IS-LM Model

Dudley Cooke

Trinity College Dublin

Dudley Cooke (Trinity College Dublin) IS-LM Model 1 / 67

Reading

Mankiw and Taylor (2008), Macroeconomics: Chapter 10.1 and .2

and 11.1

Dudley Cooke (Trinity College Dublin) IS-LM Model 2 / 67

Plan for Next Three Lectures

IS curve
LM curve
ISLM equilibrium
Fiscal/monetary policy in ISLM model
Policy applications

Dudley Cooke (Trinity College Dublin) IS-LM Model 3 / 67

Basic Assumptions

Closed economy.
Exogenously fixed nominal price level, P.
Inflation expectations are exogenous.
r = i.
There is one basket of goods.

Dudley Cooke (Trinity College Dublin) IS-LM Model 4 / 67

Keynesian Cross and Investment Demand

Keynesian Cross shows planned expenditure (E p ):

Ep ≡ C + I + G

or,
hhlds firms govt.
z }| { z }| { z}|{
Ep ≡ C( Y − p
| {z } ) + I (r ) + G
T
disp. incm

Consumption rises with income: ∆C /∆Y > 0.

National saving, S, is composed of private saving (Y − T − C ) plus
government saving (T − G ). Saving equals investment:

S =I

Dudley Cooke (Trinity College Dublin) IS-LM Model 5 / 67

Keynesian Cross and Investment Demand

Keynesian Cross:

1 Equilibrium: planned expenditure=spending (income): E p = Y .

2 Higher r , lowers I , which lowers E p s.t. E1P < Y0 . Y decreases from
Y0 to Y1 to reach equilibrium - ∆I ‘faster’ than ∆Y .1

Investment Demand:

1 Investment falls with the real interest rate (I p is planned investment):

∆I p /∆r < 0.
2 We typically assume there is a linear relationship.

1 We also see that investment is more variable than output in the data.
Dudley Cooke (Trinity College Dublin) IS-LM Model 6 / 67
Investment Demand Schedule

r, Interest Rate


r1

r0

I p (r)

I1 I0 I, Investment


Dudley Cooke (Trinity College Dublin) IS-LM Model 7 / 67

Keynesian Cross
Equilibrium: planned expenditure=spending
(income): E p = Y .

E p , Planned Expenditure

Y = Ep
E p (r0 )

E p (r1 )

ΔI


Y1 Y0 Y , Output

Higher r, lowers I, which lowers E p s.t. E1

P <
Dudley Cooke (Trinity College Dublin) IS-LM Model 8 / 67
From the Keynesian cross to IS Curve

IS Curve: combinations of real output (GDP) and (real) interest rate

such that planned and actual expenditures are equal.

E p = E (Y , r , G , T ) ≡ C (Y − T ) + I p (r ) + G

Totally differentiating C (Y − T ) + I p (r ) + G w.r.t. Y and r (assuming

fiscal policy (G and T ) is fixed) yields:

∆Y = ∆E p = CY ∆Y + Ir ∆r

where 0 < CY < 1 is the MPC (and slope of the planned exp. line in the
Keynesian cross diagram) and Ir < 0. So,

(∆Y ) /(∆r )|IS = Ir /(1 − CY ) < 0

Dudley Cooke (Trinity College Dublin) IS-LM Model 9 / 67

The IS Curve
Diagram: IS Curve

r, Interest Rate


r1 Ep < Y

r0
Ep > Y

IS

Y1 Y0 Y , Output


At points where E pIS-LM

= Model
Y the goods market is
Dudley Cooke (Trinity College Dublin) 10 / 67
Slope of the IS Curve

A given change in the interest rate will have a bigger impact on

output the flatter the IS curve. That is, if either:

1 The interest sensitivity of planned expenditure (via investment Ir ) is

high ⇒ planned expenditure line shifts further, so output falls
further.
2 The marginal propensity to consume out of disposable income (CY ) is
large ⇒ higher MPC implies steeper planned expenditure line, so
output must fall further in response to a given downward shift of the
planned expenditure line to return to planned = actual expenditure.

Dudley Cooke (Trinity College Dublin) IS-LM Model 11 / 67

Shifting the IS Curve

Assume r = i is fixed. Increase in government purchases G (in

general, change autonomous spending).
Recall, Y = C (Y − T ) + I p (r ) + G , then differentiate w.r.t. Y and
G , with T , I p (r ) fixed ⇒ ∆Y = CY ∆Y + ∆G , and rearrange ⇒ the
government purchases multiplier is:

(∆Y ) /(∆G )|r = 1/(1 − CY ) > 1

Magnitude of govt purchases multiplier: 0 < (CY = MPC ) < 1 so

1/(1 − CY ) > 1.
Intuition: An increase in G raises private income. This raises
consumption, which itself raises private income. Thus there is a
multiplied effect of government spending.

Dudley Cooke (Trinity College Dublin) IS-LM Model 12 / 67

Government Spending
Diagram: Government Spending

E p , Planned Expenditure

Y = Ep
E p (.; G1 )
p
E1
ΔG E p (.; G0 )

p
ΔC
E0

i = r Y0 Y1 Y , Output

i0

I S(G1 )
I S(G0 )

ΔY Y , Output

Dudley Cooke (Trinity College Dublin) IS-LM Model 13 / 67

Comments

The Keynesian cross and IS curve actually show the same thing. The
difference is that we represent them in different spaces.
The Keynesian cross in (E p , Y ) space and the IS in (r , Y ) space.
In (r , Y ) space, we have to hold the interest rate fixed when we look
at the effects of government spending.
To complete the analysis we also need to include the money and
bonds markets alongside the goods market.
In this case, we can see what impact changes in fiscal policy have on
things such as the interest rate.

Dudley Cooke (Trinity College Dublin) IS-LM Model 14 / 67

Money and Bonds Portfolio

Agents have access to two assets

1 they hold money (M) to spend on goods.

2 they hold bonds (B, e.g. consols: pay fixed yearly amount ($ 1)
forever, starting next year) to save, i.e. spend in the future

Real financial wealth, A, is given by:

A = M s /P + (PB /P )B s

where, s denotes stock.

A is basically a reflection of the portfolio of an individual.

Dudley Cooke (Trinity College Dublin) IS-LM Model 15 / 67

The Bond Market

B s is the total number of bonds issued (= volume of bonds) and the

price of bonds is PB (so PB B s is nominal value of bonds)
Assume the demand for bonds is given by:

Bd ( i , Y )
+ +

If income rises so does the demand to hold bonds - some income is

held in cash to buy goods now, some in bonds, to buy goods later.
If the interest rate rises (‘payoff’ from holding the bond) the price of
bonds falls and bond demand increases.

Dudley Cooke (Trinity College Dublin) IS-LM Model 16 / 67

The Money Market

Money stock equals currency and liquid bonds (regular bonds are
illiquid within period).
Treat supply of nominal money balances M s as exogenous. Recall
that the price level, P, is exogenous (and assumed fixed).
The demand for money is given by:

(M/P )d = L( i , Y )
− +

∆L/∆i ≡ Li < 0 ⇒ if the interest rate goes up you put more into
bonds (money bears no interest).
∆L/∆Y ≡ LY > 0 ⇒ if your income goes up, you consume more
now. To do this you need more money.

Dudley Cooke (Trinity College Dublin) IS-LM Model 17 / 67

More on Money Demand

Revision comments:
i = opportunity cost of holding money (i.e. you could put you money
into bonds and get interest back).
L(·) stands for Liquidity Preference.
The role of money more generally is as a ...
medium of exchange: intermediary used in trade to avoid a pure
barter system.
store of value: measurement of the market value of goods.
unit of account: able to be reliably saved, stored, and retrieved
M s , i.e. valueless money, is called Fiat money.2

2 Think: not gold coins.

Dudley Cooke (Trinity College Dublin) IS-LM Model 18 / 67
Equilibrium

If the money market is in equilibrium so is the bond market. That is:

h i
M s /P − L(i, Y ) = (PB /P ) B d (i, Y ) − B s

If M s /P > M d /P ≡ L(i, Y ) ⇒ B d (i, Y ) > B s etc. Given Y and P,

i falls to clear the market.
We also need to impose that money demand equal money supply:

M s /P = (M/P )d

Now we have an LM curve:

M s /P = L(i, Y )

where M s is exogenous.3
3 We usually think of M s as M0 (notes and coins). Other types include M2, M4,

which are broader definitions.

Dudley Cooke (Trinity College Dublin) IS-LM Model 19 / 67
LM Curve

LM curve shows combinations of real output and interest rate such

that the money market is in equilibrium, for a given price level.

M s /P = L(i, Y )

Again: ∆L/∆i ≡ Li < 0 and ∆L/∆Y ≡ LY > 0.

Note: we use the nominal interest rate for the LM and the real
interest rate for the IS.

1 the nominal interest rate affects individuals portfolio decision b/w

money and bonds.
2 the real interest rate affects firms investment decisions.
..... however, we have assumed i = r .

Dudley Cooke (Trinity College Dublin) IS-LM Model 20 / 67

Changes in Output

Although output is endogenous, we can ask what happens if it

changes.
Suppose Y1 > Y0 . This raises money demand, L(i, Y ).
As M s /P is fixed M s < M d and L needs to fall.
However, M s < M d is consistent with B d (i, Y ) < B s .
As B s is fixed, the price of bonds falls, which is equivalent to a higher
interest rate (that is, a higher payoff to holding a bond).

Dudley Cooke (Trinity College Dublin) IS-LM Model 21 / 67

Money Market and LM Curve
Diagram: Money Market and LM Curve

i, Interest Rate

i1

L(i, Y1 )
i0

L(i, Y0 )

M0 /P M/P , Real Balances

i = r

M0 /P > L LM

i1

i0
M0 /P < L

Y0 Y1 Y , Output

Dudley Cooke (Trinity College Dublin) IS-LM Model 22 / 67

Slope of the LM Curve

LM curve slopes upwards in (i, Y ) space. This can be seen by totally

differentiating M s /P = L(i, Y ) w.r.t. i, Y :

0 = Li ∆i + LY ∆Y

So, (∆i )/(∆Y )|LM = −LY /Li > 0 as LY > 0 and Li < 0.

A given change in income will have a smaller impact on the interest

rate along the LM curve, i.e. the LM curve is relatively flat, either

1 the lower the income sensitivity of money demand, LY (i.e., the

increase in money demand is less when output rises).
2 the higher the interest sensitivity of money demand, Li .

Dudley Cooke (Trinity College Dublin) IS-LM Model 23 / 67

Shifting the LM Curve

So far, M s has been fixed. However, it can also be a policy variable.

Suppose there is an increase in M s .
This implies M s > M d at the current interest rate and so individuals
prefer to buy bonds rather than hold this extra cash.
This raises the price of bonds and i falls to clear the market.
People are now happy to hold the additional money balances and the
money market returns to equilibrium at a lower interest rate.

Dudley Cooke (Trinity College Dublin) IS-LM Model 24 / 67

Shifting the LM Curve
Diagram: Shifting the LM Curve

r, Interest Rate

ΔM


Δi

L(i, Y0 )

M0 /P M1 /P M/P , Real Balances

LM (M0 /P )

LM (M1 /P )
i0

i1

Y0 Y , Output

Dudley Cooke (Trinity College Dublin) IS-LM Model 25 / 67

The Short-Run ISLM Equilibrium

Recap

1 IS Curve gives combinations of real output (GDP) and real interest

rate such that planned and actual expenditures on real output are
equal.
2 LM Curve gives combinations of real output and nominal interest
rate such that the money market is in equilibrium, for a given price
level.

As r = i we can plot the IS and LM conditions in (i = r , Y ) space as

we have two conditions in two unknowns.

Dudley Cooke (Trinity College Dublin) IS-LM Model 26 / 67

Stability of the ISLM Equilibrium

Goods Market:
Y = E p = C (Y − T ) + I p (r ) + G
If Y > E p goods demand is too low, firms accumulate unwanted
inventories (I0 > 0) , so they cut back on production. And
Y < E p ⇒ I0 < 0.

Money Market:
(M d /P ) = (M s /P ) = L(i, Y )
If M s > M d , then bond supply is less than bond demand and the interest
rate falls to clear the market. With M d > M s , it is the opposite.

Equilibrium:
M s = M d and Y = E p .

Dudley Cooke (Trinity College Dublin) IS-LM Model 27 / 67

IS-LM Equilibrium
Diagram: ISLM Equilibrium

r, Interest Rate

LM
A

B
r∗ C

D
IS

Y∗ Y , Output

A: M/P > L and Y IS-LM

>E p
Dudley Cooke (Trinity College Dublin)
Model 28 / 67
Round-up of ISLM so far ...

ISLM captures the demand side and is short-run focused.

Firms invest and households consume, hold money and save (buy
bonds to consume later).
ISLM is not really concerned with production - i.e. where the goods
come from. That is the supply side.
Next we want to consider policy options.

Dudley Cooke (Trinity College Dublin) IS-LM Model 29 / 67

Functional Forms

In macro we often adopt specific functional forms:

They (can) make things easier
They allows us to get explicit solutions (and quantify things)
See the Appendix of Ch. 11 in Mankiw’s textbook for a similar
analysis to that below.

Dudley Cooke (Trinity College Dublin) IS-LM Model 30 / 67

Elasticities

We assume the IS and LM equations have the following form:

ms − p = ky − ei

y = a + δ (y − t ) + h0 − γr + g
The parameters we have chosen measure the elasticities.4
In this course we will usually work with these types of equations.
Advantage - we can solve the model explicitly and find the multipliers
for policy
Disadvantage - we may lose some intuition.

4 Note: we have switched from (mostly) upper case to lower case letters.
Dudley Cooke (Trinity College Dublin) IS-LM Model 31 / 67
Exogenous/Endogenous Variables and Parameters

1
i= [a − (1 − δ) y − δt + h0 + g ] : IS
γ
1
i = [ky − (ms − p )] : LM
e

Endogenous: i = interest rate and y = output.

Exogenous: ms = money supply, p = price level, a = autonomous
consumption, t = taxes, h0 = autonomous investment, g =
government spending.
Parameters: γ =interest elasticity of investment, δ < 1 = MPC,
e = interest semi-elasticity of money demand, k = income elasticity
of money demand.

Dudley Cooke (Trinity College Dublin) IS-LM Model 32 / 67

Parameter Restrictions and The Quantity Equation

The velocity equation is usually written as MV = PY . In logs,

m − p = y − v.
The simplest case is V constant and equal to one. In that case,

m−p = y

We can now see that our LM is a generalized version of this equation,

with e = 0 and k = 1.
For example, we can measure the impact of changes in the interest
rate on liquidity by varying the interest elasticity of money demand, e.

Dudley Cooke (Trinity College Dublin) IS-LM Model 33 / 67

ISLM Equilibrium
Diagram: ISLM Equilibrium (with slopes)

r, Interest Rate

sl: −(1 − δ)/γ LM

r∗ sl: κ/

IS

y∗ y, Output


Dudley Cooke (Trinity College Dublin) IS-LM Model 34 / 67

Equilibrium

With functional forms the solution to the ISLM model can be written
in the following way:

h γ i
y ∗ = Ω a − δt + h0 + g + (ms − p )
e
  −1
γk
where Ω ≡ 1 − δ + >0
e

1
i∗ = [ky ∗ − (ms − p )]
e
Now we can see exactly how monetary and fiscal policy affect
output and the interest rate and the significance of the elasticity
assumptions.

Dudley Cooke (Trinity College Dublin) IS-LM Model 35 / 67

Fiscal Policy: ↑ g

An increase in government spending raises output. But there are two

mechanisms at work.
Through the IS curve output goes up (this direct effect, via the
Keynesian cross, is 1/ (1 − δ)).
However, the interest rate goes up. And we know that reduces
investment. This lowers output (an indirect effect).
  −1
Overall, output rises by Ω = 1 − δ + γk e < (1 − δ)−1 . So,
fiscal policy is said to crowd out investment.

Dudley Cooke (Trinity College Dublin) IS-LM Model 36 / 67

Fiscal Policy and Crowding Out
Diagram: Fiscal Policy and Crowding Out

r, Interest Rate

LM

r0

I S(g1 )

I S(g0 )

y0 y1 y, Output


Government Purchases Multiplier



Effect on Output WITH Crowding Out

Here g0 → g1 , where g0 < g1 .

Dudley Cooke (Trinity College Dublin) IS-LM Model 37 / 67
Monetary Policy: ↑ ms

Suppose there is an increase in ms .

This implies ms > md at the current interest rate and so individuals
prefer to buy bonds rather than hold this extra cash.
However, i falls to clear the market such that people are happy to
hold additional money balances and the money market returns to
equilibrium at a lower interest rate.
This impacts the goods market, as a reduction in i stimulates
investment.
This raises expenditures, and subsequently y .
We call this the ‘monetary transmission mechanism’.

Dudley Cooke (Trinity College Dublin) IS-LM Model 38 / 67

Monetary Trnasmission Mechanism
Diagram: Monetary Transmission Mecha-
nism

r, Interest Rate

LM (m0 − p)

LM (m1 − p)

r0
r1

IS

y0 y1 y, Output

Here m0 → m1 , where m0 < m1 and p is fixed.

Dudley Cooke (Trinity College Dublin) IS-LM Model 39 / 67

Analytics

The lowering of the interest rate by printing money is sometimes

called the ‘liquidity’ effect.
The effect on output of policy is given by:

∆y ∗ /∆ms = (γ/e) Ω > 0

The change in the interest rate is:

1
∆i ∗ /∆ms = [k (∆y /∆ms ) − 1] ≶?
e
= − (1 − δ) / [(1 − δ) + γk ] < 0

Mechanism: ∆ms , p = p → ∆i → ∆I → ∆y .

Dudley Cooke (Trinity College Dublin) IS-LM Model 40 / 67

Fiscal vs. Monetary Policy

Another question we can now ask - what are the relative powers of
fiscal and monetary policy on output?
Since we have adopted functional forms we can answer this type of
question.
∆y ∗ /∆g Ω
=
∆y /∆m
∗ s (γ/e) Ω
e interest elasticity of money demand
= =
γ interest elasticity of investment
If e > γ, then fiscal policy (as studied here) is more effective than
monetary policy.

Dudley Cooke (Trinity College Dublin) IS-LM Model 41 / 67

On Keynesian vs. Monetarists

What do Keynesians think (roughly): e is large and γ is small (poss.

‘animal spirits’) ⇒ fiscal policy is more important.
What do Monetarists think (roughly): the opposite! They think e is
small, the LM is steep and monetary policy is more important.
This makes knowing (i.e. estimating) the elasticities very important.
But that turns out to be difficult.

1 data issues
2 stability of money demand over time.

Dudley Cooke (Trinity College Dublin) IS-LM Model 42 / 67

Should we take this seriously?

We note that monetary policy may not be ∆ms . Central banks tend
to use short-term interest rates.
The same point holds for fiscal policy. We can interpret ∆g as
building roads or hospitals. However, fiscal policy is many other
things (including changes in taxation).
Also, we think of monetary policy as happening quickly (the central
bank sets i in the UK once a month and it’s effects last up to three
years).
Fiscal policy can take much longer to implement and be set
permanently or temporarily (Obama’s recent fiscal stimulus).

Dudley Cooke (Trinity College Dublin) IS-LM Model 43 / 67

More Policies - Germany in the 1990s
More Policies - Germany in the early 1990s

• Unification demands (i.e. a rise in g) along-

side inflation fears kept in place by a mon-
etary contraction (i.e. a fall in m)

r, Interest Rate LMpost

LMpre
r1 

r0

I Spost

I Spre


y1 y0 y, Output

Output Change: Ambiguous

Dudley Cooke (Trinity College Dublin) IS-LM Model 44 / 67

US in the 1990s
US in the early 1990s

• Fiscal consolidation (here, modeled as a

drop in g) whilst the Fed tried to avoid the
recession and allowed a ‘monetary easing’
(here, increase m).

r, Interest Rate
LM (m0 , low ) LM (m1 )
LM (m0 , high )

LM (m1 )

r0

I S(g0 )
I S(g1 )
y0 y, Output
low  

high  

Dudley Cooke (Trinity College Dublin) IS-LM Model 45 / 67

Other Examples

UK in the early 1980s:

Thatcher govt. elected in 1979. ‘Right-wing’ policies of fiscal
prudence and anti-inflation policies.
This caused a large recession in the UK.
US in early 2000s:
George W. Bush’s large tax cuts and relatively lax monetary policy
Complicated by the US economy borrowing heavily from abroad to
maintain high consumption levels.
Not brave enough to comment on the Irish situation!

Dudley Cooke (Trinity College Dublin) IS-LM Model 46 / 67

Recap

We built an ISLM model based on Keynesian Cross and Money and

Bond market equilibrium.
We used it to study Fiscal (and crowding out) and Monetary Policy.
It helps to clarify the policy options and potential pitfalls.
But - what about the difference between real and nominal interest
rates? What role does that play, if any?

Dudley Cooke (Trinity College Dublin) IS-LM Model 47 / 67

Fisher Equation

Nominal Interest Rate is the interest rate expressed in units of

money (it )
It tells us how much money we have to pay in the future in exchange
for having one more unit of money today (1 + it )
Real Interest Rate is the interest rate expressed in terms of a basket
of goods (rt )
It tells us how many goods we have to give up in the future in
exchange for having one more basket of goods today (1 + rt )
The real interest rate is important since agents consume goods and
not money.

Dudley Cooke (Trinity College Dublin) IS-LM Model 48 / 67

Nominal and Real Interest Rates

Suppose you borrow money today to buy a good of price Pt . Then

you have to repay (1 + it +1 )Pt next year.
In terms of goods (real terms), next period, you need to deflate by
what you expect the price level to be. That is, Pte+1 .
Thus, you expect to payback, in real terms,
Pt
(1 + it +1 )
Pte+1

It follows that the one-year real interest rate is,

Pt
(1 + rt +1 ) = (1 + it +1 )
Pte+1

Dudley Cooke (Trinity College Dublin) IS-LM Model 49 / 67

Nominal and Real Interest Rates

Expected inflation is defined as

Pte+1 − Pt
πte+1 =
Pt

We find:(1 + rt +1 ) = (1 + it +1 ) / 1 + πte+1 . But if it +1 and πte+1

are small, then,

(1 + it +1 ) / (1 + πte+1 ) ≈ 1 + it − πte+1

and,

rt +1 = it +1 − πte+1

This is the Fisher Equation.

Dudley Cooke (Trinity College Dublin) IS-LM Model 50 / 67

Ex-ante and ex-post real interest rates

The real interest rate actually captures two periods. This is reflected
in Pt and Pte+1 .
When we borrow/lend we don’t know what inflation will be over the
period.
This leads to two concepts of the real interest rate (dropping t’s).

1 r when the loan is made (or r e ): ex-ante rate: r e = i − π e

2 r once the inflation rate is realized: ex-post rate: r = i − π

These will only be the same if our expectation is correct.

Dudley Cooke (Trinity College Dublin) IS-LM Model 51 / 67

Implications of the Fisher Equation

We distinguish three cases:

π e = 0 ⇒ r = i (used above)
πe > 0 ⇒ r < i
πe = i ⇒ r = 0

1 Notice that i ≥ 0 (referred to as the Zero Lower Bound) but r ≶ 0.

US has recently had r < 0.
2 Fisher Hypothesis: nominal interest rate changes one-for-one with
the rate of change of the money supply (no effect on the real interest
rate).

Dudley Cooke (Trinity College Dublin) IS-LM Model 52 / 67

Fisher Effect (Source: Mankiw)
Fisher Effect (Source Mankiw)

• i is on the vertical axis and inflation re-

sponds 1-for-1 with the growth in the money
supply.

• Data appear to support the Fisher effect.

Dudley Cooke (Trinity College Dublin) IS-LM Model 53 / 67
The ISLM Model and the Fisher Equation

We use the same model as before - we eliminate r from the IS using

the Fisher equation.

ms − p = ky − e (r + π e )

y = a + δ (y − t ) + h0 − γr + g

We assumed that expectations are exogenous. What happens if

expectations change?
There are real effects, that is, output changes. Expected inflation
influencing output is called the Mundell-Tobin effect.

Dudley Cooke (Trinity College Dublin) IS-LM Model 54 / 67

The US Depression - 1930s

Since the decline in income in 1930’s coincided with falling interest

rates some suggest there was a contractionary shift in the IS curve.
Causes:

1 A downward shift in the consumption function (i.e., the C (Y − T )

part of the Keynesian cross)?
2 A large drop in housing investment? - there was a residential
investment boom in the 1920s ⇒ overbuilding.
3 Amplification: banks also failed. Bad loans were made and not paid
back. This lowered investment demand and loans to businesses.

Dudley Cooke (Trinity College Dublin) IS-LM Model 55 / 67

Japan’s Liquidity Trap (most of the 1990s)

Keynes argued that during a depression, such as the US in the 1930s,

monetary policy would be ineffective at influencing aggregate demand.
Why?
Monetary Policy works by lowering the nominal interest rate.
However, we know that there is a zero lower bound problem, that is,
i ≥ 0.
If output is very low (i.e. in a depression) we can’t keep reducing the
nominal interest rate.
Paul Krugman suggested the same thing happened in Japan in the
1990s.

Dudley Cooke (Trinity College Dublin) IS-LM Model 56 / 67

Some more details on Japan

In the 1980’s the Japanese economy was booming. However, there

was a stock market bubble.

1 Eventually there was a drop in stock prices and the wealth of

individuals dropped significantly.
2 Banks, trying to make profits, had lent to risky companies. They
failed and this magnified the effect of the stock price fall (a ‘credit
crunch’).
3 We might think of this as a shock (negative) hitting the IS curve, via
investment and consumption.
4 This also coincided with a low interest rate period.

Dudley Cooke (Trinity College Dublin) IS-LM Model 57 / 67

The The
NIKKEI Indexin
crunch Japan can be seen from the
NIKKEI Index of shares.

Dudley Cooke (Trinity College Dublin) IS-LM Model 58 / 67

 by the Fisher equation.

Liquidity Trap in the ISLM Model

Diagram: Japan’s Liquidity Trap

r, Interest Rate

LM

r0 = i0 = 0

IS
y∗ y y, Output

... where y is Full Employment

Japan hit it = 0. People also thought πte = 0 or slightly positive.

Then rt ' 0, by the Fisher equation.
Dudley Cooke (Trinity College Dublin) IS-LM Model 59 / 67
Policy Recommendations

We have a recession situation (i.e. y ∗ < y ) and there are a limited

number of policy options.

1 Massive Fiscal Expansion

2 Create Inflation Expectations (i.e. π e > 0)
3 Large ↑ m could boost trade, via the exchange rate (we haven’t
covered this yet)

Dudley Cooke (Trinity College Dublin) IS-LM Model 60 / 67

 Fiscal stimulus seems the obvious policy. One
Fiscal Policy
problemRevisited
is, just how long can a government we
keep it’s spending money?

r, Interest Rate

LM (m0 − p; π e)

r0 = i0 = 0
IS(g1)

IS(g0)
y0 y1 y y, Output

Dudley Cooke (Trinity College Dublin) IS-LM Model 61 / 67

 Despite the problems Japanese policymakers
Fiscal Policy in Japan
attempted to provide the economy with a boost
via fiscal policy.

Dudley Cooke (Trinity College Dublin) IS-LM Model 62 / 67

Monetary/Quantitative Easing

Japan did also attempt to stimulate the economy via monetary policy,
a policy called “quantitative monetary easing” - essentially this
involved trying to boost liquidity in financial markets.
This did have some (limited) impact.
However, by then, expectations we fixed at π e ' 0.
So, what if we could somehow alter π e ?

Dudley Cooke (Trinity College Dublin) IS-LM Model 63 / 67

‘Unusual’ Monetary Policy

How is all of this relevant for today? In 2000s we saw globally low
interest rates, which lead to a bubble. There was also excess lending
by banks. Similar to Japanese problem. Now the UK is using QM.
As is the Eurozone and Fed.
Alternative ideas:

1 One other (unusual) option is to attempt to raise π e . For a given i,

the real interest rate will fall, boosting investment. However, π e is
endogenous. We have assumed it is exogenous. So this solution
creates other problems
2 Another idea put forward - along the same lines - is that π e can
also affect output because the nominal interest rate rises less than
one-for-one with the rate of change of the money supply (contradicts
the Fisher Hypothesis) as agents change money for bonds (i.e.
portfolio reallocation), itself altering the interest rate.

Dudley Cooke (Trinity College Dublin) IS-LM Model 64 / 67

 The other (unusual) option is to attempt to
Manipulating
raise π eInflation Expectations
. For a given i, the real interest rate
will fall, boosting investment.

r, Interest Rate

LM (m0; π e = 0)

LM (m0; π e > 0)

r0 = i0 = 0
r1 = i0 − π e < 0
IS
y0 y y, Output

However, π e is endogenous. We have assumed

Dudley Cooke (Trinity College Dublin) IS-LM Model 65 / 67
 (contradicts the Fisher Hypothesis) as agents
Mundell-Tobin
change Effect
money for bonds (i.e. portfolio reallo-
cation), itself altering the interest rate.

r, Interest Rate

LM (m0; π e = 0)

LM (m0; π e > 0)

r0 = i0 − 0 A

r2 = i1 − π e C

r1 = i0 − π e > 0 B

IS

y0 y1 y, Output

Dudley Cooke (Trinity College Dublin) IS-LM Model 66 / 67

Roundup

Investment demand and Keynesian cross

IS curve
Money market and LM curve
Fiscal and monetary policy
Liquidity trap

Dudley Cooke (Trinity College Dublin) IS-LM Model 67 / 67