15

Política monetaria en el marco intertemporal

En su mayor parte, hemos ignorado el papel del dinero y, por lo tanto, la política monetaria en nuestro estudio hasta ahora.
Esto se debe a que los principales problemas que hemos estado considerando, en particular, la idea de una toma de
decisiones óptima por parte de agentes representativos, que conducen al consumo de referencia (ocio y consumo),
condiciones de ahorro óptimo, no requieren una consideración explícita del dinero.

Nuestra falta de inclusión significativa del "dinero" también se debe en parte al hecho de que ha resultado un tanto
difícil construir un marco simple para los tres roles distintos que desempeña el "dinero" en la sociedad moderna.
Durante siglos (o tal vez milenios) se ha pensado que esos tres roles distintos son:

1) Medio de intercambio (que evita los problemas del intercambio de trueque, que es
casi imposible en economías desarrolladas)

2) Unidad de cuenta (por ejemplo, si gasta dólares estadounidenses en una tienda de EE. UU., Las etiquetas de precio serán

denotado en números de dólares estadounidenses, en lugar de, por ejemplo, en números de bolígrafos)

3) Almacén de valor (si se usara una pieza de fruta para hacer pagos, una pieza de fruta
obviamente se descompone muy rápidamente, en el mejor de los días o semanas, lo que implica que su valor se erosiona

rápidamente, mientras que un pedazo de papel fibroso y seguro que muestra el retrato de George Washington

probablemente durará décadas)

A pesar de la dificultad teórica de incorporar los "cómo" y los "porqués" de sociedades o países particulares o
épocas que se basan en una definición comúnmente entendida de "dinero", se trata casi por completo de dinero
alrededor del cual se divide la escuela de pensamiento RBC y Surge una nueva escuela de pensamiento keynesiana.
Para ilustrar la diferencia fundamental entre las dos teorías y, por lo tanto, la división fundamental en la teoría
macroeconómica moderna, necesitamos desarrollar un concepto de equilibrio del mercado monetario , que a su vez
requiere tanto demanda de dinero como oferta de dinero. Tomaremos un atajo y un enfoque ampliamente utilizado,
que es el función de dinero en la utilidad (MIU) marco para generar demanda de dinero.

En pocas palabras, el enfoque MIU simplemente inserta dinero (real), es decir, el poder adquisitivo de las unidades monetarias,

como un argumento para la función de utilidad del consumidor representativo.

218 Capítulo 15

Sin embargo, antes de llegar a la economía de las recomendaciones de política a corto y largo plazo que
surgen del marco de la MIU, repasemos los vínculos entre los mercados monetarios y los mercados de
bonos.

Mercado de bonos del gobierno

Ya debe estar familiarizado con los conceptos presentados brevemente en esta sección. Pero debido a que la conexión entre los

mercados monetarios y los mercados de bonos es crucial para comprender cómo funciona la política monetaria, parece

conveniente un breve resumen. Asumimos que los bonos son todos bonos del gobierno. 1 En tiempos "convencionales", la Reserva

Federal implementa sus decisiones de política a través de compras en el mercado abierto o ventas de bonos del gobierno de los

Estados Unidos. Además, asumimos que todos los bonos son bonos nominales, lo que significa que cada unidad de un bono paga

una cantidad fija de moneda.

Hablaremos de un mercado único de bonos del gobierno dentro de un país, a pesar de que existen muchos tipos diferentes
de bonos emitidos por los gobiernos, que se distinguen principalmente por su vencimiento y su valor nominal. Un enlace longitud
de madurez es el tiempo desde la emisión hasta que el valor total del bono se reembolsa al tenedor del bono, mientras que el
bono valor nominal es el valor total que se paga al vencimiento. Por ejemplo, el gobierno de los Estados Unidos emite letras del
Tesoro a un mes, letras del Tesoro a tres meses, letras del Tesoro a seis meses, notas del Tesoro a dos años, notas del Tesoro
a tres años, notas del Tesoro a cinco años y notas del Tesoro a diez años de varios valores faciales 2

Los bonos son simplemente préstamos , un punto que a menudo se malinterpreta. Independientemente de la duración del

vencimiento y el valor nominal de un bono, un bono del gobierno es simplemente un préstamo que el tenedor de un bono le otorga al

gobierno para pagarlo en una fecha posterior con intereses. El monto a pagar en la fecha especificada previamente es el valor nominal

del bono.

Debido a que el valor nominal no se reembolsa hasta un período de tiempo futuro, la cantidad que un tenedor de bonos

estaría dispuesto a pagar en el periodo actual para un enlace de valor nominal FV

dólares es algo menos que FV dolares La razón de esto es simplemente el descuento temporal de los valores futuros. Por
ejemplo, $ 100 dentro de un año probablemente valga menos de $ 100 para usted en este momento; en otras palabras,
es probable que esté dispuesto a aceptar algo menos de $ 100 en este momento en lugar de no recibir nada ahora y $
100 dentro de un año. . Debido al descuento de tiempo, el período t precio (indicado PAGS tuberculosis ) de un bono de
vencimiento a un período está relacionado con su valor nominal FV t + 1 y la tasa de interés nominal yo t , que representa el
componente de interés entre período t y periodo t + 1. La relación entre estos tres objetos es

1. También existen bonos corporativos (bonos emitidos por compañías) y, por lo tanto, mercados para bonos corporativos, que son mercados importantes. Sin
embargo, para propósitos de política monetaria estándar o “convencionales”, es bastante irrelevante qué tipos de bonos existen, por lo que ignoraremos los
mercados de bonos corporativos.

2. También hay muchas otras longitudes de vencimiento de los bonos del gobierno de los Estados Unidos.
Política monetaria en el marco intertemporal 219

Ptuberculosis
FV = + +1 .
1 yo
tt

La forma en que se escribe esta expresión hace que parezca que define el precio de un bono. Pero una interpretación
común de esta expresión es que, en cambio, define la tasa de interés nominal
yo t porque en cualquier momento se conoce el valor nominal de un bono y la cantidad que los demandantes de bonos están dispuestos

a pagar. Así conocimiento de PAGS tuberculosis y FV t + 1 puede considerarse como definitorio yo t .

Algebraicamente podemos enfatizar esta relación simplemente reorganizando la expresión anterior para aislar la
tasa de interés nominal, que es

FV +1 1 .
yo
t = -
P tb t

Estas dos ecuaciones son obviamente equivalentes entre sí.

También incluimos otros tres puntos de simplificación en aras de la facilidad del análisis posterior.

1) El valor nominal siempre es igual a FV = 1; por lo tanto, podemos soltar el subíndice de tiempo y el

tedioso de escribir FV .

2) En la práctica, hay dos tipos principales de bonos: bonos de cupón y bonos de cupón cero. Un bono de cupón es
aquel que realiza pagos de intereses (llamados "pagos de cupón") al tenedor del bono en momentos específicos
antes de un pago final del valor nominal en la fecha de vencimiento, mientras que un bono de cupón cero no ofrece
pagos intermedios antes del pago del valor nominal en la fecha de vencimiento. Por conveniencia, supondremos
que todos los bonos son bonos de cupón cero porque no importa ni para el análisis a corto ni a largo plazo.

3) Los reembolsos de bonos nominales siempre se pagan íntegramente a tiempo.

El último punto dice que el gobierno nunca incumplirá sus obligaciones de bonos nominales, lo cual, si nos acercamos al

gobierno de los Estados Unidos, es cierto.

Función de dinero en la utilidad (MIU) y demanda de dinero

A continuación traigamos el marco de horizonte infinito. La aplicación particular de los activos financieros cuando
consideramos por primera vez el marco de horizonte infinito fue la fijación de precios en el mercado de valores. Pero un
tema más amplio que surge del análisis anterior es sobre precio de activos en general, independientemente del tipo particular
de activo financiero en consideración. En el marco expandido de período infinito aquí, habrá tres tipos distintos de activos:
acciones, dinero y bonos. La Figura 15.1 muestra esta clase más rica de activos financieros y el momento de los eventos.
Matemáticamente, aumentamos el período representativo del consumidor. t función de utilidad para incluir ahora demanda de
dinero como argumento, en particular, la cantidad demandada de
 220 Capítulo 15

El individuo elige El individuo elige El individuo elige

óptimamente el óptimamente el óptimamente el
consumo real consumo real consumo real

Connecticut y elige ct + 1 y elige de manera ct + 2 y elige de manera

de manera óptima óptima la cartera de activos óptima la cartera de activos
la cartera de ( a + 1, B t + 1, M t + 1) para ( a + 2, B t + 2, M t + 2) para el
activos ( una t, B t, comienzo de período t + 2 comienzo del período t + 3
Recibe M t) Recibe Recibe

ingresos
re ingresos re ingresos re

nominales para principio de nominales nominales

Yt periodo t + 1 Yt+1 Yt+2

a-1 en B t a+1 a+2

Bt-1 MtD Bt+1 Bt+2
re re re
Mt-1 Mt+1 Mt+2

...
Inicio del
Período t Período t + 1 Período t + 2 Período t + 3
horizonte de
planificación
económica.

Recibe cartera de activos ( una t - 1, si Recibe una cartera de activos elegida Recibe una cartera de activos elegida óptimamente ( una
re re re
t - 1, METRO t - 1) óptimamente ( una t, si t, METRO t) t + 1, si t + 1, METRO t + 1)
incluidos los ingresos por incluidos los ingresos por dividendos incluidos los ingresos por dividendos
dividendos e intereses e intereses e intereses

NOTA: La planificación económica ocurre durante toda la vida útil.

Figura 15.1
Cronología de eventos en el marco monetario de período infinito

dinero, que es la esencia del modelo MIU. Supongamos que el período representativo del consumidor t la función de utilidad
es

••• ••• ,
uc Mt ,
PAGS
t tD

en el cual MPtD / / t es la demanda del consumidor por real saldos monetarios, es decir, para el pur-

persiguiendo el poder que una demanda nominal dada METRO tD tenencias proporciona. En general, el argumento de la demanda de

dinero real es un sustituto de los diversos roles que juega el dinero en diferentes períodos de tiempo, como se describió anteriormente.

Debido al factor de descuento subjetivo β ∈ ( , )

0 1 ( que se traslada de nuestro anterior

análisis del marco de período infinito), el utilidad con descuento de por vida desde la perspectiva del comienzo del
período t puede ser declarado como

••• ••• + ••• tD+ 1 ••• + •• tD+ 2 ••• + ••• tD+ 3 ••• +
uc Mt , uc Mt + 1 , β 2 uc•• Mt + 2 , β 3 uc Mt + 2 ,
tD
β ...
PAGS
t PAGS
t +1 PAGS
t+2 PAGS
t+3

∞
•• ts+ D •••;
∑β uc•• Mts+ ,
s
=
s=00
PAGS
ts+
Política monetaria en el marco intertemporal 221

la segunda línea escribe la función de utilidad de por vida de valor presente de manera compacta usando el operador de suma Σ .

Durante cada período de tiempo, se produce un "reequilibrio" óptimo entre los tres activos de la cartera. Esto se
describe en el período t restricción presupuestaria del consumidor,

PC
t t PBMS
+ tb t a YMB
+ tD + t t = +t tD− 1 + t−1 + ( SD
t +a t ) t −1,

in which, as in the basic asset-pricing framework, P t is the nominal price of consumption,

S t is the nominal price of a one unit of stock in period t , and D t is the nominal dividend per share in period t. Notice
the timing of the budget constraint: in period t , the consumer chooses nominal money holdings to carry into
period t + 1. 3 ( And see also figure 15.1 .) In turn is implied that the period t + 1 flow budget constraint is

Pt +c1 t +1 + Ptb+B1 Mt +S a
1 + tD+ 1 + t +1 t +1 = Yt +M1 +B S tD + t + ( t +1 + D ta+ 1 ) tt ,

the period t + 2 flow budget constraint is

Pt +c2 t +2 + Ptb+B2 Mt +S2 a

+ tD+ 2 + t +2 t +2 = Yt +M2 B
+ tD+ 1 + t +1 + ( S t + 2 + D t + 2 ) a t+1,

and so on, for periods t + 3, t + 4, t + 5, ... .

Optimal Choice

The sequential Lagrange problem stated in nominal terms is

••• ••• + ••• tD+ 1 ••• + •• tD+ 2 ••• + ••• tD+ 3 ••• +
ucM ucM β 2 u ••c M β 3u c M
tD
t, β t +1 , t+2 , t+2 , ...
Pt P t +1 Pt + 2 Pt + 3
+ λ t ⋅ [ Yt M
+B tD− 1 + t −1 + ( St D+ t ) aat − 1 − Pc
t tP−B M
tb t S a
− tD − t t ]
+ βλ t +1 ⋅ [ Yt +M1 +B S tD + t + ( t +1 + D ta+ 1 ) tt − Pt +c1 t +1 − Ptb+B1 Mt +S a
1 − tD+ 1 − t +1 t +1 ]
+ βλ ⋅ [ Yt +M2 B
+ + + ( S t + 2 + D ta+ 2 ) Pt +c2 Ptb+B2 Mt +S2 a− − ]
2
t +2 tD+ 1 tt + 1 t +1 − t +2 − tD+ 2 t +2 t+2

+ ... .,

which should look familiar to you — it is simply an extension of the sequential Lagrange function in our earlier
study of stock-market pricing. The first-order conditions with respect to c t , a t , B t , and M tD are, respectively,

••• tD ••• −
u1 c Mt , λ t Pt = 0 ,
Pt

− λ t St + βλ t +1 ( S t +1 + D t +1 ) = 0 ,

− λ t P tb + βλ t + 1 0=,

3. Mechanically, we know this because it is M tD , rather than M tD − 1 , that appears on the left-hand side of the budget constraint, and the
left-hand side represents “ outlays ” in period t .
222 Chapter 15

and

••• tD ••• 1
u 2c Mt , ⋅ − λ tβλ+ = t +1 0.
Pt P t

The first condition states the usual result that the marginal utility of consumption equals the Lagrange multiplier
(scaled by the price level P t ). The second first-order condition is our familiar stock-pricing equation. The third
first-order condition is that on bond holdings. In the fourth first-order condition, the 1/ P t term arises because
each individual can choose his/ her nominal money holdings, but takes the aggregate price level P t as given.
Because real, not nominal, money demand is the second argument of the utility function, the chain rule is
required, which generates the 1/ P t term.

These four first-order conditions taken together generate many rich insights about link- ages between bonds
markets and stock markets, between bond markets and monetary markets, and are the foundation of possible
ideological divides between whether or not changes in monetary policy affect short-run macroeconomic
conditions or long-run macro- economic conditions, or both. The following sections describe these insights in
turn. As you will see, we will go back and forth between “ macroeconomic theory ” and “ finance theory ” —
given the richness of the framework, the “ intersection ” between the two appar- ently different strands of
thought turns out to be a very clear intersection.

Pricing Kernel and Asset Prices

Delving back into a bit of finance theory, we can rearrange the first-order condition on bond holdings to get

βλ λt +1 .
P tb =
t

This already sheds a lot of light on the intersection of macro and finance! Recall from our study of
stock-pricing that βλ λ t +1 / was
t defined as the “ pricing kernel ” of the

economy.

Here it is! The price of a nominal bond equals the pricing kernel times one. 4 Or, stated from the opposite
perspective, the pricing kernel of an economy equals the price of a short-term riskless nominal bond. 5

4. The “ one ” here is simply the payoff of the nominal bond in our model — that is, we assumed that the face value, hence the payoff, of
the bond is FV = 1.

5. The “ riskless ” component was mentioned above, so we can think of these nominal bonds as US government nominal bonds.
Monetary Policy in the Intertemporal Framework 223

Note that P tb in the expression above is of the same general form as the stock-pricing equation we
encountered earlier — the price of an asset ( P tb ) depends on a pricing kernel and a future payoff (which is
simply FV = 1). Bonds are thus priced using the general type of asset-pricing equation we used to price stocks.
Continuing, the first-order condition on a t gives us

βλ λt 1+
St = ( S t +1 + D t +1 ),
t

which is our usual stock price condition. From what we now know, we can alternatively express the stock price
as

St P= S tb ( t +1 + D t + 1 ),

which explicitly shows a crucial linkage between bond prices and stock prices. Stock prices can thus be said to be
keyed (partially) off of bonds prices.
The big-picture, finance-theoretic, lesson to take away here is that asset-pricing equa- tions invariably have
the same general form, regardless of what specific type of asset is being considered. That general form is

Price of asset in current period

= (Pricing kernel) (Asset-a
× pppropriate future returns)

Fisher Equation

We can obtain the exact Fisher equation as an implication of optimal choices in this model, rather than as a
relationship which we so far have seemingly “ assumed ” to be true. To see this, begin with the last expression,
SPS t = tb ( t +1 + D t + 1 ). Divide this expression
through by the nominal price level P t ( which is distinct from the nominal price of a bond
P tb ) to get

St ( t +1 + D t + 1 ).
= tb
Pt P S Pt

Next, on the right-hand side, multiply and divide by P P t +1 / t +1 (which is, of course, just
multiplying by one, which is always a valid operation to conduct … ) to arrive at

St ( t +1 + D t +1 ) P tP+ 1 .
= tb ⋅
Pt P S Pt +1 t

The real price of stock purchased in period t is S P t / t (because it is divided by the current
price level), while the real payoff in period t + 1 of the stock purchased in period t + 1 is
( S t +1 + D tP
+1 ) / t +1 ( because it is divided by the future price level). The period-( t + 1) real
224 Chapter 15

payoff divided by the period- t real price is defined as the real return on the asset — that is, it is the object we have
heretofore been calling the real interest rate. 6

Letting r t denote the real interest rate between period t and period t + 1, we therefore have that

( S t +1 + D tP
+1 S
)/P
/ t +1
1 + =r t .
t t

With this, we can write the previous expression as

1 t +1
= +( 1 r t P) ⋅ .
P tb Pt

Only one more step remains in obtaining the exact Fisher relation from first principles. To finish the algebra, note
that, by construction and based on our definitions, 1 tb/ P= +
1 it
, and
Pt +P1 t/ =+
1 π t. + 1
The previous expression can thus be rewritten as

1 + it = +
(1 rt ) (1 + π t + 1 ) ,

which is the exact Fisher relation .

The economic intuition behind the Fisher equation is that it links the returns available on nominal assets
(nominal bonds) and the returns available on real assets (stocks). The linkage is through inflation; once the
nominal returns of bonds are adjusted by inflation, the returns on nominal bonds are exactly equal to the returns
on stocks, provided that financial markets are “ operating well. ”

This type of idea — that, once returns are converted into comparable units, they are equal- ized when
markets are behaving rationally — goes by the terminology of no-arbitrage in finance theory. No-arbitrage
relationships are key building blocks of more advanced finance theory; we defer richer consideration of issues
stemming from such relationships to a more advanced course on finance theory.

The exact Fisher equation emerges naturally in any model featuring both nominal assets and any type of real
asset, not just stocks. This brings us back full circle to our initial study of the two-period consumption – savings
model, in which we asserted the exact Fisher equation.

6. Stocks are considered to be “ real ” assets because their payoff is generally not fi xed in currency terms, whereas bonds are considered to
be “ nominal ” assets because their payoff is generally fi xed in currency terms (nonindexed bonds, at least).
Monetary Policy in the Intertemporal Framework 225

Nominal Interest Rates and Money Demand

Next let ’ s consider how the nominal interest rate i t affects macroeconomic conditions. So far we have not
exploited the information contained in the first-order conditions with respect to consumption or money
holdings, but now we finally will. Rewrite the first-order condition on nominal money holdings from above as

u 2c( Mt ,P tD t )
−=
λ t− βλ .t + 1
Pt

We know from the first-order condition on bond holdings that βλ t +1 = λ t Ptb ; inserting this
in the previous expression gives

u 2c( Mt ,P tD t )
−=
λ t− λ t Ptb .
Pt

Dividing through by λ t ,

u 2c( Mt ,P tD t )
− 1= − Ptb .
λ t Pt

Next we can use the first-order condition on consumption to replace the λ t Pt term on the
left-hand side, giving us

u 2c( Mt ,P utDc M t )
=−
1 Ptb .
P1 ( t , tD t )

The term on the left-hand side now is just the MRS between real money demand and consumption — that is,
it is the ratio of the marginal utility of (real) money to the marginal utility of consumption.

As for the right-hand side of this expression, because P tb = 1 1 / (+ i t ) , it can be stated as

u 2c( Mt ,P u tDc M t ) 1
=−
1 +.
P1 ( t , tD t ) 1 it

One final algebraic simplification gives us the consumption – money optimality condition

u 2c( Mt ,P u tDc M t ) i ti
=+,
P1 ( t , tD t ) 1 t

which states that the MRS between period- t real money and period- t consumption equals a function of the nominal
interest rate at the representative agent ’ s optimal choice. This opti- mality condition is completely analogous to
the consumption – leisure optimality condition
226 Chapter 15

Slope = – i / ( 1+ i )
Consumption

Optimal choice

M D/ P

Figure 15.2
Consumption – money demand optimality condition

and the consumption – savings optimality condition with which we have become familiar. The consumption –
money optimality condition states that when consumers are making their optimal choices, they choose
consumption and real money holdings in such a way as to equate their MRS between consumption and money
demand to a function of the nominal
interest rate.
Except for interpretation, the indifference curve – budget constraint diagram in figure
15.2 ought to look familiar by now.
Also as in, say, the consumption – leisure analysis we can translate the optimal choices for any particular
nominal interest rate i in the indifference curve – budget constraint diagram in figure 15.2 to a market diagram.
Figure 15.3 traces the quantity of money demanded as a function of its price i . To get from figure 15.2 to figure
15.3 , conduct the following thought experiment: successively lower the nominal interest rate i in figure 15.2 .
Along the money demand axis, it seems to be the case that M D successively increases. If this is true, this
generates the clear downward-sloping portion of the money demand func- tion in the money-market space of
figure 15.3 . Continuing the thought experiment, suppose i is extremely small — for example, i =

0.0125. It is apparent from the consumption – money optimality condition that the budget line is extremely
flat. If i were to hit exactly zero or turn strictly negative, the optimality condition would make no sense at all.
Because of the strict equality sign in the consumption – money optimality condition, it would imply that the
MRS between con- sumption and real money demand was negative, which violates (at least 99.9999 percent
of the time) basic microeconomic principles. Casual inspection of figure 15.2 , which has the “ usually shaped
indifference curves ” that are strictly convex to the origin, also visually confirms this.
Monetary Policy in the Intertemporal Framework 227

MS M s ( post–Great Recession)
i

MD

ZLB region

M/P

Figure 15.3
Money market whereby money demand ( M D ) increases as nominal interest rate i decreases. Nominal interest rates can never fall below
zero.

Hence emerges the “ zero lower bound ” (ZLB) restriction on nominal interest rates, which states exactly
what we concluded: nominal interest rates can never fall below zero. The ZLB restriction is clear in the
money-market space in figure 15.3 .

Functional Form for Preferences

To facilitate both the short-run and long-run monetary policy analysis, as well as to formalize what was
seemingly casually concluded immediately above, let ’ s specialize our utility function to

••• ••• = M tD
ucM
tD
t, ln c t + ln .
Pt Pt

This functional form displays strictly convex to the origin indifference curves in the indif- ference curve space of
figure 15.2 . And none of the policy conclusions we reach below depend on this particular functional form, but it
allows for ease of algebraic manipulations to come.

The marginal utility functions associated with this utility form are obviously 7

tD
••• ••• = 1
u1c Mt ,
Pt ct

7. Verify this for yourself. Also note well that there is no use of the chain rule here — the chain rule was already used to obtain the
consumption – money optimality condition, regardless of the precise utility functional form.
228 Chapter 15

and

••• tD ••• = 1
u 2c Mt , .
Pt M tDP / t

This means that the period- t consumption – money optimality condition can be written as

ct M i ti
=+ ,
P tD / t 1 t

which is diagrammable in figure 15.2 . Or, recasting it in money-market space,

M tD 1 i ti •••
= + ••• ct,
Pt t

which is diagrammable in figure 15.3 .

With all of this now in place, we are ready to examine two long-standing questions in monetary analysis, one
a short-run issue, the other a long-run issue. Both the long-run and short-run issues center on the question of
whether or not monetary policy is neutral. Monetary policy is said to be neutral with respect to the economy if
changes in mone- tary policy do not affect real aggregate outcomes in the economy. Symmetrically, monetary
policy is said to be nonneutral with respect to the economy if changes in monetary policy

do affect real aggregate outcomes in the economy.

Monetary Policy I: The Short Run

To consider neutrality and nonneutrality in the short run, we first have to define more rigor- ously what the short
run is, and then look at the ordering of events within that short run. A natural interpretation of “ short run ” in our
multiple-period model is one period of time,
which we label “ period t . ”
What about the “ ordering of events ” within that one period of time? Figure 15.4 zooms in on period t and
diagrams one example. The two main aspects around which the short-run neutrality debate revolves are whether
or not an “ unexpected change ” in Federal Reserve monetary has occurred and the fact that money markets
almost universally clear quickly. 8

Figure 15.4 contains both of these aspects.

Suppose that a monetary policy “ shock ” has occurred. For the sake of concreteness, suppose that the
money supply in period t , M tS , unexpectedly turns out to be larger than markets had earlier (earlier within the
short-run period t , to be more precise, and as figure
15.4 shows) anticipated. The motivation is likely meant to “ boost aggregate demand. ”

8. The vastly liquid and continuously operating money market fund (which directly corresponds to the money market in our analysis) had
failed to clear only three times in their 37-year history up until the 2008 fi nancial crisis.
 Monetary Policy in the Intertemporal Framework 229

Consumers make Federal Reserve Regardless of whether or not

optimal choices of c meets and money shock occurred,
determines actual M S period- t
t
and nominal t consumption–money
money demand M D, of economy. optimality condition
t
must still hold.
taking as given If different from
some expected M S, a t Thus M t = M D = Mt S t
expectation “money shock” has must occur to ensure
a t –1 at
about M S. t occurred. monetary market equilibrium.
B t –1 Bt
(equilibrium) M t –1 (equilibrium) M t

Period t

Question: if monetary policy shock occurred,

what prices or quantities adjust during period t to
ensure consumption–money optimality condition
holds?

Figure 15.4
Timing of events within a given period. Second half of timeline emphasizes that money-market equilibrium is achieved in every period
time of time.

Regardless of policy motivation, by definition of money market equilibrium,

M t M tD M tS
= =
Pt Pt Pt

must be true, regardless of whether a policy shock has occurred. 9 In turn the (equilibrium) money demand
function (based on the particular functional form described above) requires that

M t 1 i ti •••
= + ••• ct.
Pt t

For this expression to hold with equality, a nominal money supply shock requires that P t
adjusts or c t adjusts or i t adjusts, or any combination thereof. Notice that whatever it is that adjusts to maintain
money-market equilibrium, it all occurs in the short run. That is, all of these prices and quantities are dated
period t .

9. Note that in equilibrium, we drop the S and D superscripts because the very defi nition of equilibrium is that supply = demand.
230 Chapter 15

To simplify the analysis, and because it has been empirically true in the United States from late 2008 until at
least 2015, suppose the short-term nominal rate is i t = 0. In terms of figure 15.3 , the economy has hit the zero
lower bound. The unanticipated monetary stimu- lus then has to affect nominal prices P t in the short run or
consumption quantity demand c t
in the short run, or both.
Let ’ s paint the two polar extreme cases, first the strict Keynesian sticky price view, and then the strict RBC
flexible price view. In the strict Keynesian case, nominal prices do not adjust in the short run. Thus, feeling flush
with unexpectedly large quantities of cash, con- sumers will raise their demand for goods in the short run. “
Monetary stimulus ” has suc- ceeded in that real quantity demanded has increased, at least in the short run. Monetary
policy is thus nonneutral in the Keynesian school of thought.

In the strict flexible-price RBC case, nominal prices adjust very quickly. Provided that “ very quickly ” is
shorter than the length of period t , the unexpected increase in consump- tion demand is quickly neutralized —
the terminology is not coincidental — by a rapid increase in P t . In this case, all “ monetary stimulus ” has
created is a burst of inflation. Mon- etary policy is thus neutral in the RBC school of thought.

Figure 15.5 illustrates these two extreme cases from the perspective of the period-t goods market. The
aggregate goods demand function necessarily shifts outwards due to an unexpected increase in the nominal
supply — the (equilibrium) money demand expression

M t 1 i ti •••
= + ••• ct
Pt t

shows this.

Pt
AS ( in RBC view)

AS ( in Keynesian view)
P* t

AD post-policy shock

AD pre-policy shock

GDP t

Figure 15.5
Following a positive shock to monetary policy, aggregate demand shifts outward
Monetary Policy in the Intertemporal Framework 231

Whether or not this leads to a temporary increase in equilibrium GDP depends entirely on the shape of the “
short-run aggregate supply function. ” For macro-relevant lengths of period t ( which is typically quarterly because
GDP accounts are compiled and referenced for the January – March quarter, April – June quarter, the July –
September quarter, and the October – December quarter), data suggest that an empirically relevant slope of
aggregate supply is strictly positive — somewhere between the extremely flat Keynesian AS function and the
extremely vertical RBC-style AS function. For modern macroeconomists, this begs the question: What are the
microeconomic reasons for “ partial ” nominal price stickiness? We sidestep this issue for now, and return to it
later in the more advanced New Keynes- ian theory in chapters 22 through 24. For the remainder of the analysis
here, we consider the effects of monetary policy in the long run.

Monetary Policy II: The Long Run

We have been considering an infinite-period framework. As we were able to do in our earlier, simpler,
infinite-period model absent money, it is useful to consider steady states. In our explicitly monetary model
here, considering the steady state will starkly reveal a relationship important to all of monetary theory, a
relationship between inflation and the rate of growth of the nominal money supply of the economy. This way
of thinking about inflation commonly goes under the name of “ monetarism ” or the “ quantity theory of
money. ” 10

Let ’ s continue to use the utility function

••• ••• = M tD
ucM
tD
t, ln c t + ln ,
Pt Pt

but, just as in the consideration of short-run effects above, none of the conclusions we reach depend on this
particular functional form.
For the sake of not having to turn back several pages, recall that the money demand expression is

M tD 1 i ti •••
= + ••• ct.
Pt t

10. One of the most-often quoted sayings by the late Milton Friedman, the 1976 Nobel laureate in economics, is that “ infl ation is everywhere
and always a monetary phenomenon, ” which has commonly been interpreted to mean that it is the actions of the central bank of an economy
(in particular, how the central bank manages the money supply of an economy) that alone determine the rate of infl ation in the economy. As
we are about to see, more precisely, only in the steady state (i.e., in the “ long run ” or “ on average ” ) is infl ation a purely monetary
phenomenon.
232 Chapter 15

A completely analogous condition holds in period t − 1 (or period t − 2, or period t + 1, etc.):

M tD− 1 1 i ti − 1 •••
= + ••• c t −1.
Pt − 1 t−1

Let ’ s combine these time- t and time-( t − 1) versions of the consumption – money optimality condition by dividing
the former by the latter; doing so gives us

M tDP M
/ t ct •••1 + i ti ••• + •••i ti − 1 •••
= .
P tD− 1 / t−1 c t −1 t 1 t −1

Reorganizing terms a bit, we have

M t P tP− 1 ct •••1 + i ti ••• + •••i ti − 1 •••

= .
M t −1 t c t −1 t 1 t −1

From our usual definition of inflation, we have

P tP− 1= + 11π.

t t

Now define the growth rate of nominal money in an analogous way. Specifically, define

M t
μt = −1
M t− 1

as the growth rate μ of the nominal money stock of the economy between period t − 1 and period t. As an
example, if the nominal money supply does not change between period t − 1 and period t , the nominal money
growth rate is μ t = 0 .
Using our definitions of the inflation rate and the money growth rate, we can rewrite the money growth rate
equation as

1 + μ πt ct •••1 + i ti ••• + •••i ti − 1 •••

.
1 += t c t −1 t 1 t −1

Now let ’ s consider the steady state. Recall our definition of a steady state as a state of the economy when all
real variables settle down to constant values over time, but nominal variables need not. Let ’ s make the latter
part of this concept a bit more precise than we did earlier: it is only nominal level variables that need not settle
down to constant values in the long run. For example, the nominal price level of the economy need not settle
down to a constant value in the long run. The same is true of the level of the nominal money supply of the
economy. Yet nominal growth rate variables do settle down to constant values in the long run. That is, the growth
rate of a nominal variable is considered to be a real variable. Moreover
Monetary Policy in the Intertemporal Framework 233

interest rates, regardless of real or nominal, also settle down to constant values in the steady state.

Applying this more precise concept of a steady state to the previous expression above, we see that all of the
variables contained in it settle down to constant values in the long- run: that is, c
t − 1= = ct c , i t − 1= = i t i , μ t − 1= = μ tμ , and π t − 1= = π tπ . Imposing these
steady-state values and canceling terms, we obtain

1 + μ
1,
1 + =π

or more simply,

π μ= . (1)

Expression (1) captures the essence of the monetarist school of thought within macro- economics, stating
that (in the long run — i.e., in the steady state) the inflation rate of the economy is governed by the rate of growth
of the money supply.
The rate of growth of the money supply is controlled by an economy ’ s central bank because it is ultimately
the economy ’ s sole (legal) supplier of money. The higher is the growth rate of money in an economy, the
higher is (in the long run) the economy ’ s inflation rate. Hence monetary policy is nonneutral in the long
run. This long-run monetarist per- spective is universally accepted by modern-day RBC oriented
macroeconomists and modern-day New Keynesian oriented macroeconomists — both camps acknowledge
that in the long run, nominal prices do adjust. 11 The neutrality debate is entirely about the short run.

We will next examine even further the causes and consequences of monetary policy in both the short run and
the long run, with a special focus on the interactions between mon- etary policy and fiscal policy. This monetarist
linkage will be in the background of many of the causes and effects we discuss there.

Chapter 15 Problem Set Questions

1. Deriving a money demand function. Denote by φ( , ) ct i t the real money demand

function. Here you will generate particular functional forms for φ ⋅ ( ) using the MIU
model we have studied.

11. This view apparently would not have been shared by Keynes himself, to whom the famous phrase “ In the long run, we ’ re all dead! ”
is attributed.
234 Chapter 15

In an MIU model, recall that the consumption – money optimality condition can be expressed as

u mt i ti
=+ ,
u ct 1 t

where u mt denotes marginal utility with respect to real money balances (what was named u 2 in our look at
the MIU model) and u ct denotes marginal utility with respect to consumption (what was named u 1 in our
look at the MIU model). In each of the fol- lowing, you are given a utility function and its associated
marginal utility functions. For each case, construct the consumption – money optimality condition and
use it to generate the function φ ⋅
( ) . In each case, your money demand function should end up
being an increasing function of c t and a decreasing function of i t . ( Note: Be careful to make the distinction
between real money holdings and nominal money holdings. The marginal utility function u mt is marginal
utility with respect to real money holdings.)

••• t ••• = •••M t ••• .

a. u c Mt , ln c t + ln
Pt Pt

••• t ••• = M t
b. u c Mt , 2 ct +2 .
Pt Pt

1−. σ
••• t ••• = t •••
c. u c Mt , c tσM⋅ •••
Pt Pt

2. M1 money and M2 money. Consider an extended version of the infinite-period MIU

framework. In addition to stocks and nominal bonds, suppose that there are two forms of money: M1
and M2. M1 money (which we will denote by M t 1 ) and M2 money (which we will denote by M t 2 ), both of
which directly affect the representative con- sumer ’ s utility. The period- t utility function is assumed to
be

1 2 1 2
••• t M Pt ••• = M t M t
ucM
t , , ln c t + ln + κ ln ,
Pr t Pt Pt

which has three arguments. The Greek lowercase letter “ kappa ” ( κ ) in the utility func- tion is a number
between zero and one, 0 ≤ κ≤ 1 , over which the representative con-
sumer has no control. The period- t budget constraint of the consumer is

Pc
t tM+M Bt S
1
+ a Yt M
2
+ t + = t+ t t
1
t−1 + (1 + i tMM
−1 )
2
t−1 + ( 1 + i tB
−1 ) t −1 + ( St D+ a t ) t− 1 ,

where i t denotes the nominal interest rate on bonds held between period t and t + 1 ( and hence i t − 1 on
bonds held between t − 1 and t ) and i tM denotes the nominal interest rate on M2 money held between period
t and t + 1 ( and hence i tM − 1 on M2 money held
Monetary Policy in the Intertemporal Framework 235

between t − 1 and t ). Thus note that M2 money potentially pays interest, in contrast to M1 money, which pays
zero interest.

As always, assume the representative consumer maximizes lifetime utility by opti- mally choosing
consumption and assets (i.e., in this case choosing all four assets optimally).

a. Using the functional form for utility given in this problem, what is the marginal
rate of substitution between real M1 money and real M2 money? (Hint: You do not need to solve a
Lagrangian to answer this — all that is required is using the utility function.) Explain the important
steps in your argument.

b. A sudden, unexplained change in the value of κ would be interpretable as which

of the following: a preference shock, a technology shock, or a monetary policy shock? Briefly
explain. Let φ 2 ( , ,
ct i i t tM )
denote the real money demand function for M2 money. Note the three arguments to the
function φ 2 (.) . Using the first-order conditions of the representative consumer ’ s Lagrangian, generate the
function φ 2 ( , , c t i i t tM ()i.e., solve for real M2
money demand as a function of c t , i t , and i tM ). Briefly explain (economically) why i tM
appears in this money demand function. (Note: You must determine yourself which are the relevant
first-order conditions needed to create this money demand function.)

3. The yield curve. An important indicator of markets ’ beliefs/expectations about the

future path of the macroeconomy is the “ yield curve, ” which, simply put, describes the relationship
between the maturity length of a particular bond (recall that bonds come in various maturity lengths) and
the per-year interest rate on that bond. A bond ’ s “ yield ” is alternative terminology for its interest rate. A
sample yield curve is shown in the following diagram:

February 9, 2005 (USD)

5.00000

4.50000

4.00000
Yield (%)

3.50000

3.00000

2.50000

2.00000
0.00 5.00 10.00 15.00 20.00 25.00 30.00

Borrowing period (years) Yield curve as on

236 Chapter 15

This diagram plots the yield curve for US Treasury bonds that existed in markets on February 9, 2005: as it
shows, a 5-year Treasury bond on that date carried an interest rate of about 4 percent, a 10-year Treasury
bond on that date carried an interest rate of about 4.4 percent, and a 30-year Treasury bond on that date
carried an interest rate of about 4.52 percent.

Recall from our study of bond markets that prices of bonds and nominal interest rates on bonds are
negatively related to each other. The yield curve is typically discussed in terms of nominal interest rates (as in
the graph above). However, because of the inverse relationship between interest rates on bonds and prices
of bonds, the yield curve could equivalently be discussed in terms of the prices of bonds.

In this problem, you will use an enriched version of our infinite-period monetary framework from this
chapter to study how the yield curve is determined. Specifically, rather than assuming the representative
consumer has only one type of bond (a one- period bond) he can purchase, we will assume the
representative consumer has several types of bonds he can purchase — a one-period bond, a two-period
bond, and, in the later parts of the problem, a three-period bond.

Let ’ s start just with two-period bonds. We will model the two-period bond in the sim- plest possible way:
in period t , the consumer purchases B tTWO units of two-period bonds, each of which has a market price P tb
TWO ,
and a face value of one (i.e., when the
two-period bond pays off, it pays back one dollar). The defining feature of a two- period bond is that it
pays back its face value two periods after purchase (hence the term “ two-period bond ” ). Mathematically,
then, suppose that the representative con- sumer has a lifetime utility function starting from period t :

•••M t ••• + •••M t + 1 ••• + •••M tt + 2 •••

ln c t + ln β ln c t + 1 + β ln β
2
ln c t + 2 + β 2 ln
Pt P t +1 Pt + 2

•••M t + 3 •••
+ β
3
ln c t + 3 + β3 ln ...,
Pt + 3

and his period- t budget constraint is given by

Pc
t tP+B P
tb t +
,
B
tb TWO tTWO + M tS+a=Yt+Mt B Bt t−1 + t−1 + −2
tTWO + ( St D+ tt )at− 1 .

(Based on this, you should know what the period t + 1 and period t + 2 and period t +
3, etc., budget constraints look like.) This budget constraint is identical to that dis- cussed in the chapter,
except for the terms regarding two-period bonds. Note carefully the timing on the right-hand side — in
accordance with the defining feature of a two- period bond, in period t , it is B tTWO

−2 that pays back its face value. The rest of the notation
is just as discussed in the chapter, including the fact that the subjective discount factor (i.e., the measure of
impatience) is β < 1.
Monetary Policy in the Intertemporal Framework 237

a. Qualitatively represent the yield curve shown in the diagram above in terms of
prices of bonds rather than interest rates on bonds. That is, with the same maturity lengths on the
horizontal axis, plot (qualitatively) on the vertical axis the prices associated with these bonds.

b. Based on the utility function and budget constraint given above, set up an appro-
priate Lagrangian in order to derive the representative consumer ’ s first-order con- ditions with respect
to both B t and B tTWO ( as usual, the analysis is being conducted from the perspective of the very
beginning of period t ). Define any auxiliary nota- tion that you need in order to conduct your analysis.

c. Using the two first-order conditions you obtained in part b, construct a relation-
ship between the price of a two-period bond and the price of a one-period bond. Your final
relationship should be of the form P tb TWO ,
= ,...
and on the right-hand-
side of this expression should appear (potentially among other things), P tb . ( Hint: In order to get P tb into
this expression, you may have to multiply and/or divide your first-order conditions by appropriately
chosen variables.)

d. Suppose that the optimal nominal expenditure on consumption ( Pc ) is equal to

1 in every period. Using this fact, is the price of a two-period bond greater than, smaller than, or
equal to the price of a one-year bond? If it is impossible to tell, explain why; if you can tell, be as
precise as you can be about the relationship between the prices of the two bonds. (Hint: You may
need to invoke the con- sumer ’ s first-order condition on consumption.)

e. Now suppose there is also a three-period bond. A three-period bond purchased in

any given period does not repay its face value (also assumed to be 1) until three
periods after it is purchased. The period- t budget constraint, now including one-, two-, and
three-period bonds, is given by

Pc
t tP+B P
tb t +
,
B
tb TWO tTWO
,
+ P tb THREE tTHREE
B + M tS+a t t

Yt M B
=+ t −1 + tt − 1 + B tTWO
− 2 + B tTHREE
−3 + ( St D+ a t ) t −1 ,

where B tTHREE is the quantity of three-period bonds purchased in period t and

,
P tb THREE its associated price. Following the same logical steps as in parts b, c, and d above (and
continuing to assume that nominal expenditure on consumption ( Pc )
is equal to one in period every period), is the price of a three-year bond greater than, smaller than, or
equal to the price of a two-year bond? If it is impossible to tell, explain why; if you can tell, be as
precise as you can be about the relationship between the prices of the two bonds. (Hint: You may
need to invoke the consum- er ’ s first-order condition on consumption).

f. Suppose β = 0.95. Using your conclusions from parts d and e, plot a yield curve in terms of bond prices
(obviously, you can plot only three different maturity lengths here).
238 Chapter 15

g. What is the single most important reason (economically, that is) for the shape
of the yield curve you found in part f? (This requires only a brief, qualitative/ conceptual
response.)

4. The cash-in-advance (CIA) framework. A popular alternative to the money in the

utility function (MIU) framework is one in which money holdings directly facilitate transactions — that is,

provide a medium of exchange role, one of the basic functions of “ money ” that the MIU framework

captures in only a shortcut form. Suppose that in any given time period, there are two types of consumption

goods: cash goods and credit goods. “ Cash goods, ” denoted by c t

1, are goods whose purchase in

period t requires money, while “ credit goods, ” denoted by c t 2, are goods whose pur-
chase in period t does not require money (i.e., they can be bought “ on credit ” ). The market nominal price
of each type of good is identical, P t .

The representative consumer consumes both cash and credit goods. Specifically, suppose the
period- t utility function is u c c n ( 1t , 2t , t ) , with n t denoting the individual ’ s
labor during period t . ( Suppose that total hours available in any given time period is 168 hours per week
and that the only possible uses of time are labor or leisure.) The consumer ’ s period- t budget constraint is

Pc
t 1Pc
t +M tP 2B Pw nt M
t + + =Btb t t t t + t− + 1 t − 1,

Income is earned from labor supply (with w t denoting the market determined real wage in period t , which is
taken as given by the individual), and for simplicity, suppose that there are no stock markets (hence
one-period riskless bonds and money markets are the only two asset markets). The consumer ’ s bond and
money holdings at the start of period t are B t - 1 and M t - 1 , and at the end of period t are B t and M t . The individual
’ s budget constraints for period t + 1, t + 2, … , are identical to the above, with the time subscripts
appropriately updated. As always, suppose that the representative consumer ’ s subjective discount factor
between any pair of consecutive time periods is β ∈( , )
01.

In addition to the budget constraint, in each time period the representative consumer also has a
cash-in-advance constraint,

Pc M
t 1 t= t.

The cash-in-advance (CIA) constraint captures the idea that in order to purchase some goods, a certain
amount of money (or more generally, “ liquidity ” such as checkable deposits) has to be held. In
principle, the CIA constraint is an inequality constraint (specifically, Pc M
t ≤
1t t ), but in analyzing this problem, you are to assume that it
always holds with equality, as written above. From the standpoint of the analysis, you will conduct,
because the CIA is technically an inequality constraint, you may not
substitute the CIA constraint into the budget constraint.
Monetary Policy in the Intertemporal Framework 239

The consumer ’ s budget constraints and CIA constraints for period t + 1, t + 2, ... , are identical to the above,
with the time subscripts appropriately updated.

a. Set up an appropriate sequential Lagrangian from the perspective of the beginning

of period t . Define any new notation you introduce.

b. Based on the Lagrangian constructed in part a, derive the first-order conditions

(FOCs) with respect to period- t choices, c t 1, ct 2, n t , B t , and M t . Define any new
notation you introduce.

c. Using the FOCs from above, derive the cash good/credit good optimality condi-
tion, which should have a final form u c c n1 (u
1 ct c
, n2 t , t )( 21 t , 2t , t ) (=) f i t . Note that
f ( i t ) is a function that depends only on the nominal interest rate. You are to deter- mine f ( i t ) as part of
this problem; there should be no other variables or parameters at all on the right-hand side of the cash
good/credit good optimality condition you derive. Show clearly the important steps in the algebra.

d. Suppose now that the utility function is u c c n ( 1t , 2t , t ) ln= c1t + ln c 2 t + v (n) t , in

which v ( n t ) is some unspecified function of labor. Taking into account the fact that the CIA
constraint holds with equality at the optimal choice, derive, based on this utility function and your
work above, the real money demand function,

M t
= ...
Pt

e. Recall that in the MIU framework, the consumption – money optimality condition
was

Marginal utility of real money holdings i ti

=+ .
Marginal utility of consumption 1 t

Compare, in terms of economics, this optimality condition to the optimality condition you obtained in part c
above for the CIA framework by briefly commenting on the similarities and differences between the
results predicted by the two frameworks. (Note: This does not mean restate in words the mathematics;
rather, offer two or three thoughts/critiques/etc., on how and why the two frameworks do or not capture
the same ideas, how and why the two frameworks perhaps are or are not essentially identi- cal to each
other, etc.)

5. Monetary policy in the MIU model. In this question you will analyze, using indiffer-
ence curve – budget constraint diagrams, the implications of alternative nominal interest rates on the

representative consumer ’ s choices of consumption and real money balances. Recall that, with an

instantaneous utility function u c M P ( ,t t / t ) (where, as usual, c t

denotes consumption and M P t / t denotes real money balances), the consumption-
money optimality condition (which we derived in this chapter) can be expressed as
240 Chapter 15

u mct (M, t P u ct /M t ) i ti
=+ ,
Pc t ( , t t / t ) 1 t

where, again as usual, i t is the nominal interest rate, u c (.) denotes the marginal utility of consumption, and u
m(.) denotes the marginal utility of real money balances.

a. Suppose that the central bank is considering setting one of two (and only two)
> i t . On the indifference map
1
positive nominal interest rates: i t 1 and i t 2 , with i
2
t

shown above, qualitatively (and clearly) sketch relevant budget lines and show the consumer ’ s
optimal choices of consumption and real money under the two alternative policies. Note in the
diagram the point on the vertical axis marked “ fixed ” — this denotes a point that must lie on every
budget constraint. Clearly label your diagram, including the slopes of the budget lines.
Consumption

FIXED

Real money

b. You are a policy adviser to the central bank, and any advice you give is based
only on the goal of maximizing the utility of the representative consumer. The central bank asks
you to help it choose between the two nominal interest rates i t 1

and i t 2 ( and only these two). Referring to your work in the diagram above, which nominal interest rate
would you recommend implementing? Briefly explain.

c. Suppose instead that the central bank is willing to consider setting any nominal
interest rate, not just either i t 1 or i t 2 . What would your policy recommendation be? Briefly justify your
recommendation, and also in the diagram (of part a) sketch and clearly label a new budget line
consistent with your policy recommendation.