3 lnflation, Unemployment,

_J
1 and Monetary Rules

The aim of this chapter is to move beyond the models presented in Chapter 2 so that
we can analyse inflation and the monetary mies that are used by contemporary central
banks-i.e. when lhe central bank sets the interest rate to stabilize the economy around
an inflation target. ln order to do this, we first introduce Phillips curves and show how
they are derived from the labour market diagram. The key to understanding the Phillips
curves is that when unemployment is at the equilibrium levei, inflation is co nstant;
when unemployment is lower, infl ation goes up and when unemployment is higher,
inflation goes down. This suggests that it might be possible to mn the economy at lower
unemployment but with a higher rate of inflation. However, as we shall see, such a trade-
off is only possible in the short mn because wage setters are concerned about the real
wage and they will react when their real wage is eroded by inflation by pushing up their
wage claims. ln the Iong mn, there is no trade-offbetween unemployment and inflation:
this is shown by a vertical long-mn Phillips curve.
Central banks are now typically viewed as operating monetary policy through adjusting
the interest rate in order to keep the economy dose to its inflation target at the equilibrium
levei of output. Since inflation is only constant at equilibrium output, this appears logica l.
The reason the central bank plays an active role in managing the economy is because the
economy is subject to ali kinds of disturbances that shift either inflation away from its
target or output away from the equilibrium levei or both. These shocks o r disturbances
produce changes in inflation that are persistent and costly to elimina te so there is a case
for the central bank intervening to try to minimize the fluctuati ons.
lf we take as an example a rightward shift of the IS curve as th e result of an invest-
ment boom, this will push inflation up. The higher inflation will get incorpora ted intu
future wage- and price-setting decisions and will o nly be squeezed out of the econo my
by a period of unemploym ent above the equilibrium. The aim of the central bank is tu
minimize the cost to the economy of this shock in term s of the higher intlation caused by
it and the higher unemployment required to reduce it. lt will raise lhe int erest rate soas
to dampen aggregate demand and guide the economy back toward target intlation anti
equilibrium unemployment. Thi s behavi our of the central bank is oft en described as th e
use of a 'reaction function' or moneta ry policy mie.
By combining the Phillips curve, the JS curve and the central ban k's mo netary mi e, we
have the so-ca lled 3-equatio n mudei, IS-l'C-MR . This mudei is a textbook versiun of the
one used in contemporary central banks anel in more atlvanced treatments of moneta ry
macroeconomics. ln Chapt er S, mo netary policy is cxplored in grea ter depth and we
 68 THE MACROECONOMIC MODEL

. . . trate based monetary policy

mvestigate problems that can arise with the use of an mteres
rule. · · ·
ences of inflat10n mertia are
ln section 1, the Phillips curve is developed and the consequ k'
. t d The central ban s monetary
explained. ln section 2 the 3-equation model 1s presen e · d
rule is derived and the ;S-PC-MR model is used to analyse inflation ªnd aggregate emaod
shocks. We look at the relative costs of 'cold turkey , an d •gra dualist' disinflation . . .strate-
.
· ·
g1es. ln section 3 we Iook at how the inflat10n rate at t e h medium-run equ1hbnum 1s
determmed . . . . th e 1·nterest rate based monetary
under two contrasting monetary po 11C1es.
rule anda pohcy where the central bank targets the money supply · Section 4 shows how
·
inflation can be analysed using the standard IS/ LM model.
ln th1s· chapter, we model the supply s1de · usmg · 1mper
· fectly competitive. labour and
product markets and adopt the simplification that the PS curve is flat. ~his make~ the
examples more straightforward and rarely makes a difference to the basic mechamsms
under review but when it does, this is pointed out. An Appendix to the chapter presents
a famous model of inflation and unemployment using a competitive labour market and
a central bank that targets the growth of the money supply. This is Milton Friedman's
model of the 'natural rate of unemployment' and is provided so the reader can see the
parallels with the model presented in the chapter.

1 lnflation and Phillips curves

Inflation is the rate of change of prices, which means the price levei today reflects the
pattern of past inflation. Unless inflation is negative, which is called deflation, the price
levei does not fali. If P is today's price levei and P_1 is last period's price levei, then the
rate of inflation over the past year is rr:
_ P-P_ 1
'll'=-p--·
-1

Fig. 1.4 in Chapter I displays the inflation rates in severa! countries from the 1960s to
the present: whilst inflation in the past decade has been low at 2 to 3% in many of the
OECD countries, it was much higher often at rates between 10 and 20% in the 1970s.
Policy makers are concerned about inflation. High inflation tends also to be volatile
and this creates uncertainty and undermines the way in which prices convey informa-
tion. As we shall see, it is costly to reduce high inflation-i.e. an increase in unemploy-
ment is normally required to bring inflation down. These reasons lie behind the so-called
inflation-targeting regimes of central banks. For a fuller discussion of inflation and its
costs, see Chapter 5. The objective of this section is to explore the or·g· f • fl . d
1 ms o m ation an
inflationary pressure.

1.1 lnflation inertia

ln this chapter we introduce an important assumption used fo d . . .
. . . . . . r mo e 11 mg of mflation
expectat1ons and mflat10n mertia. The we1ght of evidence that h .
. . . . as accumulated on mfla-
twn dynam1cs m many countnes over the past decades sugge t h .
s s t at changes m output
 INFLATION , UNEMPLOYMENT, MONETARY RULES 69

(and employment) are followed by changes in inflation, which is summarized by saying

that output leads inflation. 1 Consistent with this evidence is a standard model in which
inflation depends on
• past inflation, n1 - 1,

• the gap between current unemployment and the ERU.

Some recent models have suggested to the contrary that inflation leads output-
i.e. inflation goes up not because output has gone up but rather, inflation goes up in
anticipation of a future increase in output. This means that the increase in inflation is
observed before the increase in output takes place. One such model is introduced in
section 1. 7. However there is no strong empirical evidence to support replacing the stan-
dard view with this one.
There are two broad interpretations of the past inflation term: one in terms of expecta-
tions and the other in terms of inertia. ln this book, we mainly use the second. However, a
common way of rationalizing the inclusion of past inflation is to assume tbat wage setters
expect inflation this period to continue to be what it turned out to be last period. This is
interpreted as an example of the adaptive formation ofexpectations. As applied to inflation,
expectations are formed adaptively when expected inflation this period is equal to last
period's expected rate of inflation plus a correction term to take account of the amount
by which last period's forecast was proved wrong:

(adaptive expectations)

where a is a positive constant less than or equal to one. lf past forecasting mistakes are
fully corrected so that a = 1, we have

(simple adaptive expectations)

This form of adaptive expectations therefore provides an interpretation of the lagged

inflation term, n _ 1, as a determinant of inflation. However, this is an unintelligent and
thus implausible way of forming expectations: why look entirely to the past when form-
ing a view about the future?
A more palatable and realistic interpretation ofwhy past inflation is included as a deter-
minant of inflation is in terms of the inertia that characterizes wage and price setting in
a complex economy. A common view is that wage setters incorporate past inflation into
their current money wage claim in order to make up for any erosion in living standards
(i.e. in the real wage) that has taken place since the previous wage round. We shall gen-
erally assume that wage setters are not able to incorporate expected future changes to
inflation in their current bargain. The theoretical and empirical debate about inflation
inertia and expectations hypotheses is reviewed in Chapter 15, where a more general
model that includes bot/1 past inflation and expected future inflation is developed.
ln this chapter, we work with the standard model in which the inflation inertia term
is defined as rr 1 = rr _ 1. Since 'inertial' inflation sounds rather clumsy, we often refer to

' For r~cent evidence, see Christi ano, Eichenbaum, and Evans (2005); also Estrella and Fuhrer (2002). and
Muellbauer and Nunziata (2004).
70 THE MACROECONOMIC MODEL
f inertia in its interpretation.
, hasize the role o
lagged inflation and denote it by rr to emp
Wehave:
7í = rr1 + a(y - r,)
+ a (y-y,).
current ~nflation inflat:~ (nertia
.outputt·gapaugmented Phillips curve)
(iner 1a-
l d inflation term is interpreted
This is the inertia-augmented Phillips curve. When th e agge . expectations (rrE = 71' _ 1)
as reflecting expected inflation on the assumption of adapuve '
· sed·
the term Expectations-augmented Phillips curve 1s u ·

rr •= rrE + a(y - y,)

rr L i + a(y - y,) .
current lnflation expected inflation out put gap
(expectations-augmented Phillips curve)

We now tum to extending the macro model by deriving the Phillips curves.

1.2 Deriving Phillips curves

As we have seen in Chapter 2, there is a unique unemployment rate at which the labour
market is in equilibrium. At this equilibrium, the WS and PS curves cross, which means
both wage and price setters are content with the prevailing real wage and have no incent-
ive to alter their behaviour. Our task here is to examine the link between unemploy-
ment and inflation: why is inflation constant at the equilibrium rate of unemployment
and why, when unemployment deviates from equilibrium, will this be accompanied by
changing inflation?
We begin with a numerical example in which inflation in the economy is 4% per year
and unemployment is attheERU: employment isE1 and the real wage is w 1 (see Fig. 3.1).
Money wages are set so that the WS curve represents the real wage. To keep the real
wage constant at w1 , wage setters require a money wage rise that will make up for the
rise in prices over the past year. Since prices have risen by 4% (i.e. lagged inflation is 4%),
money wage rises of 4% will therefore be set. Firms then set prices according to their
pricing rule. Since prices a_re setas a constant mark-up on unit Jabour costs (W / >.), this
depends on how much umt labour costs have risen. Since

P=(l+íl) - : ,

w -T
b.P b.W b.>.
P = ·
ln the sim pie
. case. where average labour productivity (>.) is
· constant, the only determinant
o f ch anges m umt 1abour costs is money wage changes:

-w ·
b.P b.W
p
 INFlATION, UNEMPLOYMENT, MONETARY RULES 71

w WScurve
p
Wz - -- -- - - -- -
: :1 _-_-_-_----~~-~p-_~,~-=~2~_%
____-r:--- ~ ::::::::~ _;J~~p5_:.'.CU~rv
~e

Eo E2 E

Figure 3.1 Upward pressure_on inflation when employment is above equilibrium employment;
downw ard pressure on 1nflat1on when employment is below equil ibrium employment

Table 3.1 Constant, rising, and falling inflation

lnflation (% per year) and employment

Period Employment Lagged inflation 'gap' Wage inflation Price inflation

o E1 4 o 4 4

Case 1: Constant inflation

1 E1 4 o 4 4

2 E1 4 o 4 4

3 E1 4 o 4 4

Case 2: Rising inflation

4 2 6 6
1 Ez
6 2 8 8
2 Ez
8 2 10 10
3 E2
Case 3: Falling inflation

4 -2 2 2
1 Eo
2 -2 o o
2 Eo
3 Eo o - 2 -2 - 2

Money wages have risen by 4% and so unit labour costs have risen by 4%. Therefore firms
must raise prices by 4% in order to keep their profit margin unchanged. We have deduced
that price and wage inflation remain unchanged at 4% per year and the economy remains
at point A (see Fig. 3.1). This is illustrated as Case 1 in Table 3.1 .
Suppose now that employment is at E2 , which is higher than the equilibrium and that
lagged in.flation is 4%. With employment of E2, the real wage on the WS curve lies 2%
above w (shown by the 'gap' in Fig. 3.1 and Table 3.1). Money wages will have to be
1
raised by 4% to keep the real wage unchanged; they will be raised by another 2% to take
72 THE MACROECONOMIC MODEL
th . . by 4% + 2% = 6%). With
e real wage up to W2 on the WS curve (money wages wtll nse t reserve their profit
a money wage rise of 6%, firms will increase prices by 6% so ~sfl o _P is higher than .
10
margin. Thus, we observe wage and price inflation of 6%-i.e. at!O~ h • bin
the . . wages and pnces ave nsen y
th prev,ous period. The other thing to note is that stnce . • ated by wage setters.
e sarne amount, the real wage is w 1 and not the W2 ant1ctp .
· h ·se in 10flat1on . from 4% to
U nemployment lower than the ERU is associated wtt ª n . .. .
6%. But inflation will not remain at 6% because the labour market is not m equthbnum.
Th al . 1wage of w 2 • When wages are
e re wage outcome was w 1 yet wage setters requ1re a rea
next set, the 'catch-up' wage increase is 6% to cover past inflation so money wages will
· b • ·11 b aised by 8%. We observe the
nse Y8% to take the real wage to w 2 • ln tum, pnces Wt e r . . ..
phenomenon that unemployment below the equilibrium rate is assoc1ated w1th nsmg
inflation (see Table 3.1). nd
If we take the converse case of unemployment above the ERU ª employment of
Eo, the sarne reasoning gives the result that inflation is falling. At h~gh unem~l~yment,
workers' power in the labour market is weakened, which is reflected 10 th e positive slope
of the WS curve. At employment of less than E,, the WS curve is below the PS curve.
ln such a situation, the increased competition for jobs means workers will not be able
to get money wage increases of the full 4% that would maintain their real wage. The
money wage increase set by wage setters is 4% minus the 'gap' of 2%, i.e. 2%. With their
costs rising by 2%, firms raise their prices by 2%, keeping their profit margins unchanged.
Thus inflation falls to 2% (see Table 3.1). Since wages and prices have increased by the
sarne amount, nothing happens to the real wage. ln the following period, the negative
gap between the WS curve and the prevailing real wage persists. Once again, the market
conditions mean that wage setters will be unable to maintain the real wage of Wt: the
money wage increase is 2% minus the 2% gap, i.e. money wages will not rise at ali. ln
period 3, as we can see in the table, money wages and prices fall by 2%. Unemployment
above the ERU is accompanied by falling inflation.
These results can be illustrated using a Phillips curve diagram. This has inflation on
the vertical axis and output on the horizontal one. Just like the SAS curves of Chapter 2,
which refered to the price levei not the inflation rate, the Phillips curves are intimately
related to the wage-setting and price-setting curves and are best understood when drawn
on a diagram directlybelow the WS/PS diagram (see Fig. 3.2).
ln reality, there is nota one-for-one relationship between a rise inoutput anda fali in
unemployment. When output rises, workers who have been kept on the pay-roll but have
not been fully employed (e.g. working shorter than normal hours) may be fully utilized
with the result that higher output does not-at least initially-entail a rise in employ-
ment. This is called labour hoarding. Second, even if employment rises, unemployment
does not necessarily fall if the new jobs are taken by those who were not previously in the
Jabour force. People of working age who are neither employed nor unemployed are called
economically inactive and the decision of whether or not to participate in the Jabour
market is responsive to economic conditions.
The combination of labour hoarding and changes in the labour force mean that a 1%
change inoutput growth above or below its trend tends to be associated with respectively
a fali or rise in the unemployment rate of less than 0.5 percentage points. This empír-
ica! relationship between changes in the growth rate relative to its trend and changes
 INFLATION , UNEMPLOYMENT, MONETARY RULES 73

w
p

2%
-2% PS curve

WScurve

Eo Ez E
Inflation (1T)
1
Vertical Phillips curve (VPC) ---+

Phillips curves:
PC (1T 1=6%)
PC(1T 1=4%)
PC (1T 1=2%)
PC (1T 1=0%)

Figure 3.2 Deriving the Phillips curves

in the unemployment rate is called Okun's Law. The responsiveness of unemployment

to changes in growth is lower in countries with tighter regulations on hiring and fir-
ing (as observed in many continental European countries) and with stronger traditions
of lifetime employment, such as is characteristic of Japan. For ease of exposition, we
shall nevertheless show changes in employment and unemployment moving one-for-
one with changes in output.
An inertia- or expectations-augmented Phillips curve is defined as a feasible set of infla-
tion and output pairs for a given rate of lagged inflation, 1r = 1r _ 1. We construct the
1

1
Phillips curves for the examples in Table 3.1. With 1r = 1r_ 1 = 4%, we have identified
three feasible inflation and output pairs: at E1 , 1r = 4%; at Ez, 1r = 6%; and at Eo, 1r = 2%.
This is the Phillips curve for 1r1 = 4% and we show it in Fig. 3.2. To emphasize that the
Phillips curve is only defined for a given 1r1 , it is labelled PC( 1r1 = 4%). Using the data in
2

Table 3.1, we can plot the other Phillips curves and they are shown in Fig. 3.2.

2 Note that in the alternative interpretation of the role of lagged inflation in the Phillips curves, expected

inflation is equal to lagged inílation (7rE = 7r _ i) and the fa mily of Phillips curves (e.g . PC(1l = 4%) ) is
described as 'expectatio ns augmented'.
74 THE MACROECONOMIC MODEL

Each Phillips curve is defined by two characteristics: d ti the height

. . ast inflation rate an . xes .
( 1) the Iagged inflation rate, whJCh 1s equal to th e P f utput assooated w1th the
. . b the levei o 0
of the Phillips curve on a vertJCal lme a ove
ERU ' and . 3 The Phillips curves w1
·11 b e steepe r 1"f
(2) the slope of the WS curve, which fixes 1ts slope.
the WS curve is steeper and vice versa.
·t ·s useful to express them
l 1
. f Ph 1ºIJ"pscurves,
1
As well as a diagrammatic representat10n
h
°
gued that mone
y wages increase by lagged
in equation formas in section 1.1. We ave ar a between the existing real wage
inflation plus an amount to dose the percentage g P p . es increase by the sarne
. the WS curve. nc
(1.e. on the PS curve) and the real wage on d ºfference between current
· function of the 1
amount as wages. The percentage gap is ª d fvity employment and
.· ·th constant pro uc 1 '
and equilibrium employment; in a dd1twn, wi . i·near expression:
.
output are proport1onal, . • • we sh a 11 use the s1mp 1e 1
so for s1mphoty

1r = 1r 1 + a(y - y,)
(3.1 )
7r 1'- 1 +a(y -y,)
current inflati on inflation in ertia outp ut ga p

· · o f outpu t from the medium-run equilibrium

where a is a positive constant. The deVIat10n .
should be thought of in percentage terms (th1s · 1s
· approxi·mately equal to .the d1fference
.
between y and y, if they are measured in logs). If output is above the me~mr:1-run e~m-
librium, then y - y, is positive and this will raise inflation above last penod s mflati~n.
Similarly if y _ y, is negative, inflation will fali below last period's inflation; only w1th
y = y, is inflation constant at last period's rate. We can see from this expression that
the Phillips curve shifts up or down whenever lagged inflation changes and that its slope
depends on a , which in tum reflects the slope of the WS curve.

1.3 Phillips's original curve

we now consider the interesting case where average inflation in the economy over many
years is zero and the economy is buffeted about by unpredictable or random shocks
to aggregate demand. By random, we mean that some shocks are positive and some are
negative so that output is sometimes above and sometimes below the ERU. if these shocks
are short-lived, it is reasonable to expect that wage setters view a temporary rise in prices
as a fleeting experience and do not incorporate past inflation into their wage claims. If
so, the economy would be observed at the points shown in Fig. 3.3. This is one way to
present the original 'Phillips curve' published by A. W. Phillips using data for the UK
4
between 1861 and 1957. Fig. 3.4 reproduces Phillips's plot for the period between 1861
and 1913. Note that since Phillips has unemployment on the horizontal axis measured
from Ieft to right, the Phillips curve is downward sloping.

J Note that if lhe PS curve was downward ~l~ping, the slope of the Phillips curve wo uld be sleepe r, refl ecling

lhe slope o f both the WS and lhe PS curves. l h IS m1rrors lhe discussion in Ch a pter 2 aboul lhe slo pe o f lhe SAS
curve. 4 Phillips ( 1958).
 INFLATION , UNEMPLOYMENT, MONETARV RULES 75

1T

X
6%

4%
e
X

X
2%

Figure 3,3 The 'original' Phillips curve

10
.,;
8
.,
"""'3: 6
.,
. .. .
>,

e:"'
o ., 4
-
E~
o o.
QJ
. • ... . . ..
... •. . ..
2
•-i..
e:
"'e.,
.e:: o .. . •• . +•
.
õ
"'
e,:
-2 . .
•
-4
o 3 3 4 5 6 7 8 9 10 11
Unemployment, %

Figure 3.4 Phillips' original curve : UK 1861-1913

Source: Economica, 1958.

1.4 Phillips's original curve may exist but it cannot be exploited

Having seen a Phillips curve of the original type, we ask the following question:

• Can the government use policy to shift the economy from its customary position at
point A to a higher levei of activity at point B (in Fig. 3.3)? The policy maker may be
tempted to doso in order to reduce unemployment at the cost of a small rise in
inflation.
Suppose that the government increases the money supply (to shift the LM cu rve to the
right). The interest rate falis and output is raised via higher investment spending. Unlike
the random shocks that shifted the economy from A to D to B to C to A etc. in the world
76 THE MACROECONOMIC MODEL
. d 1·fferent situation: the governrnent
of the original Phillips curve, we have here ªquite. . with 1ower unemployrnent
. . fers a s1tuat10n
1s seeking to keep output at y 2 because it pre nc for the economy to be near
nd
even if there is some inflation. Eventually, th e te e they will build the 2% p.a.
. O f workers-1.e.
point B is likely to figure in the calculattons . . f m zero to 2% p.a. At this
. d. fl fon w1U nse ro
mflation into their wage claims: expecte m ª 1 ment disappears because
. . fl t'10 n and unemp1oy
pomt, the long-run trade-off between tn a . h 1't olicy of holding output
. . . t ersists w1t s P
the Ph1lhps curve sh1fts up. lf the governmen P . Case 2 in Table 3.1 and the lower
at Y2, the Phillips curve will continue to sh1ft up as in .
.h • reasing inflat10n.
unemployment will be associated w1t ever-mc . . onomics called the
. · f something m ec
This example
.. provides an important . 11lustrat1on
. . ·° .
d by Philhps seeme d to suggest
Lucas cntique. s White the relationsh1p ongmally eSttmate . d
1 ment th1s co 11apse as soon
the existence of a trade-off between inflation and unemp oy '
. . e only existed because govern-
as the government tried to exploit it. A stable Ph111 1ps curv .
• 1 Th tusion that the pohcy maker
ments did not systematically try to make use of 1t. e cone
· 11· bove the levei of output at the
cannot choose any point other than one on a vertICa me a .
· · rh·11· curve This is a useful remmder
ERU, has led to the not10n of a vertical long-run 1 1ps ·
when analysing policy, and is shown in the Phillips curve diagram as VPC. 1t may be
noted, that if governments cease to try to run the economy at unemployment below the
ERU, the original Phillips curve of a stable trade-off may well reappear in the data.

1.5 Disinflation is costly

An important implication of the Phillips curve, 1r = 1r -1 +a(y-y,), is that if the authorities
wish to reduce the rate of inflation, there will be a cost as there will be a period in which
unemployment is above the ERU. We assume that the economy is in a situation in which
inflation is high and constant; unemployment is at the ERU. If inflation is to be lower than
it was last period, unemployment must be pushed up above the ERU . This is because-as
the Phillips curve shows-inflation is equal to what it was last period plus an amount that
depends on how far the economy is from the ERU. Remember that a slacker labour market
with higher unemployment means that the current real wage cannot be maintained and
this will trigger lower wage and price inflation. Another way to put this is that when
unemployment is above the ERU, there is a negative gap between the WS curve and the
PS curve.
From the Phillips curve equation,

7r=L1+ a (y-y,)
=> (1r - 'lr- t) = a (y - Ye)
lf(1r - 1r_1) < O,
then a (y - y,) < O
=> y < y,.

s This was one ol lhe contributions for which Robert Lucas was -1 d.
(Lucas (1976)) . ci e m lhe award of the Nobel prize in 1995
 INFLATION, UNEMPLOYMENT, MONETARY RULES 77

1T

VPC

8 ____ __________ _____ _!3_

7
F/
6 --- - - -- -- ----- - \ ----

5
F''.
4
3
1TT =2

Yc y

Figure 3.5 Disinflation is costly

We can also demonstrate the point using an example and a diagram. 6 ln Fig. 3.5, the
economy is at point B with high inflation of 8% (on PC(rr1 = 8)). lt is assu med that
the central bank wishes to reduce inflation to its target rate, which is 2%. The Phillips
curve shows the feasible inflation and output pairs, given last period's inflation, of 8%.
The only points on the curve with intlation below 8% are to the left of B, i.e. with high er
unemployment. With Phillips curves like this, disinflation will always be costly. lt should
be clear that if the influence of the past on wage setters could be wiped away, then the
possibility of a costless disinflation emerges: this amounts to saying that the policy maker
could shift the PC(rr1 = 8) down, e.g. to PC( rr 1 = 2). ln this case, the economy would
jump from B to A.
Returning to the Phillips curve PC(rr 1 = 8), we assume that the central bank has chosen
to raise unemployment to point F. lnflation falis to 6% anda new Phillips curve PC( rr1 = 6)
arises. The central bank can then choose a point on PC(rr 1 = 6), point F'. This leads to
a downward shift in the Phillips curve, which is not shown. Eventually, lhe ob jective of
inflation at 2% is achieved and the economy remains at the ERU.

1.6 Disinflation and central bank preferences

Although any point on the initial Phillips curve such asPC(rr 1 = 8) in Fig. 3.5 is feasible,
the question naturally arises as to which point along this curve the central bank would
choose when implementing a policy of disintlation. We assume that the aim of the central
bank is to get the economy to the ERU with an inflation rate of 2%. 7 AI one ex treme, the

l> Note that we have specified a linear Phillips curve, which is rcflccted in the diagrarn. ln St:'Cti on 2.6 we return
to questi on of the implications of linear as compared with convex Phillipscurves.
i Thi ssounds innocent enough but as weshall see in Chaptcr 5, if the central bank aim s for an un employmen t

rate below th e ERU, it will no t be able to get infl ation down to its target of rrr :;; 2'Hi. Th ls is lmown as the
inílation-bias problem of monetary policy.
78 THE MACROECONOMIC MODEL

1r VPC

- -]3_
8

6
5 PC(rr 1=3)
4
PC(rr 1=2)
3

e
y
Ye

Figure 3.6 Disinflation and central bank preferences

• Th' b ·ngs inflation down to the target

central
· bank could choose point C (see Fig. ·3.6).
· is n loyment as indicated by the low
m the next period but at the cost of a steep nse m unemp .
output at point C. Once inflation has fallen to 2%, the Phillips curve shifts to PC(~ = 2)
and the central bank can safely cut the interest rate to stabilize output at Y•· The economy
would move from B to C to A.
A less harsh approach would be for the central bank to choose point F as in the case
discussed above. The fali in inflation and the rise in unemployment in the first period
would be less than if C were chosen. We can see that as we change the preferences of the
central bank from a lower to a higher willingness to sacrifice a rise in unemployment in
order to get a given reduction in inflation, we move along the Phillips curve from B to
C. We summarize this by saying that the inflation aversion of the central bank rises as it
chooses a point closer to C.
The central bank's preference between a deviation of inflation from 1rT or of unem-
ployment from the ERU can be represented with indifference curves. The indifference
curves of two different central banks are shown in Fig. 3.6. The indifference curves of the
more inflation-averse central bank are flatter. This central bank chooses point D and the
other one chooses point F on PC( 1r = 8). The more inflation-averse central bank chooses
1

D because it is willing to sacrifice

. . a bigger increase in unemploymen t t o get ·m fl a t·10n
down by more.
.. When the Ph1lhps
. . curve shifts down , the cent ra I b an k c h ooses 1ts
• pre-
ferred
. . pos1t1on where the md1fference
. curve is tangential t o e new Ph1lhps curve an d
th . .
m th1s way the economy ad1usts to point A · The inflatio n averse-central bank guides the
economy down the path from D to A. ln exactly the sarne wa th . .
central bank with steeper indiffere . Y, e less mflat1on-averse
nce curves gmdes the econo d
F to A. For both central banks since th • my •own• the path rfromd
, e1r most preferred p ·t·
y = y, the indifference curves shrink t . OS! ion 1s w1th 1r = 1r an
' o a pomt at A. The d . .
indifference curves is explained in detail in Ch envation of the central bank
required here. apter S; ª graphical illustration is all that is
 INFLATION, UNEMPLOYMENT, MONETARY RULES 79

1.7 Costless disinflation and rational expectations

We noted above that if the influence of the past on wage setters was absent, it would be
possible for the economy to jump from point B to point A in Fig. 3.5 without any rise in
unemployment. What additional assumptions about the economy do we need to make
in order to eliminate the cost of disinflation?

• inflation inertia is absent. This means there are no nominal rigidities in the economy
in wage- or price-setting behaviour or institutions that operate to produce inflation
inertia. And adaptive expectations play no role in wage setting. Instead, the so-called
rational expectations hypothesis holds. Hence, we have

7r = 7rE + a(y - y,) +E

where E (called epsilon) is a random shock term.

• when the central bank announces a new low inflation target, 1r T, this is believed by ali
market participants. Another way to say this is that the central bank's policy
announcement is credible. For the target to be credible, it must be consistent with

7r = 7rT + E

i.e with inflation remaining constant at the target, apart from unforeseen shocks.

Rational expectations of inflation mean that the onlydifference between what agents in
the economy expect inflation to be and what it turns out to be is something random: they
do not make systematic errors. Another way to express this is that the agent's subjective
expectation of inflation is equal to the 'objective' expectation given ali the available
information about the structure of the economy and about policy that is available at the
time the expectation is formed. Because ali of the information available is used in forming
the expectation, there is no correlation between the error term and the information
available when the expectation is formed. When applied to inflation in our model, the
rational expectations hypothesis means that:

(rational expectations of inflation)

where E is the 'objective' expected value and the Phillips curve is now:

7r = 7rT + a(y - Yc) + E. (Phillips curve; rational expectations)

This says that with rational expectations, when output is at equilibrium, inflation is at
target apart from a random shock. We can express this another way, by rearranging as
80 TttE MACROECONOMIC MODEl

follows
l T )
y-y,= -(rr-rr -(
Q

(ir - ;l ).
=
(3.2)
Hence.y = y, + _!_ ( ir - rrE)
o inflation surprise

(Lucas surprise supply equation)

. · Ued the Lucas surprise supply

where f; (pronounced 'ksi') is a random error. Th 1s 1s ca .. .
equation after its originator Robert Lucas.8 Output only deviates from eqmhbnum when
· · ' h ·ce levei) 'Surprise' or unantici-
there IS a surpnse to inflation (or more generally to t e pn ·
· · · · h' t ut away from equilibrium because
pate d mflation IS normally mterpreted as pus mg ou P . . . .
a firm will find it difficult to know whether this is a rise in economy-wide mflatwn or m
the price of its product relative to the general price levei. If it is a rise in economy-wide
inflation, then the firm will not want to make a supply response; however, if the higher
inflation reflects a rise in demand for the firm's product relative to other firms, it will want
to increase its supply. ln general, it will not be possible for firms to distinguish with cer-
tainty between a general anda relative price change with the result that some increased
supply response will occur following an inflation 'surprise'.
We should note the possibility that wage and price setters may suspect that an increase
in inflation is the result of a deliberate attempt of the central bank or the govemment to
run the economy at unemployment below the equilibrium, so doubts may surface about
the credibility of the inflation target. This is another way of expressing the discussion
in section 1.4 about the Lucas critique and whether the Phillips curve trade-off can be
exploited by the government. ln such a case, we say that the inflation target of 2% is
not credible. Even with rational expectations and in the absence of inflation inertia,
there would not be a costless drop in inflation: rather, the government would have to
demonstrate its preference for lower inflation even at the cost of a rise in unemployment
by choosing a point to the left of B in Fig. 3.5. The issue of the credibility of monetary
policy is explored in more depth in Chapter 5.
We can summarize by saying that in a world where available information is used by
ali market participants to form their expectations (rational expectations), where there
are no sources of inflation inertia and where government commitment to the inflation
target is credible, then random shocks to inflation known as surprises will lead output
to deviate from its equilibrium levei. This means, in contrast to the standard model with
inflation inertia, that there will not even be temporary movements along the Phillips
curve in response to announced changes in government policy: disinflation will not be
costJy. The following distinction between the two approaches should also be noted:
• when using inertia-augmented Phillips curves , causality goes fro m a d ev1a
· t 10n
· ·tn
output from equilibrium (e.g. because of a change in aggregate demand) to a change
8
Lucas (1 972, 1975). For a very clear d iscussion , see Hoover (1 988).
 INFLATION, UNEMPLOYMENT, MONETARY RULES 81

in inflation relative to lagged inflation.

MD -+ liy relative to y, -+ Ci1r rela tive to 1r _ 1

• in the Lucas surprise supply equation, causality goes from a deviation in inflation
from its expected value to a change inoutput rela tive to equilibrium.

Cirr relative to 1r E -+ liy relative to y,.

The Lucas supply equation highlights the fact that under the assumptions of ratio-
nal expectations, policy credibility, and the absence of inflation inertia, since it is only
unanticipated changes in inflation that can affect output, systematic monetary policy
is ineffective in altering the levei of activity in the economy. There is also no need for
systematic policy because the economy returns directly to equilibrium once a shock to
inflation has disappeared. These are strong assumptions. As we shall see in the rest of
this chapter and in Chapter 5, because of the presence of inflation inertia, a variety of
economic disturbances shift the economy shift the economy away from equilibrium and
it does not return costlessly to equilibrium of its own accord. As a consequence, central
banks do engage in systematic monetary policy to stabilize the economy.

2 Monetary rules and the 3-equation 15-PC-MR model

We are now in a position to introduce the so-called 3-equation model. The three equations
are:
(1) the IS equation,
(2) the Phillips curve equation, and
(3) the monetary mie derived from the government or central bank's policy trade-off
between output and inflation.
ln this chapter, the emphasis is on gaining familiarity with the 3-equation model by
using diagrams. The equations are set out for the sake of clarity. ln Chapter 5, a more
extensive analysis of monetary policy that makes use of the equations is undertaken and
in Chapter 15, the further steps needed to develop the more sophisticated versions of
this model that have become the workhorse of modem monetary macroeconomists are
explained. Here we emphasize the development of a good intuitive feel for the adjustment
processes.
We have already shown the IS and the Phillips curves in diagrams and set them out
in explicit equation form. The monetary policy mie has not yet appeared explicitly in a
diagram oras an equation. However, as we shall see in the next section, the monetary
policy mie can be represented in the Phillips curve diagram by joining the points F, F', A
in Fig. 3.5. Similarly, in Fig. 3.6, ifwe join the points F, F',A, we have the monetary policy
rule for a less inflation-averse central bank and for a more inflation-averse central bank
by joining the points D, D', and A.