Chapter 13

IS-LM model with rational

expectations

13.1 Introduction
In this and the following chapters the focus is shifted from long-run macro-
economics to short-run macroeconomics. The long run models considered
in previous chapters concentrated on mechanisms that are important for the
economic development over a time horizon above, say, 10 years. The basic
view is that for the economic trend, the supply side such as capital accumu-
lation, population growth and technical progress matter a lot for output and
employment, and monetary factors have only little influence. Focusing on the
trend, long-run analysis is often performed as if no further disturbances in the
economic structure and environment occurred. In the next four chapters we
shall deal with the short run, that is, mechanisms governing the adjustment
processes with a time horizon less than, say, three years.1 Here, the demand
side, monetary factors, nominal rigidities and expectational errors matter a
lot. This will prepare us for the final part of this text, namely the chapters
dealing with the medium run. These chapters are intended to bridge the gap
between the long run and the short run. The purpose of medium-run theory
is to explain the fluctuations around the trend (“business cycles”). Here the
1
These number-of-years figures should not be understood as more than a rough indi-
cation. Their appropriateness will certainly depend on the specific problem at hand.

425
426 CHAPTER 13. IS-LM MODEL WITH RATIONAL EXPECTATIONS

dynamic interaction between demand and supply factors and the resulting
adjustment in relative prices play a central role.
As a first approach to short-run theory this chapter presents a simple
dynamic version of the basic Keynesian model, the IS-LM model. The IS-
LM model was constructed by Hicks (1937) in an attempt to summarize the
analytical content of Keynes’ General Theory of Employment, Interest and
Money (1936). Although there has been plenty of debate about whether it
succeeded to do so (see, e.g., Leijonhufvud 1968), the IS-LM model has re-
mained a cornerstone of mainstream short-run macroeconomics. The demand
side of the large macroeconometric models which governments, financial in-
stitutions and trade organizations use to predict macroeconomic evolution in
the near future is built on the IS-LM model. Empirically the IS-LM model
does quite well (see, e.g., Gali 1992, Rudebusch and Svensson 1998). At
the theoretical level the IS-LM model has been criticized for being ad hoc,
i.e., not derived from first principles (the behaviour of households and firms,
given preferences, technology and market structure). In recent years, how-
ever, micro-foundations of (different elaborate versions of) the IS-LM model
have been provided (see McCallum and Nelson, 1999, Sims 2000, Walsh 2003
and Woodford 2003).
The IS-LM framework is based on the simplifying idea that for short-run
analysis of effects of demand shocks it is legitimate, as a first approximation,
to treat the nominal price level as an exogenous constant. This idea rests
on the empirical observation that in the short run, firms − at least in the
manufacturing and service industries − usually let output do the adjustment
to changes in demand while keeping prices unchanged. The firms seem able
to do that because there is typically enough excess capacity in industrialized
societies. Indeed, three of the most salient short-run features of macroeco-
nomic time series in the industrialized market economies are the following
(Blanchard and Fischer, 1989, and Nishimura, 1992):

1) Shifts in aggregate demand (induced, e.g., by changes in the money

supply) are largely accommodated by changes in quantities rather than
by changes in nominal prices (nominal price rigidity); at least this holds
for manufacturing and services.
13.1. Introduction 427

2) Large movements in quantities are often associated with little or no

movement in relative prices, including real wages (real price rigidity).

3) By contrast, prices are sensitive to changes in cost (cost-based prices).

The so-called New Keynesian approach consists of different attempts at

explaining these “stylized facts” and incorporating them in elaborate ver-
sions of the IS-LM model. This approach will be the topic in later chapters.
One “old Keynesian” approach, to which Keynes’ General Theory (1936) be-
longs, is the aggregate supply-aggregate demand (AS-AD) framework, where
the level of nominal wages is an exogenous constant in the short-run, while
the price level is endogenous and cost-determined. The demand side of the
AS-AD framework constitutes the IS-LMmodel. The model is often, how-
ever, interpreted as an independent model in its own right, the simplest “old
Keynesian” model, based on the simplifying assumption that nominal wages
and prices are fixed in the short-run. The model then explores the impli-
cations of this. Thereby, output (and employment) becomes demand-driven
− quantities are the equilibrating factors, not relative prices which do not
move. This is a complete upside-down compared to long-run theory, where
output and employment are regarded as mainly supply driven − with relative
prices as the equilibrating factors. The important relative prices in macro-
economics are the level of real wages at labour markets and the level of real
rates of interest at the asset markets.
The next section presents the traditional static version of the IS-LM
model (with exogenous expectations). This serves as an introduction to
the succeeding section which analyses a modernized, dynamic version of the
IS-LM model with endogenous forward-looking expectations. In the next
chapter the framework is extended to the open economy. We consider the
Mundell-Fleming model for a small open economy and the Blanchard-Fischer
version of Dornbusch’s overshooting model. Later chapters discuss micro-
foundations. For now we take nominal rigidities as given. The purpose is to
make the internal logic of the IS-LM framework under rational expectations
transparent.
428 CHAPTER 13. IS-LM MODEL WITH RATIONAL EXPECTATIONS

13.2 The traditional IS-LM model

The traditional, static IS-LM model is well-known from elementary macro-
economic textbooks. Here we provide a refresher and show the link to the
more advanced, dynamic version.
We consider a closed economy with a private sector, a government sector
and a central bank. Output demand is given as

Y d = C(Y p , r) + I(Y, r, K) + G + εD , (13.1)

0 < CY P ≤ CY p + IY < 1, Cr + Ir < 0, IK < 0,

where Y is output (GDP), r is the ex ante (short-term) real rate of interest, K

is the capital stock, G is public spending on goods and services, Y p is private,
disposable income and εD is some unspecified demand shock. Disposable
income is
Y p ≡ Y − T, (13.2)
where T is real net tax revenue (equal to gross tax revenue minus transfers).
It is convenient to have two independent fiscal variables, one representing
government spending and another representing how it is financed, hence we
shall treat both G and T as exogenous variables. A balanced primary budget
is the special case T = G.
It is a weakness of the model that private financial wealth does not enter
as an argument in the consumption function C(·). The dynamic version of
the model in the next section remedies this. The role of r and K in the
investment function I(·) is consistent with the q-theory of investment. The
third argument in the investment function, current output, should rather be
expected future demand. This is a weak point which the New Keynesian
versions (Chapter 23) attempt to improve.
Changes in the capital stock are relatively small in the short run, hence
we ignore them and suppress the explicit reference to K. Inserting (13.2)
into (13.1) we can write aggregate demand as

Y d = D(Y, r, T ) + G + εD , where (13.3)

0 < DY = CY p + IY < 1, Dr = Cr + Ir < 0, (13.4)
DT = −CY p ∈ (−1, 0).
13.2. The traditional IS-LM model 429

The demand for money balances (currency and checkable deposits) is

given by
M d = P · (L(Y, i) + εL ), LY > 0, Li < 0, (13.5)
where P = the output price level, i = the (short-term) nominal rate of
interest and εL is a “liquidity preference” shock. There is a definitional link
between r and i, namely,
r = i − πe, (13.6)
where π = the (forward-looking) inflation rate (∆P/P ) and the superscript
e denotes expected value. Therefore, r denotes the ex ante − or expected −
real rate of interest. This, rather than the ex post real rate of interest, is the
relevant argument in the consumption and investment functions.
The IS-LM model studies the interaction between the output market and
the financial markets. It is for simplicity assumed that there are only two
assets, interest-bearing bonds and money. That is, financial wealth is partly
held in the form of money, partly in the form of bonds. Then, clearing
on the bond market implies clearing on the money market and vice versa.2
“Clearing” just means that the rate of interest has adjusted such that the
existing stocks of bonds and money are willingly held. So it is enough to
consider only one of these markets. Usually the money market is considered.
The standard IS-LM model assumes clearing at both the output and the
money market:

Y = D(Y, i − π e , T ) + G + εD , (IS)
M
= L(Y, i) + εL . (LM)
P
These two equations constitute the traditional IS-LM model.
The interpretation of the model is that output quickly adjusts to shifts
in demand, while the price level remains constant within the relatively short
time horizon considered. Behind this is the implicit assumption that the level
of employment, say N, required to produce Y, is feasible (given an aggregate
production function Y = F (K̄, N), where K̄ is the given capital stock). The
conception is that in the real world usually N < N̄, where N̄ is labour
2
By the “money market” is meant that abstract place where the money stock (supply)
“meets” money demand.
430 CHAPTER 13. IS-LM MODEL WITH RATIONAL EXPECTATIONS

supply. The difference, N̄ − N, is interpreted as some sort of involuntary

unemployment. By involuntary unemployment is meant a situation where
some people are without a job, but are willing to take a job at the going
wage or even a lower wage. Theories of efficiency wages, insider-outsider
configurations and wage bargaining explain how such a situation can be an
equilibrium (a state of balance).
Of course the assumption of a completely constant price level is an exag-
geration. One should think about this as wage and price level adjustments
taking place relatively slowly through an expectations-augmented Phillips
curve. Empirically, as mentioned above, there are many indications that
output movements are normally faster than price movements.
Notice that the IS equation postulates “equilibrium” or “clearing” at a
flow market: so much output per time unit matches the demand per time
unit for this output. In contrast, the LM equation postulates “equilibrium”
or “clearing” at a stock market: so much money at a point in time. The
endogenous variables are Y and either i or M, depending on which of the
two is the instrument used by the central bank. In Walrasian (classical and
neoclassical) thinking, the equilibrating variables are the prices, but here
they are: a quantity (production) and either a relative price, the interest
rate, or another quantity, the money supply. A third version of the model,
however, includes an interest rate rule so that both i and M are endogenous.
The variables P, π e , G, T, εD , εL and, in the traditional versions, either M
or i are exogenous. Usually π e = 0 is assumed, in view of the fixed price level.
The fiscal policy instruments are G and T ; the monetary policy instruments
are either M or i. In the analysis below we assume that the choice of policy
(what instruments to use and what particular value to give each instrument)
has to be decided before the shock is known.

Money stock as instrument (or at least intermediate target)

Here we consider M as exogenous. This is the traditional case. The idea

is that by buying and selling bonds on the bonds market the central bank
affects the money supply. In reality, however, the central bank has no direct
control over the money supply. It has control only over the monetary base,
13.2. The traditional IS-LM model 431

not the bank-created money (checkable deposits etc.). The money supply
can at best be an intermediate target for monetary policy, that is, a target
that can be reached in some average-sense over the medium term. Thus,
considering M as the monetary instrument does not fit entirely well with a
short-run model. Yet, we follow the standard approach and treat M as if it
could be directly and immediately controlled by the central bank.
The IS curve is the locus of combinations of Y and i that are consistent
with clearing at the output market, i.e., consistent with the IS equation.
Indeed, since Dr < 0, D is a monotonous function af i and so the IS equation
determines i as an implicit function of Y, π e , G, T and εD :

i = iIS (Y, π e , G, T, εD ).

The partial derivatives can be found by taking the total differential on both
sides of (IS):

dY = DY dY + Dr (di − dπ e ) + DT dT + dG + dεD . (13.7)

We find ∂i/∂Y|IS by putting dπ e = dG = dT = dεD = 0 and reordering:

∂i 1 − DY
= < 0, (13.8)
∂Y |IS Dr
where the sign comes from (13.4).
Similarly, the LM curve is the locus of combinations of Y and i that are
consistent with clearing at the money market, i.e., consistent with the LM
equation. Indeed, since Li < 0, L is a monotonous function af i and so the
LM equation determines i as an implicit function of M/P, Y and εL :
M
i = iLM (Y, , εL ).
P
The partial derivatives can be found by taking the total differential on both
sides of (LM):
M
d = LY dY + Li di + dεL . (13.9)
P
We find ∂i/∂Y|LM by putting d(M/P ) = dεL = 0 and reordering:

∂i −LY
= > 0. (13.10)
∂Y |LM Li
432 CHAPTER 13. IS-LM MODEL WITH RATIONAL EXPECTATIONS

ISπ e ,G ,T ,ε LM M
D
,ε L
P

Figure 13.1:

Fig. 13.1 shows the downward sloping IS curve and the upward sloping
LM curve. A solution (Y, i) of the model is unique and we can write Y and
i as implicit functions of all the exogenous variables:
M
Y = f (π e , G, T,
, εD , εL ), (13.11)
P
M
i = g(π e , G, T, , εD , εL ). (13.12)
P

Comparative statics How do the endogenous variables, Y and i, vary

with the exogenous variables? To find out, we calculate the partial derivatives
of the implicit functions f and g. A convenient method is the following.
There are given two equations, (13.7) and (13.9), and two new endogenous
variables, the changes dY and di. The changes, dπ e , dT, dG, dεD , d(M/P )
and dεL , in the exogenous variables are our new exogenous variables. The
system is simultaneous (not recursive). We first reorder (13.7) and (13.9)
so that dY and di appear on the left-hand side and the differentials of the
exogenous variables appear on the right-hand side:

(1 − DY )dY − Dr di = −Dr dπ e + DT dT + dG + dεD ,

M
LY dY + Li di = d − dεL .
P
From this linear system we find dY and di by Cramer’s rule:
13.2. The traditional IS-LM model 433

¯ ¯
¯ −D dπ e + D dT + dG + dε −D ¯
¯ r T D r ¯
¯ ¯
¯ M
d P − dεL Li ¯
dY =
∆
(−Dr dπ + DT dT + dG + dεD )Li + Dr (d M
e
P
− dεL )
= , (13.13)
∆
and

¯ ¯
¯ 1−D e
−Dr dπ + DT dT + dG + dεD ¯
¯ Y ¯
¯ ¯
¯ LY dM
P
− dεL ¯
di =
∆
(1 − DY )(d M
P
− dε e
L − (−Dr dπ + DT dT + dG + dεD )LY
)
= (13.14)
,
∆
where the determinant ∆ is defined by
¯ ¯
¯ 1 − D −D ¯
¯ Y r ¯
∆ = ¯ ¯ = (1 − DY )Li + Dr LY
¯ LY Li ¯
= (1 − CY P − IY )Li + (Cr + Ir )LY < 0.

Finally, we find the partial derivatives of f and g, respectively, w.r.t. the

real money supply, M/P, by putting dπ e = dG = dT = dεD = dεL = 0 above
and reordering. We get
∂Y Cr + Ir
M
= fM/P = > 0,
∂( P ) (1 − CY p − IY )Li + (Cr + Ir )LY
∂i 1 − CY p − IY
M
= gM/P = < 0,
∂( P ) (1 − CY p − IY )Li + (Cr + Ir )LY

in view of (13.4). Such partial derivatives of endogenous variables w.r.t.

exogenous variables are called “multipliers”. This is because the effect on,
e.g., Y of a small increase in M/P is determined as dY = (∂Y /∂( M P
))d( M
P
),
M
that is, the partial derivative acts as a multiplier on the increase, d( P ), in
the exogenous variable.3
The intuitive interpretation of the signs of the multipliers is the following.
The central bank increases the money supply by an open market purchase
3
Instead of using Cramer’s rule, in the present case we could just substitute di, as
434 CHAPTER 13. IS-LM MODEL WITH RATIONAL EXPECTATIONS

of bonds held by the private sector. Immediately after this the supply of
money is higher than before and the supply of bonds available to the public
is lower. At the initial interest rate there is now excess supply of money and
excess demand for bonds. But the attempt of agents to get rid of their excess
cash in exchange for more bonds can not succeed in the aggregate because
the supplies of bonds and money are given. Instead, what happens is that
the “price” of bonds goes up, that is, the interest rate goes down, until the
available supplies of money and bonds are willingly held by the agents.
How do shocks affect output under the present money stock targeting
rule? To see that, we put, first, dπ e = dG = dT = d M P
= 0 in the dY
equation (13.13). When in addition we put dεL = 0 (or dεD = 0), we find
the partial derivative of Y w.r.t. εD (or εL ) :

∂Y Li
= fεD = (13.15)
∂εD (1 − CY p − IY )Li + (Cr + Ir )LY
1 1
= ∈ (0, )
1 − CY p − IY + (Cr + Ir )LY /Li 1 − CY p − IY
∂Y −Cr − Ir
= fεL = < 0. (13.16)
∂εL (1 − CY p − IY )Li + Dr LY

As expected, a positive demand shock is expansionary, while a positive liq-

uidity preference shock is contractionary.

The short-term interest rate as instrument

Now we consider i as exogenous and M as endogenous. This version of the

model corresponds better to how central banks operate in practice. The
short-term interest rate has, in fact, the character of an instrument, since it
is a variable that the central bank can control. The central bank announces
its desired level of the short-term interest rate and adjusts the monetary base
through open market operations such that the actual short-term interest rate
determined from (13.9), into (13.7) and then find dY from this equation. In the next step,
the found solution for dY can be inserted into (13.9), which then gives the solution for di.
However, if Li were a function that could take the value zero, this procedure might invite
a temptation to rule this out by assumption. That would imply an unnecessary reduction
of the domain of f and g. The only truly necessary assumption is that ∆ 6= 0, and that is
automatically satisfied in the present problem.
13.2. The traditional IS-LM model 435

equals the announced interest rate. Instead of the upward-sloping LM curve

we get a horizontal line in the (Y, i) diagram in Fig. 13.1.
The implication is that the model is now recursive. Since M does not
enter the IS equation, Y is now given by this equation independently of the
LM equation. Indeed, the IS equation determines Y as an implicit function

Y = h(πe , G, F, i, εD , εL ). (13.17)

Comparative statics We find the partial derivatives of the h function

from (13.7). For example, the partial derivatives w.r.t. the interest rate, a
demand shock and a liquidity preference shock, respectively, are

∂Y Cr + Ir
= < 0,
∂i 1 − CY p − CY
∂Y 1
= > 1, (13.18)
∂εD 1 − CY p − CY
∂Y
= 0.
∂εL

A positive liquidity preference shock does not decrease output, because it

is immediately counteracted by a higher money supply (see below) so that the
interest rate does not increase. The liquidity preference shock is “cushioned”
by the monetary policy. On the other hand, a positive demand shock has a
larger effect on output than in the case of money stock targeting (compare
(13.18) to (13.15)). This is because in that case a dampening rise in the
interest rate was allowed to take place (the so-called financial crowding-out
effect). From a similar analysis, Poole (1970) concluded that:

• a money stock targeting rule is preferable if most shocks are output

demand shocks, while

• an interest rate targeting rule is preferable if most shocks are liquidity

preference shocks (money demand shocks).

Finally, the solution for the money supply is

M = P · (L(Y, i) + εL ),
436 CHAPTER 13. IS-LM MODEL WITH RATIONAL EXPECTATIONS

where the solution for Y from (13.17) can be inserted. By taking the total
differential on both sides we get
M
dM = P (LY dY + Li di + dεL ) + dP, (13.19)
P
which implies, e.g.,
∂M ∂Y Cr + Ir
= P (LY + Li ) = P (LY + Li ) < 0,
∂i ∂i 1 − CY p − CY
∂M 1
= P LY > 0,
∂εD 1 − CY p − CY
∂M
= P > 0,
∂εL
∂M M
= > 0.
∂P P

The balanced budget multiplier What is the effect on Y of an increase

in G leaving T unchanged (i.e., ∆G is financed by a corresponding increase
in the budget deficit)? We find this from (13.7) by putting dπ e = di =
dT = dεD = 0. We get ∂Y/∂G = 1/(1 − DY ) = 1/(1 − CY p − IY ) > 1,
in view of (13.4). What if the increase in G is financed by a corresponding
increase in T ? To get an answer, we first find, again from (13.7), ∂Y /∂T
= −CY p /(1 − CY p − IY ) < 0. The total combined effect of an equal increase
in both G and T is given by
∂Y ∂Y 1 − CY p
+ = ≥ 1.
∂G ∂T 1 − CY p − IY
This is called the balanced budget multiplier, and in case IY = 0 it is exactly
equal to 1. If the interest rate was allowed to increase, as in the case with
exogenous M, then the balanced budget multiplier would be smaller. But
with the present interest rate targetting rule this financial crowding-out effect
is eliminated.

An interest rate rule

Suppose the central bank conducts stabilization policy by using the interest
rate rule
i = i0 + i1 Y, i1 > 0,
13.3. Dynamic IS-LM model with forward-looking expectations 437

where i0 and i1 are policy parameters.4 If the LM curve in Fig. 13.1 were a
straight line, that diagram is again valid. Yet, since the interpretation of the
LM curve is in this case different, we should rather call it the IR curve (IR
for interest rate). Anyway, now both i and M are endogenous.
Inserting into (IS) gives

Y = D(Y, i0 + i1 Y − π e , T ) + G + εD .

By taking the total differential on both sides we find

∂Y ∂Y 1 1
= = ∈ (0, ),
∂G ∂εD 1 − CY p − IY − i1 (Cr + Ir ) 1 − CY p − IY
∂Y (Cr + Ir )Y
= < 0,
∂i1 1 − CY p − IY − i1 (Cr + Ir )
∂Y Cr + Ir
= − > 0,
∂π e 1 − CY p − IY − i1 (Cr + Ir )
∂Y
= 0.
∂εL
We see that all multipliers become ≈ 0, if the policy reaction coefficient i1
is large enough. In particular undesired fluctuations due to demand shocks
can be damped this way.
The corresponding changes in i are given as ∂i/∂x = i1 ∂Y /∂x for x =
G, εD , i1 , π e and εL . From (13.19) we find the corresponding changes in M as
∂M/∂x = P (LY + i1 Li )∂Y/∂x for x = G, εD , i1 and π e ; finally, from (13.19)
we have again ∂M/∂εL = P > 0.

13.3 Dynamic IS-LM model with forward-looking

expectations
The most important weakness of the above model is probably the absence of
endogenous forward-looking expectations. This motivated Blanchard (1981)
to develop a dynamic extension of the IS-LM model. The important elements
are:
4
Notice that an “interest rate targeting rule” is a limiting case of such an “interest rate
rule”, the case i1 = 0.
438 CHAPTER 13. IS-LM MODEL WITH RATIONAL EXPECTATIONS

• The focus is manifestly on short-run mechanisms. Therefore the model

allows for a possible deviation of output from demand − the adjustment
of output to demand takes time (so that in the meanwhile changes in
inventories occur). In this way the model becomes dynamic.

• Agents have endogenous forward-looking expectations. Indeed, agents

have rational expectations (model consistent expectations). Since there
are no stochastic elements in the model, this is the same as perfect
foresight.

• There is a distinction between the long-term interest rate and the short-
term interest rate, i.e., the structure of interest rates can be studied.
These elements lead to a richer model than (IS)-(LM) above. The dy-
namic model conveys explicitly the central message of Keynesian theory,
namely that the key equilibrating role is taken by output changes generated
by discrepancies between aggregate demand and supply.5 In addition, letting
agents have endogenous forward-looking expectations (instead of exogenous
expectations) is an important step towards a realistic model of macrodynam-
ics. Finally, the distinction between the long-term interest rate and the short-
term interest rate opens up for a concise indicator of expectations. While it
is the short-term interest rate which the central bank can control, it is the
long-term rate which is a major determinant of investment and consump-
tion. Firm’s investment in physical capital is normally an endeavour with a
long time horizon. Similarly, the decision concerning the trade-off between
consumption and saving tends to be based on a quite long-run perspective.

13.3.1 Basic framework

Let R denote the real long-term interest rate (below defined more precisely
as the internal rate of return on a consol). To obtain a better description of
aggregate demand we replace r by R in (13.1) from the static model:
Y d = D(Y, R, T ) + G, where
0 < DY = CY p + IY < 1, DR = CR + IR < 0, DT = −CY p ∈ (−1, 0).
5
This expounding of the central message in Keynes’ General Theory is contended by,
e.g., Patinkin (1976, 1979).
13.3. Dynamic IS-LM model with forward-looking expectations 439

In order not to have too many balls in the air at the same time, the model
ignores the stochastic term εD in (13.1) (as well as εL in the money demand
function). Since the model is in continuous time, including stochastic terms
would require the use of stochastic differential calculus.
By having consumption demand depending negatively on the long-term
interest rate, the model realistically include important wealth effects on con-
sumption.6 On the production side, given that the aim is short-run analysis,
it is natural to let the adjustment of output to demand take time. Hence,
we replace the IS equation from the static model by the error-correction
specification
dYt
Ẏt ≡ = λ(Ytd − Yt ) (13.20)
dt
= λ(D(Yt , Rt , T ) + G − Yt ),

where λ > 0 is the adjustment speed (here assumed constant) and t is time.
The rest of the model is straightforward:
Mt
= L(Yt , it ), LY > 0, Li < 0. (13.21)
P
Rt = 1/Qt , (13.22)
rt ≡ it − π e , (13.23)
1 + Q̇et
= rt , (13.24)
Qt
where Q is the real price of a long-term bond, here identified as a consol
paying to the owner a constant stream of one unit of account per time unit in
the indefinite future. Equation (13.21) is the usual equilibrium condition for
the money market. Financial markets − and i − adjust very fast; therefore;
therefore, it is not unreasonable to assume clearing at any instant. Equation
(13.22) tells us that the long-term interest rate at time t is the reciprocal of
the real market price of a consol at time t. This is just another way of saying
6
Our presentation of the model is close to that in Blanchard and Fischer (1989). In the
original Blanchard (1981) paper, however, the forward-looking variable is Tobin’s q rather
than the long-term interest rate, R. But since the (real) long-term interest rate can, in
this context, be considered as the inverse of Tobin’s q, there is essentially no difference.
Wealth effects come true whether the souce is interpreted as changes in Tobin’s q or the
long-term interest rate.
440 CHAPTER 13. IS-LM MODEL WITH RATIONAL EXPECTATIONS

that the long-term rate Rt is defined as the internal rate of return on the
consol, i.e., that number Rt which satisfies the equation
Z ∞
1
Qt = 1 · e−Rt (s−t) ds = . (13.25)
t Rt
That is, the discount rate Rt which transforms the payment stream on the
concol into a present discounted value at time t equal to the market price of
the consol at time t is the long-term rate. Inverting (13.25) gives (13.22).
Equation (13.23) defines the (expected) short-term real rate of interest.
Finally, equation (13.24) is an equilibrium condition (or a no-arbitrage con-
dition) for the two bond markets, saying that the expected real rate of return
on the consol (including a possible capital gain or capital loss) must be equal
to the real rate of return on the alternative asset, the short-term bond. In
general, in view of the higher risk associated with long-term claims, pre-
sumably a positive risk premium should be added on the right hand side of
(13.24). We assume rational expectations, however, i.e., model consistent
expectations. And since there is no uncertainty in the model (no stochastic
elements), rational expectations implies perfect foresight and therefore the
risk premium vanishes.7 The assumption of perfect foresight also implies that
Q̇et = Q̇t . And since the price level P is assumed constant in the model, we
have π et = π t = 0 for all t. Therefore, equation (13.23) reduces to rt = it for
all t.
The exogenous variables are G, T, P and either M or i, depending on the
choice of monetary policy instrument.
In any case, the model can be reduced to two coupled first order differen-
tial equations in Y and R. The first of these equations is (13.20) above. As
to changes in Q, from (13.24), with Q̇ = Q̇e , and (13.22) we have

1 Q̇ Ṙ
+ = R − = r = i, (13.26)
Q Q R
in view of (13.23) with π e = 0 (from now, the dating of the variables is
suppressed unless needed for clarity). By reordering,

Ṙ = (R − i)R. (13.27)
7
If a constant risk premium α were added, the dynamics of the model will only be
slightly different in the case α > 0 compared to the case α = 0.
13.3. Dynamic IS-LM model with forward-looking expectations 441

As in the static IS-LM model there are two different cases to consider: case
1 where the money stock is the policy instrument (or at least an intermediate
target), and case 2 where the interest rate is the policy instrument. Case 2
is by far the simplest one and as argued above, it is in some sense closest to
what modern monetary policy is about. Case 1 is also of interest, however,
both because of its historical appeal and because it yields very impressive
dynamics. In addition, case 1 has some affinity with what happens under an
interest rule.
Before considering the two cases separately, we shall emphasize an equa-
tion which is very useful for the economic interpretation of the dynamics.
Assuming no speculative bubbles (see below), the ”no-arbitrage” formula
(13.24) is equivalent to a statement saying that the market value of the con-
sol is equal to its fundamental value, that is, the present value of the future
payments on the consol, where present value is calculated with the market
short-term interest rate as the rate of discount:
Z ∞ U
s
Qt = e− t rτ dτ ds, so that (13.28)
t
1 1
Rt = = R ∞ U s r dτ .
Qt t
e t τ ds
−

This equivalence result can be obtained by solving the differential equation

(13.24), given that there are no speculative bubbles. In other words: the
long-term rate, Rt , is a kind of average of the (expected) future short-term
rates, r. The higher are these, the lower is Qt , and the higher is Rt .

13.3.2 Money stock as instrument

Here we consider M as exogenous, so that the short-term rate, i, is endoge-
nous. Then, equation (13.21) gives i as an implicit function of M/P and Y ,
i.e.,
M
r = i = i(Y, ), with iY = −LY /Li > 0, iM/P = 1/Li < 0. (13.29)
P
Inserting this into (13.27) we have
M
Ṙ = (R − i(Y, ))R, (13.30)
P