) o(c
t+1
) (:
t+1
:)
= 1 +i
t
oc
t+1
:
t+1
j
which can be used in (44) to obtain, after some rearrangement of terms, the
log-linearized Euler equation
c
t
= 1
t
c
t+1

1
o
(i
t
1
t
:
t+1
j)
26
References
Chari, V.V., Lawrence J. Christiano, and Patrick J. Kehoe (1996): Op-
timality of the Friedman Rule in Economies with Distorting Taxes, Journal
of Monetary Economics 37, 203-223.
Cooley, Thomas F. and Gary D. Hansen (1989): Ination Tax in a Real
Business Cycle Model, American Economic Review 79, 733-748.
Correia, Isabel, and Pedro Teles (1999): The Optimal Ination Tax,
Review of Economic Dynamics, vol. 2, no.2 325-346.
Friedman, Milton (1969): The Optimum Quantity of Money and Other
Essays (Aldine Press, Chicago, IL).
King, Robert G., and Mark Watson (1995): Money, Prices, Interest
Rates, and the Business Cycle, Review of Economics and Statistics, vol 58,
no 1, 35-53.
Lucas, Robert E., (1996): Nobel Lecture: Monetary Neutrality, Journal
of Political Economy, vol. 104, no. 4, 661-682.
Lucas, Robert E. (1982): Interest Rates and Currency Prices in a Two-
Country World, Journal of Monetary Economics 10, 335-359.
Patinkin, Don (1987): "Neutrality of Money," in J. Eatwell, M. Milgate
and P. Newman, The New Palgrave: A Dictionary of Economics, W.W.
Norton, New York.
Sidrauski, Miguel (1967): "Ination and Economic Growth," Journal of
Political Economy, vol. 75, 796-816.
Svensson, L.E.O. (1985): "Money and Asset Prices in a Cash-in Advance
Economy," Journal of Political Economy, 93, 5, 919-944.
Walsh, Carl E. (2003): Monetary Theory and Policy, Second Edition,
MIT Press.
Woodford, Michael (1990): The Optimum Quantity of Money, in B.M.
Friedman and F.H. Hahn, eds. Handbook of Monetary Economics, vol II,
Elsevier-Science.
Woodford, Michael (2003): Interest and Prices: Foundations of a Theory
of Monetary Policy, Princeton University Press.
27
Exercises
1. Optimality Conditions under Non-Separable Leisure
Derive the log-linearized optimality conditions of the household problem
under the following specication of the period utility function with non-
separable leisure.
l (C
t
. `
t
) =
1
1 o
[C
t
(1 `
t
)
i
]
1o
2. Alternative Interest Rules for the Classical Economy
Consider the simple classical economy described in the text, in which the
following approximate equilibrium relationships must be satised

t
= 1
t

t+1

1
o
(i
t
1
t
:
t+1
j)
and
:
t
= i
t
1
t
:
t+1

= j +o 1
t

t+1

and where
t
and, hence, :
t
, are determined independently of monetary pol-
icy. Next you are asked to analyze, in turn, two alternative monetary policy
rules and their implications. When relevant, we assume that the money
market clearing condition takes the form
:
t
j
t
=
t
j i
t
+
n
t
where
n
t
is a stochastic money demand disturbance.
a) Strict Ination Targeting.
(i) Derive an interest rate rule that guarantees full stabilization of ina-
tion, i.e. :
t
= :

for all t where :

is an ination target assumed to be "close

to" zero (so that the log-linearized equilibrium conditions remain valid).
(ii) Determine the behavior of money growth that is consistent with the
strict ination targeting policy analyzed in (i).
28
(iii) Explain why a policy characterized by a constant rate of money
growth :
t
= : will generally not succeed in stabilizing ination in that
economy.
b) Price Level Targeting.
(i) Consider the interest rate rule
i
t
= j +c
j
(j
t
j

)
where c
j
0, and j

is a (constant) target for the (log) price level. Determine

the equilibrium behavior of the price level under this rule. (hint: you may
nd it useful to introduce a new variable j
t
= j
t
j

the deviation of the

price level from targetto ease some of the algebraic manipulations).
(ii) Consider instead the money targeting rule
:
t
= j

Determine the equilibrium behavior of the price level under this rule.
(iii) Show that the money targeting rule considered in (ii) can be com-
bined with the money market clearing condition and rewritten as a price-level
targeting rule of the form
i
t
= j + (j
t
j

) +n
t
where is a coecient and n
t
is a stochastic process to be determined.
(iv) Suppose that the central bank wants to minimize the volatility of the
price level. Discuss the advantages and disadvantages of the interest rate rule
in (i) versus the money targeting rule in (ii) in light of your ndings above.
3. Nonseparable Preferences and Money Superneutrality
Assume that the representative consumers period utility is given by:
l
_
C
t
.
`
t
1
t
. `
t
_
=
1
1 o
_
(1 0) C
1i
t
+0
_
`
t
1
t
_
1i
_1
1

`
1+,
t
1 +,
a) Derive the optimality conditions of the associated consumers problem.
b) Assume that the representative rm has access to a simple technology
1
t
= `
t
and that the monetary authority keeps a constant money growth

n
. Derive the economys steady state equilibrium under the assumption of
perfect competition.
29
c) Discuss the eects on ination and output of a permanent change in
the rate of money growth
n
, and relate it to the existing evidence.
4. Optimal Monetary Policy in a Classical Economy with an
Exact Equilibrium Representation
Consider a version of the classical economy with money in the utility func-
tion, where the representative consumer maximizes 1
0

1
t=0
,
t
l
_
C
t
.
AI
1I
. `
t
_
subject to the sequence of dynamic budget constraints
1
t
C
t
+`
t
+Q
t
1
t
_ `
t1
+1
t1
+\
t
`
t
1
t
Assume a period utility given by:
l
_
C
t
.
`
t
1
t
. `
t
_
= log C
t
+ log
`
t
1
t

`
1+,
t
1 +,
(45)
Suppose there is a representative perfectly competitive rm, producing
the single consumption good. The rm has access to the linear production
function 1
t
(i) =
t
`
t
(i), where productivity evolves according to:

t1
= (1 +
o
) exp
o
t

with
o
t
is an i.i.d. random process, normally distributed, with mean 0 and
variance o
2
o
.
The money supply varies exogenously according to the process
`
t
`
t1
= (1 +
n
) exp
n
t
(46)
where
n
t
is an i.i.d., normally distributed process with mean 0 and variance
o
2
n
. We assume that
n
t
evolves exogenously, outside the control of the
monetary authority (e.g., could reect shocks in the monetary multiplier that
prevent the monetary authority from fully controlling the money supply.).
Finally, we assume that all output is consumed, so that in equilibrium1
t
= C
t
for all t.
a) Derive the optimality conditions for the problem of households and
rms.
b) Determine the equilibrium levels of aggregate employment, output,
and ination (Hint: show that a constant velocity
1IYI
AI
= \ for all t is a
solution)
30
c) Discuss how utility depends on the two parameters describing monetary
policy,
n
and o
2
n
(recall that the nominal interest rate is constrained to be
non-negative, i.e., Q
t
_ 1 for all t). Show that the optimal policy must satisfy
the Friedman rule (i
t
= 0 all t) and discuss alternative ways of supporting
that rule in equilibrium.
5. A Shopping Time Model (based on Walsh (2003)).
Assume that the transactions technology is such that consuming C
t
re-
quires a quantity of shopping time `
c
t
= :
_
C
t
.
AI
1I
_
, where :
c
0 and :
n
_ 0.
Hence the amount of time diverted from leisure is given by `
t
+`
c
t
, where `
t
denotes hours of work. Let the original period utility be given by \ (C
t
. 1
t
)
where 1
t
= 1 `
t
`
c
t
denotes leisure.
a) Derive the condition determining the optimal allocation of time.
b) Derive the implied utility function in terms of consumption, hours and
real balances, and discuss its properties.
6. A Model with Cash and Credit Goods
Assume that the utility of the representative household is given by:
\ (C
1t
. C
2t
. `
t
) (47)
where C
1t
denotes consumption of a cash-good (i.e., a good that requires
cash in order to be purchased), C
2t
is consumption of a credit-good ( which
does not require cash), and `
t
is labor supply. For simplicity, let us assume
that the price of the two goods is identical and equal to 1
t
(this will be
the case if the production function of the representative rm is given by
1
1t
+ 1
2t
= `
t
and there is perfect competition). Purchases of cash-goods
have to be settled in cash, whereas credit goods can be nanced by issuing
one-period riskless nominal bonds.
The budget constraint is given by
1
t
(C
1t
+C
2t
) +Q
t
1
t
+`
t
= 1
t1
+`
t1
+\
t
`
t
+1
t
Finally, the CIA constraint is given by
1
t
C
1t
_ `
t1
+1
t
where, in equilibrium, 1
t
= `
t
, i.e. transfers to households correspond
to money tranfers made by the central bank, and which consumers take as
given. For simplicity we assume no uncertainty.
31
a) Derive the rst order conditions associated with the households prob-
lem
b) Note that whenever the CIA contraint is binding we can dene a
reduced form period utility:
l
_
C
t
.
`
t
1
t
. `
t
_
= \
_
`
t
1
t
. C
t

`
t
1
t
. `
t
_
where C
t
= C
1t
+C
2t
. Show that l
n
_ 0, given the optimality conditions de
rived in a).
32