The Political Business Cycle After 25 Years_allan Drazen

8/9/2019 The Political Business Cycle After 25 Years_allan Drazen

1/44

The National ureau of Economic Research

The Political Business Cycle after 25 YearsAuthor(s): Allan DrazenSource: NBER Macroeconomics Annual, Vol. 15 (2000), pp. 75-117Published by: The University of Chicago PressStable URL: http://www.jstor.org/stable/3585387 .

Accessed: 06/11/2014 05:21

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

of scholarship. For more information about JSTOR, please contact [email protected].

.

The University of Chicago Press and The National Bureau of Economic Research are collaborating with

JSTOR to digitize, preserve and extend access to NBER Macroeconomics Annual.

http://www.jstor.org

This content downloaded from 152.118.148.226 on Thu, 6 Nov 2014 05:21:42 AMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=ucpresshttp://www.jstor.org/stable/3585387?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/3585387?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=ucpress


2/44

AllanDrazen

UNIVERSITY OF

MARYLAND,

HEBREW

UNIVERSITY OF

JERUSALEM,

AND NBER

The

Political usiness

ycle

after5 Years

1.

Introduction

A

quarter

f

a

century

as

passed

since the

nitial

utburst f formal

theoretical

nd

empirical

workon

political

usiness

cycles,

hat

s,

on

political

eterminantsf

macroeconomic

ycles.

On the

empirical

ide,

therewasKramer's1971) nfluentialtudy f conomic eterminantsf

U.S.

congressionaloting,

ollowed

y

thework

fTufte

1975, 978)

nd

Fair

1978).1

Nordhaus's

1975)

pioneering

ormal

model

of

the

political

business

ycle

PBC)

due

to

opportunisticre-electoral anipulation

as

published xactly wenty-five

ears

go.2

Soon

after,

ibbs

1977)

pre-

sented

model

of

partisan olicymakers

that

s,

policymakersaving

different acroeconomic

oals)

n

an

environmentimilar o that fthe

Nordhaus,

but

where these

partisan

ifferences ere

the

key

driving

force. erhaps sinfluentialn timulatingesearch as the 972 residen-

tial

lection

n

the

United

tates,

n

which

ncumbent ichard

ixon

was

justifiably

iewed s

engaging

n

significantre-electoral

anipulation.3

I

wishto thank

my

discussants,

lberto

lesina,

arl

Walsh,

nd

conference

articipants

and

seminar

articipants

t

theHebrew

University

f

Jerusalem

nd the

Bank f srael or

helpful

omments,

nd StefanHubrich

or

xtraordinarily

ble research

ssistance nd

many

ery

seful

iscussions. his

research as

supported

n

partby

theMaurice alk

Institute

or conomic

esearch,

ebrew

University

f

Jerusalem.

1.

Early

work

on connections etween

olitics

nd

fluctuations

n

economic

ctivity

s

reviewednKramer1971).2. Kalecki

1943)

presented

n

early

xplicit

model of the

PBC;

the

political

ature f

economic

luctuationsas

recognized y Schumpeter

1939)

n

his

study

f business

cycles.

imultaneously

ith

Nordhaus,

indbeck

1976)

presented

similar

dea;

soon

after,

cRae

1977)

lso

presented

formalmodel f hePBC.

3.

Rogoff

1988)

called

Nixon the

all-time ero of

political

usiness

ycles,"

t least n

contemporary

.S.

history.

ufte

1978)

begins

his

famous ook on the

PBC

with

quotation

rom

814,

A

Government

s

not

upported

hundredth

art

o

much

y

he


http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp


3/44

76

-

DRAZEN

Subsequent

to

this

flurry

f

research,

here

has been a

large

amount

of

further ork.

Theoreticalresearch

has

concentrated

n

making

both

op-

portunistic nd partisanmodels consistentwithvotersbehaving ratio-

nally,

both

in

forming xpectations

bout

future

olicy

and

in

voting

on

the basis of those

expectations.

The

success of

opportunistic re-electoral

manipulation

was rationalized

by

assuming

that

there s

imperfect

nfor-

mation about

an

incumbent's

ompetence,

with

expansionary

policy

be-

fore

n

election

taken as an

indicatorof

high competence,

as

in

the

pio-

neering

work of

Rogoff

1990)

and

Rogoff

nd Sibert

1988),

and

in

papers

that

followed.

A

partisan

postelectoral ycle

was

argued

to

be consistent

with rational

xpectations

n

the

mportant

work

by

Alesina

(1987, 1988).

On the

empirical

ide therehas been

extensive

work

testing

he

original

and

subsequent

models,

and

more

generally,

ooking

for

empirical

evi-

dence of

political

determinants f

business-cycle activity.

n

his

NBER

Macroeconomics

nnual

paper

in

1988,

Alesina

presented

n excellent

um-

mary

of much of

the work

up

to

thattime.

It is

over

a

decade

since

Alesina's

paper

was

published.

It

now seems

like

a

good

time

to look

at the

past twenty-five ears

of

work

and to

evaluate

the

state

of the literature.What

is

our

current

tate

of

under-

standingof thePBC, boththeoreticallynd empirically? n whatpoints

is there

agreement

nd on

what

points

is there till

significant

isagree-

ment? How well

do the models

explain

the

data?

What

does

existing

theory

s well

as data

suggest

about directions or

future

esearch?

The

short

answer

to

these

questions

is

that we have learned

quite

a

bit,

with

agreement

on

a

number of

ssues,

but

still

ignificant

isagree-

ment on

others. On the

empirical

side,

there are

a

number

of

clear

electoral

effects

n macroeconomic variables.

However,

at

least

for

the

opportunisticmodel in developed countries,there is much less hard

evidence

than both the theoreticalmodels

and the

conventional

wisdom

about

the

prevalence

of

"election-year

conomics"

would

suggest.

Al-

though

there s

wide

(but

not

universal)

agreement

hat

aggregate

eco-

nomic conditions

affect

lection outcomes

in

the United

States,

there

s

significant

isagreement

bout

whether here s

opportunistic

manipula-

tion

that

can

be observed

in

the macro

data.

There

is

a

clear

partisan

effect

n

the United

States

(as

well as

in

some

other

countries),

with

economicactivity einglower

in

the first

art

of

Republican

than

Demo-

cratic

dministrations,

ut

still

disagreement

bout

the

underlying

driv-

ing

mechanisms.

On

the theoretical

ide,

many

of the

leading

models

have

been

criticized

for

implausibility

f

key

assumptions.

Two

key

constant,

uniform,

uiet prosperity

f

the

country

s

by

those

damned

spurts

which

Pitt

used to

have

just

in

the

nick

of

time."




4/44

ThePoliticalusiness

ycle

fter

5 Years

?

77

points,

as

I will

discuss

below,

are:

first,

he

assumption

of

seemingly

irrational

ehavior

by

the

public

in

some of the

models; and, second,

the

reliance on monetary urprises s thedrivingforce.

The

purpose

of this

paper

is twofold: first o

present

a short

review

and

critical ssessment of the

existing

iterature,

oth

opportunistic

nd

partisan

models,

the

principal

aim

being

to

point

out

what

we

know

empirically

nd

to

what

extent

existing

models

explain

the

empirical

regularities.

A

principal

conclusion

is that

models based on

manipulat-

ing

the

economy

via

monetary

policy

are

unconvincing

both theoreti-

cally

and

empirically,

hile

explanations

based

on

fiscal

policy

conform

much

better to the data and form a

stronger

basis for

a

convincingtheoreticalmodel of electoral effects n economic outcomes.

Second,

I

present

new

model of

political ycles

based on

Rogoff's

1990)

model of

political

budget cycles,

extended

to include

monetary olicy.

The

model

is

the first o

incorporate

both

monetary

nd fiscal

policy

in a

rational

opportunistic

rameworkwith

separate monetary

nd

fiscal uthorities.4

This

separation

of

monetary

policy

from the direct control of

elected

officials

s crucial for

number of reasons. It is both in

sharp

contrast

o

existing

BC

models

and

far

more

nstitutionally

ealistic

han

the

policy-

makingstructurenthosemodels. Moreover, t s crucialto thenature of

the electoral

ycle,

which

depends

on the

nteraction etween the ncum-

bent

politician

who can

influence

fiscal

policy

and

an

independent

cen-

tralbank

that

controls

monetary

ggregates

and

interest

ates,

but

may

be

pressured

to

accommodate fiscal

shocks. We also

present

some non-

parametric

mpirical

vidence

n

favor

of the

active-fiscal,

assive-monetary

(AFPM)

model of

the

opportunistic

BC.

The

roadmap

for he

paper

is as

follows.

n

the next

section

quickly

review the opportunisticPBC model based on expansionarymonetary

shocks and

present

a

conceptual

assessment. In

Section 3

the

empirical

work on

this

approach

is

summarized. n

Section

4

I1

move on

to

partisan

models

driven

by

monetarypolicy,

both

the

original

Hibbs

model and

Alesina's rational

partisan

model.

In

Section 5 the

empirical

vidence

on

partisan

effects n

macroeconomic

outcomes is

reviewed. In

Section

6,

I

sum

up

what I

consider to

be the

conceptual

and

empirical

problems

with

monetary-based

PBC

models and

present

evidence

in

favor of

a

fiscal-based

model.

In

Section

7

recentwork

on

fiscal

cycles

n

develop-

ing

countries

s

summarized,

both

theoretical

xtensionsof

the

political

budget-cycle

model of

Rogoff

1990),

and

empirical

results

supporting

the

importance

of

fiscal

nfluences n

political

business

cycles

in

a

wide

4.

Rogoff

nd

Sibert

1988)

present

a

model of

fiscal-basedPBC with

nflation

ffects,

ut

where

both tax

and

inflation

olicy

are chosen

by

a

single

authority.




5/44

78

*

DRAZEN

range

fcountries.

n

Section

,

twocentral

uestions

elated

o a

fiscal-

based PBC model

re

posed;

the

nswers

resented

motivate he

AFPM

model of Section9, which combines lection-influencediscalpolicy

with

accommodatingmonetary olicy.

n

Section

10,

I

take

a look

at

some

data

for

the United States that are consistent

with the

AFPM

model,

nd

I

present

oncluding

omments.

2.

The

Monetarypportunistic

odel

Beginning

with Nordhaus's

1975)

model,

early

models of

the

PBC,

whether pportunisticrpartisan,were based on monetary olicy s

the

driving

orce.

Expansionary

monetary

olicy

ed

to a

temporary

increase

n

economic

ctivity,

ollowed

with a

lag,

by

an

increase

n

inflation.Models

differed

n

the motivation

f

policymakers,

s well

as

in the

modeling

f

xpectation

ormation,

nd

these

differences

ed to

very

ifferent

ypes

f

politically

nduced conomic

ycles.

Nonetheless,

it is

useful

o review

monetary-based

odels

as a

group

n

assessing

their uccess

n

explaining

PBC.

All

are based

on some

variant f

a

basic three-equationramework,ne equationrepresentinghe pol-

icymaker's

bjective,

ne

giving

he relation

etween

hanges

n the

rate

f

money rowth

r

nflationn

the

one

hand

and economic

ctivity

on

theother

a

Phillips

urve),

nd

finally,

ne

specifying

ow

expecta-

tions

of

inflation

re

formed.

We

begin

with

a

brief

eview

of these

models,

rief

ecause

we

simply

want o

point

ut ome

of

heir heoreti-

cal

shortcomings

nd to

summarize

mpirical

estsof

their

bility

o

explain

olitical

usiness

ycles.

his

review,

ontained

n Sections

, 3,

4,

and

5,

is

based

on

Chapter

of

Drazen

2000a),

where fuller

reat-

ment

may

be found.

2.1 NORDHAUS'S

PPORTUNISTIC

ODEL

Nodhaus's

model

was

meant o

show

that

f

voting

werebased on

eco-

nomic

erformance

n

the

ecent

ast

nd f

xpectations

f

nflation

ere

backward-looking,

n

opportunistic

ncumbent

ho controlled

onetary

policy

would

find

t

optimal

o

nduce n

inflation-unemployment

ycle

corresponding

o

the

length

f

his

term,

with

a boom

ust

before

n

electionnd a recessionfterwards.

The

structure

f

the

economy

s summarized

y

a

nonstochastic,

expectations-adjusted

hillips

urve,

yielding

n

inflation-output

rade-

off.

=

eTM-

1T,

(1)

t

'Tt

t




6/44

The olitical

usiness

ycle fter

5 Years

?

79

where

xt

s thedeviation

f actualfrom

otential

utput

nd where

he

monetary

uthority

s assumed

o

control he nflationate

ri.5

Theobjective f thepolicymakers tomaximize isprobabilityfre-

election.

oting

ehavior s

retrospective,

n that t

depends

n

economic

performance

nder

he ncumbent

n

the

past.

Economic

erformance

n

a

period

s measured

y

the

behavior f nflation

nd

unemployment,

o

thatvoter

dissatisfaction

n

any

period

can

be

represented y

a

loss

functionfthe

form

(x,

)2

(t

-

2

~2,

=

C

+

2(2)

2 2

where

r

s the

electorate's

arget

ateof

nflation,

is the

target

ate f

economic

ctivity

relative

o

potential utput),

nd

a

is

the

relative

weight

he electorate

uts

on

output

fluctuationselative

o inflation

fluctuations.

n

opportunisticolicymaker

ill

choose

the

policy

hat

attractsmost

voters,

o

that

hese

parameters

ould be

thought

f as

representing

he

preferences

fthemedian

voter.

In

thebasicmodel, ne thenposits retrospectiveoting unctionoranelection t theend of

period

,

oftheform:

N,

N

8sSts

+Et,

(2a)

s=O

yielding

he number f

votes

Nt

as

a

function f voters's

well-being,

where

N'(.)

<

0. The

exogenous ength

f ime

etween lections

s

T

+

1

periods, < 8< 1is thefactor ithwhich oters iscount ast conomic

performance

a

"forgetfulness

oefficient"),

nd

Et

s

a

mean-zero

tochas-

tic erm

elating

conomic

erformance

o electoral utcomes.

he elec-

toral

mechanisms

not

made

more

pecific.

he standard

pportunistic

PBC

model

assumes

that

8

is

small,

n

the

sense

that

ecent

conomic

performance

ountsfarmore

heavily

n

influencing

oter

hoices han

economic

erformance

n

the

moredistant

ast.

The

stochasticlement

is

added to allow for

he

possibility

f n

incumbent

osing

heelection.

To close the model one must pecifyheformationfexpectations.

5.

In

order

o

reproduce

he

regularity

f

high

nflation

agging

he

monetaryxpansion,

one must

ecouple

money rowth

nd nflation.

simple

ssumption

long

hese

ines

is that

nflationeflects

oney rowth

n

the

previous eriod,

hat

s,

r,

=

-U,_,,

with

,

being

he

monetaryuthority's

ontrol

ariable,

nd with

he

divergence

f

ctual rom

potential

utput

epending

n

the

differenceetween he

ctual ate f

money rowth

and the

economy-wide

xpected

ate f

money rowth

A.

See

Chapter

.3 of

Drazen

(2000a)

for

recise

etails.




7/44

80

DRAZEN

Crucial to

the main

results

f the Nordhausmodel is some

form

f

adaptive

xpectations.

standard

ormulation

f

adaptive

nflationx-

pectationss:

IT=-_1

+

"+0(,t-1

"/'-t-Ir),

(3)

where 0

is a

coefficient

etween

0 and

1

representing

he

speed

with

which

xpectationsdapt

to

past

nflation. hat s

crucial

n

theforma-

tion of

expectations

s that

wr

oes

not

depend

on the

expectation

f

future

olicies,

o

that

xpectations

renotrational.

t s this

haracteris-

tic

combined

with he absence

of

any

other onnections

etween

eri-

ods)

which

gives

the incumbent

olicymaker

n

exploitable

rade-off

between

nflation

nd

unemployment

n

the

attempt

o affect

lection

outcomes.

Voter ehavior

n

the

Nordhausmodel

s

backward-looking

n

two

dimensions:

otingdepends

on

past

incumbent

erformance,

nd ex-

pectations

f

moneygrowth epend

only

on

past

inflation

ates.

The

incumbent

policymaker

lected

at

t-3

chooses

inflation

ates

r7Tt_,

t-2,

Srt_1,

and

irt

o

maximize

is

expected

vote

in the nextelection.

This

simple tructureields hefollowingehavior f ncumbents howish

to

maximize

he

probability

f

remaining

n

office.

mmediatelyreced-

ing

an

election

he

government

timulates he

economy

ia

expansion-

ary

monetary olicy.

The levels

of

monetary xpansion

nd

economic

activity

re those

that

maximize oter atisfaction

n

an

election

eriod

taken

lone.

In the

period

mmediately

fter he

election,

he

govern-

ment

everses ourse.

t

engineers

recession

ia

contractionary

one-

tarypolicy

to

bring

down

inflationary

xpectations.

he

incumbent

keeps economic ctivityow to keep expected nflationow untilthe

period

immediately

efore

he next

election,

o thata

given

rate

of

economic

xpansion

induced

by

a

monetary

urprise)

an be obtained

at

a

relatively

ow

rateof

nflation.

n

thenext

lection

ycle,

he

same

behavior

s

repeated.

Hence,

we

have

a

simple

xample

n which

the

possibility

f

nfluencing

he

probability

f

re-election,

ombined

with

the structure

f

the

economy,

ields

cycle

n economic

ctivity.

The

exact

olution

may

be

found

n

any

treatment

f the

Nordhaus

model,

for

xample,

razen

2000a, . 233-236).]

2.2 CONCEPTUAL

RITIQUE

There

are

three

general

conceptual

riticisms

f the

basic

Nordhaus

model

s

a

tool

for

xplaining

PBC.

First,

t

assumes

hat he

president

controls

monetary

olicy,

n

assumption

hat s inconsistent

ith

the

independence

f

the

Federal

Reserve.

Although

ome

observers

rgue




8/44

The olitical usiness

ycle

fter

5Years

?

81

that

decisions

on

monetary olicy

n

the United States are

strongly

nflu-

enced

by

the executive

branch,

the notion

that the

president

can

easily

use monetary policy as an electoral tool does not fit the institutional

facts.

A

more subtle

argument

s that

an

independent

Federal Reserve

may

be

especially willing

to

accommodate the executive branch's

pres-

sures

for

monetary olicy during

election

years

n

orderto

prevent harp

movements

n

interest ates

which would lead the Fed to be criticized.

We

return o this

argument

below.

A

second,

more serious

problem

with

the

Nordhaus

model is its reli-

ance on irrational

ehavior on the

part

of

voters. Voters are

naive,

not

simply

n

the

way they

form

xpectations

f

nflation,

ut also

in

the

way

they

assess

government erformance.Any

voter

who

has

lived

through

an

election

cycle

in

Nordhaus's

world

should not be fooled into

voting

for

an

opportunistic,

manipulative policymaker.

He

will

know that

the

pre-election

period

of

low

inflation

nd

high

economic

activity

will

be

followed

by

a

postelection

period

of

both

high

inflation nd

high

unem-

ployment.

He should

therefore

unish

rather han reward an

incumbent

who

engages

in

pre-electoral

manipulation.

Finally,

and more

generally,

one

may

question

the central role

as-

signed tomoving along thePhillipscurveto reduce unemployment ia

inflation

urprises.

Fiscal

policy plays

no role

in

the

PBC

in

the

model,

though

transfers

nd

other

ypes

offiscal

policy

appear

to

play

an

impor-

tant

role

in

some

episodes

of

pre-electoral

olicy

manipulation.

3.

Empirical

ests

f

he

Nordhaus

odel

There have

been

many

econometric ests of the

monetary pportunistic

PBC, bothfor conomicoutcomes and forpolicy nstruments. he most

common form f

econometric

est of

these models

in

terms

of

outcomes

is

to

run an

autoregression

of an

economic

performance

measure

on

itself,

small

set of economic

variables,

and

political

dummies to

test a

specific

heory.

Consider a

regression

of

the form:

Yt

=

aY,

+

bo

+

1bIX

+

dPDUMt

+

Et,

(4)

i=1

j

where

Y

is

an

outcome

variable

such as

GDP,

the

Xj

are other

economic

variables that

may

also

affect

Y,

such

as world

economic

activity,

nd

PDUM is a

political

dummy

variable

or

set

of

variables)

meant

to

repre-

sent a

given

political

model.

The

autoregressive pecification

or

Yt

is

adopted

as a

parsimonious

representation

f

the

time-series

ehavior

of




9/44

82"

DRAZEN

Y,,

nstead

of

using

a structural odel. For

example,

s

a

test

ofthe

Nordhausmodelon

quarterly

ata,

Alesina

nd Roubini

1992),

Alesina,

Cohen, ndRoubini1992), ndAlesina,Roubini,nd Cohen 1997)use

a

dummy

ariable hat

quals

1

in

the

election

uarter

nd in the

T-1

quarters

efore

he

lection,

nd 0

otherwise,

here

T

may

qual

4, 6,

or

8. As

the measure

f economic

ctivity they

ake the

year-over-year

growth

ateof GNP or

an

unemployment

easure,

he exact

pecifica-

tion

depending

n the

model nd data set.

3.1 THE

EFFECT

F

ECONOMIC ONDITIONS N ELECTIONS

Prior

o

discussing

he ffect

f lections n macroeconomic

ariables,

ne

must onsider he effect feconomic onditions n elections.A crucial

assumption

n

the Nordhaus

model,

or

in

any

model

of

pre-electoral

manipulation,

s that

voters

vote on

the basis of

economic

variables.

Kramer

1971)

regressed

otesreceived

y

the

ncumbent

arty

n

U.S.

congressional

lections

n

two

measures

f

performance

n

the

year

of

the election-the

growth

ate

of

real

per

capita

ncome

nd the rate

of

inflation

n

that

year-and

found

hey

were both

significant

etermi-

nants

fvote otals.

he

mportance

f

conomic onditions

or

oting

n

congressionalelectionswas confirmed yTufte 1975).6

The

most

nfluential ork

was

probably

hat fFair

1978)

updated

n

Fiar

1982,

988)],

who found imilar

esults or

he

United tates.

n

his

original

rticle,

air

ooked

at

presidential

lections

rom 916

through

1976,

arguing

hat

f

voters

hold

the

party

hat

holds

the

presidency

accountable

or conomic

vents,

heir nfluence hould

be seen most

strongly

n

presidential

lections.

Fair

found

that

the

change

n

real

economic

ctivity

n

the

year

of the

election,

s measured ither

y

the

change n realpercapitaGNP or thechange nunemploymentn the

election

ear,

oes

appear

o have an

important

ffectn votes

for

resi-

dent.

Specifically,

1%

increase

n

the

growth

ate

ncreases

he ncum-

bent'svote total

y

about

1%.

(Further

vidence

uggests

t

may

be the

growth

f

real

per

capita

GNP

in

the second

and third

uarters

f the

election

year

that s

important,

ut data limitations

revent

air from

drawing

ny

definitiveonclusions

bout

what

part

f

the

election

ear

is most

mportant

n

determining

oter

ehavior.)

Given

the

growth

f

economic ctivity,thermeasures f macroeconomicerformanceon-

tribute

ittle;

he

most

mportant

f the othermeasures

s

the nflation

rate

n

the

two-year

eriod

before he

election,

s

measured

by

the

change

n

the

GNP

deflator. second

keyfinding

f

Fair's

s that oters

6.

Though

most

studies

confirm

he

basic

results,

Stigler

1973)

concluded

that

congres-

sional

election

results

are not affected

y

economic

fluctuations.

ee

also

Okun's

(1973)'

comment

on

Stigler,

s well as

Arcelus and

Meltzer

1975)

and Bloom

and

Price

1975).




10/44

The olitical

usiness

yclefter

5

Years

?

83

appear

to have

a

high

discount

ate n

past

economic

erformance;

hey

don't ookback

more han

year

r

two.7

Numerous ther rticles ind imilar esults nthe mportancefpre-

election onditions n

voting

patterns

n both

the UnitedStates

and

other

ountries.

ooking

t

voting

r

popularity

unctions,

ewis-Beck

(1988)

found hat

he

sort f results

hat

Kramer

nd

Fair

report

or he

United tateshold

n

Britain,

rance,

West

Germany,taly,

nd

Spain

as

well. Madsen

1980)

reported

imilar esults or

Denmark,

orway,

nd

Sweden.8We summarize his

s:

REGULARITY Aggregateconomic onditionseforen election,pecifically

percapita utput

r income

rowth

and

to a lesser xtent

nflation),

ave

significant

ffect

n

voting atterns

n theUnited tates

nd

other ountries.

3.2

ECONOMICACTIVITY

Numerous conometric

ests

rovide

ittle

upport

or he

political

ycle

in

economic

ctivity redicted

y

the

Nordhausmodel.Studies

for

he

United

States

began

with

McCallum's

1978)

study

of

unemployment

fluctuationsefore lections.Altand Chrystal1983)summarizearly

empirical

tudies as

showing striking

ack of

support,

point

re-

inforced

y

results ummarized

n

Alesina,

Roubini,

nd Cohen

1997).

Faustand Irons

1999),

using

more

ophisticatedechniques,

ome

to

a

similar

onclusion.

igure

,

showing

meanrates

fGNP

growth

season-

ally adjusted)by quarter

f

the

president's

erm

n

the United tates

from

948

o

1998,

llustrates

he

point.9

Similarly,

o evidencewas found n

developed

conomies utside he

United tates or Nordhaus-styleBC for nemploymentreconomic

growth

Paldam,

1979;Lewis-Beck,

988).

Alesina,

Roubini,

nd Cohen

7.

One should

distinguish

aggregate

from ndividual

economic conditions on

voting.

Lewis-Beck

1988)

argues

that

ndividuals

vote on the basis of national

economic

perfor-

mance

(sociotropicvoting)

rather han

their

own

personal

economic situation

"narrow

pocketbook"

voting).

8. What

about

the effect f

economic conditions on the

timing

of elections when

govern-

ments

can

call

early

elections?

to

(1990)

finds evidence that

governments

n

Japan

do

not

manipulate

policies

in

anticipation

f

upcoming

elections,

but

that

hey pportunisti-

cally manipulate

the

timing

of

elections to

take

advantage

of

autonomous

economic

expansions.

Specifically, igh growth

ignificantly

ncreasesthe

probability

f an elec-

tion,

while

high

inflation

significantly

educes

it.

Chowdhury

(1993)

reports

similar

resultsfor

ndia,

with

the

government

more

ikely

o

call

early

elections when economic

times

are

good.

On the

other

hand,

Alesina, Cohen,

and

Roubini

1993),

argue

that

for

sample

of 14

OECD

countrieswith

endogenous

election

timing,

here

s

no

evidence

of

such an

effect

n

countriesotherthan

Japan.

9. A

plot

of

median

growth

rates,

or of

other measures of

aggregate

economic

activity,

or

the United

States would tell a similar

tory.




11/44

84 DRAZEN

Figure

MEAN GNP GROWTH

RATE,

948-1998

10

8

S6

0

a.

4--

2

0

0 1

2

3

4

5 6

7

8

9 10

11

12

13

14

15

Quarter

fter

aking

ffice

(1997)

reject

n

opportunistic

ycle

n

real

activity

or

sample

of 18

OECD

countries

ver

he

period

1960-1993.10

Wesummarizehegeneral onsensus hat heopportunisticBC re-

ceives ittle

upport

n

the

pre-electoral

ehavior

f GNP

or

unemploy-

ment s:

REGULARITY

Theres no

significant

ncrease

n

aggregate

conomic

ctivity

prior

o lectionsn

either

he

United

tates

r

other ECD

countries.

3.3

INFLATION

Thepostelectoralncreasen nflationredictedythe

Nordhaus

model

receives

upport

n

some countries

nd

not

n

others.

Alesina,

Cohen,

and Roubini

1992)

and

Alesina,

Roubini,

nd Cohen

(1997)

test

for

political

ycle

n inflation

measured

s

the

growth

ateof theCPI

over

the

previous

quarters), sing

he

ame

data set and

methodology

hey

used

for

GNP

growth,

nd

defining political

dummy equal

to

1 in

the

election

uarter

nd

in

the3

quarters

ollowing

he

election,

nd 0 other-

wise.

n a

pooled

cross-section,

ime-series

egression,

hey

ind

highly

10. If aggregateeconomicperformances importantn determining heway people vote

and

governments

want to

win

re-election,

why

don't

we

observe

a

clear

opportunistic

PBC?

Lewis-Beck

1988)

argues

that t

s because it

s

exceedingly

hard to time

conomic

manipulation.

Monetary

and fiscal

policy

can

be used

only

with

great mprecision,

o

that

politicians

cannot

expect

to

time

the

aggregate

stimulus

to come

right

before

an

election,

while the

risks ssociated

with a

mistimed

xpansion

are

high.

Another

xpla-

nation

s that

opportunistic

oliticians arget

ransfers

o a fraction

f

voters

with minor

effect n

aggregate

economic

activity.

he AFPM model

in Section

9

includes

both of

these

possibilities.




12/44


ycle

fter

5 Years

?

85

Figure

MEAN

NFLATION

RATE

CPI),

1960-1978

8

6-

4-

2

0-

0

1 2

3 4 5 6 7 8 9 10

11 12 13 14 15

Quarter

fter

aking

ffice

significant

oefficient f

the correct

sign

on the

political

dummy;

in

the individual

country

regressions, they

find

the

coefficient

s

of the

correct

ign

in

almost

all

the

regressions,

nd

significant

t the

10% or

higherlevel forDenmark, France,

Germany,

taly,

nd New Zealand.

Overall,

they

conclude the

PBC

effect n

inflation s

widespread

across

OECD

countries

on

the

basis of their

pooled regression)

nd

on

a

much

stronger mpirical footing

han

the

effect

n

GNP

and

unemployment.

The evidence for

he United States

s

less

clear.

n

similar

ests

to those

described

above,

Alesina, Roubini,

and

Cohen

(1997)

reject

he existence

of a

postelectoral

urge

n

inflation ver

the

period

1947-1994.

However,

the behavior of

nflation fter lections

changed

over

this

ample period.

After 979there s no evidence of a political nflationycle,whichcorre-

sponds

to

the

timing

of

the

change

in

Federal

Reserve

policy

rules

in

1979.

(See,

for

xample,

the estimated

policy

rules

in

Clarida, Gali,

and

Gertler,

000.)

Prior o this

however,

there

s more

evidence

of

a

possible

postelectoral

ncrease

in

inflation.

his

is

consistent

with

other

studies,

and

is

illustrated

n

Figures

2 and

3,

showing

mean

annualized CPI

inflation

seasonally

adjusted)

from

1960

to

1979 vs.

1979 to

1998

by

quarter

of the

president's

term.

A

graph

for1948-1979

looks

very

simi-

larto 1960-1979,but thelatter s used forbetter omparabilitywith ater

figures.)

To

summarize:

REGULARITY

In

many

OECD

countriesheres

a clear

postelectoral

ncrease

in

inflation.

n

theUnited

tates,

here

s

evidence

f

uch

postelectoral

ncrease

in

inflation

rior

o

1979,

butno

evidence

hereafter.




13/44

86

-

DRAZEN

Figure

MEAN NFLATION

ATE

CPI),

979-1998

7

6

cc5

4

2

1

0

0

1

2 3 4

5 6 7 8 9

10 11 12 13

14

15

Quarter fter akingoffice

3.4 MONETARYNSTRUMENTS

Not

surprisingly,

he

results

or

xpansionarymonetary

olicy

before

elections

mirrorhose

for

nflationfter

lections.

sing

he ame

politi-

cal

dummy

hey

did for

nflation,

lesina,Cohen,

and

Roubini

1992)

find

significant

olitical

ffector he

yearly

M1

growth

ates

n

pooled

cross-section,ime-seriesegressionsntheirample fOECD countries,

with

money

rowth

einghigher

or he

year

o

year-and-a-half

efore

elections.

n

the

country

egressions,

he

results

re

ess

strong, hough

a

num'ber

f

countries

isplay

ignificant

ffects.

For heUnited

tates,

he

ensitivity

f he

nflationesults

o the ime

period

onsidered

s seen

n

money rowth

ates

s

well.

Alesina, ohen,

and Roubini

1992)

find

nlyvery

weak

evidence f

politicalmonetary

cycle

n the

postwar eriod,

conclusion einforced

n

Alesina,

Roubini,

andCohen 1997)for heperiod 949-1994.ncontrast, rier1989) nd

Beck

1987)

oth

find

ignificant

upport

or n

office-motivatedodel

f

monetaryolicy

n

the

United tates ver he

ubperiod

960-1980.

rier,

using

U.S.

quarterly

ata from

961

o

1982,

egresses

M1

growth

n its

previous

alue,

he

full-employment

eficit,

nd

a

political ummy

peci-

fied s a

fifteen-quarter

econd-degree

olynomial

istributed

ag

on

a

dummy

which akes value of

one

n

the

lection

uarter

nd zeroother-

wise.

The

polynomial

istributed

ag

s chosen

o

conserve

n

degrees

f

freedom.) e finds hat hetimingf n electionignificantlynfluences

money rowth,

ven

whenfluctuations

n

output,

nterest

ates,

nd the

deficit

re held

constant. eck

1987)

also

finds

political

ycle

n

the

money upply

n the

United

tates ver

he ame

period.

Figures

and

5

present

mean

M1

growth

ates

seasonally

djusted)by

quarter

f the

president's

erm ver he

periods

960-1979

nd 1979-1998.

nterestingly,

Beck

finds o similar

ycle

n

monetary

nstruments,

uch

as reserves r




14/44

The

olitical

usiness

ycle

fter

5Years

.

87

Figure

MEAN

M1

GROWTH

RATE,

960-1978

8

(f6-

0

4

X.

2

0

I

0

1 2 3

4

5

6

7 8

9 10 11 12 13 14 15

Quarter

fter

aking

ffice

Figure

MEAN

M1

GROWTH

RATE,

979-1998

12

10

8-

0o 6

4-

2

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Quarter

fter

aking

ffice

the

federal funds

rate,

a

point

made clear

in

Figure

6,

giving

the mean

federal

fundsrate

by quarter

of

termfrom 959 to

1998. The difference

n

results forthe behavior of

money

growth

and

instruments

f

monetary

controlwill be centralto our model of the PBC presentedbelow. We

summarize

these

results

as:

REGULARITY

Theres evidence

f

pre-electoral

ncrease

n

money

rowth

ratesn

many

ountries.

n

theUnited

tates,

heres

a

pre-electoralffectrom

1960

o

1980,

ut

one

hereafter.

heres

no

vidence

or

heUnited

tates

f

n

electoral

ycle

n the

ederal

unds

ate.




15/44

88

-

DRAZEN

Figure

MEAN FEDERAL

FUNDS

RATE,

959-1998

8

7

6

c5

'*

4

C3

2

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Quarter

fter

aking

ffice

4.

Monetary

artisan

odels

The

basic

partisan

model startswith the

observation hat

right-wing

nd

left-wing arties

have different

ositions

on

economic issues

and

hence

differentmacroeconomic

objectives.

In

terms of the

objective

function

(2), theyhave differentreferences ver inflation nd unemployment,

both

in

inflation

nd

unemployment

argets

nd the

relative dislike

of

inflation s.

unemployment.

4.1 THE

BASIC

HIBBSMODEL

The

partisan

PBC

model

was

introduced

by

Hibbs

(1977).

To

represent

the

difference

n

interests,

we

replace

the

social loss function

2)

by

a

partisan

oss

function:

(xt

-

Xi()2

+

(

-r

T-

)2

'

=

aI

+

(5)

2 2

for

party

,

where

i1r

s

party

's

target

ate

of

nflation,

j

is

party

's target

for conomic

activity,

nd

d

is the

relative

weight

put

on

output

fluctua-

tions relativeto

inflation luctuations

y

party

.

There are two

parties,

left-wing

arty,

enoted

L,

and a

right-wing arty,

denoted

R. The

two

parties are characterizedby the following possible differencesn their

objectives.

First,

the

left-wing

arty

may

have a

higher target

for eco-

nomic

activity

han the

right-wing arty.

Second,

the

left-wing

arty

may

assign

a

larger

ost

to

deviations

of

economic

activity

rom ts

target

level

than

to deviations

of nflation

rom he

target.

Finally,

he

eft-wing

partymay

have

a

higher

nflation

arget

han the

right-wing arty,

nde-

pendent

f

the effects

n

economic

activity

ia

the

Phillips

curve,

which




16/44

The

olitical

usiness

ycle

fter

5 Years

?

89

could reflect ther

effects f

nflation iewed

differently

y

the two

par-

ties.

To summarize

the difference etween the

parties:

5-L tR

aL

>- aR

(6)

To obtain the

partisan cycles,

at least one

of

these

must hold

with

strict

inequality.

Fluctuations

n

economic

activity

nduced

by

these

partisan

differ-

ences are

generated

in

the

basic Hibbs

model

by

movements

along

an

exploitable

Phillips

curve,

where it s

assumed,

as inthebasic Nordhaus

model,

that

expectations

are not rational.

Thus,

the

left-wing arty

will

pursue

a

more

expansionary

monetary policy throughout

ts

term.n

How

long

these effects ast

depend

on

the exact

specification

f

expecta-

tions.

In

an

adaptive expectations

framework,

he more

slowly

nflation

expectations

adjust

to

actual

inflation,

he

longer

will be

the

partisan

effect.

A

basic criticism f the

original

Hibbs

model

is

the same as the one

that was applied to the Nordhaus model, namelythat t relieson mis-

taken

expectations

of

what

policy

will

be

in

order

to

get

real

effects.

Hence,

to the extent hat

t

s assumed that

monetary olicy

s

used to

hit

partisan

unemployment

nd

growth

argets,

he

explanation

fthe

politi-

cal

business

cycle

s

unsatisfactory.

4.2

ALESINA'S

RATIONAL-PARTISAN ODEL

Alesina

(1987, 1988)

introduced

rational

expectations

nto

a

monetary-

based PBC, influenced ythecriticismf models based on an exploitable

Phillips

curve.

In

his

partisan

model with

rational

expectations,

only

surprise

inflation

ffects

utput,

leading

to

Alesina's

terming

he

ap-

proach

the

rational-partisan

odel. The

rational-partisan

model can

be

represented

by

a

similar

three-equation

model to that

used

by

Nord-

haus,

retaining

he

expectations-augmented

hillips

curve

1)

but

chang-

ing

the

other

two

components.

First,

following

Hibbs,

the

motivation

f

policymakers

s

quite

different han in

the

Nordhaus

model:

they

are

purely partisan,

with

no

opportunistic

motives

and

hence

no desire to

manipulate

outcomes. To

represent

the

difference

etween

economic

effects

n

the

early part

and

the latter

part

of

an

incumbent's term

of

office,

Alesina

divides a

term f

office nto two

periods

and

assumes that

11.

As

in

the

Nordhaus

model,

the

key

assumption

here

is

that,

n

spite

of the

Federal

Reserve's formal

autonomy

n

the United

States,

monetary

olicy

reflects he

adminis-

tration's

macroeconomic

goals.




17/44

90

DRAZEN

there

is an

election

every

other

period, say

at

t,

t+2,

t+4,

.

. . .

It

is

assumed

hat

party

ares

only

bout tsown term f

office,

o

that

he

objective unctionfparty at timetmaythenbe representedyan

extended

ersion

f

5),

namely

A

(xt

-

2j)2

+(Vt.-

jCj)2

Al

aI

+

(7)

2 2

+

ai

xt+l

+

_tj)2

(t-

1

)2

for

party ,

where

?7

nd

?j

are

the

partisan argets,

i

is

the

relative

weight ut

on

output

eviations

y

party

,

and

P

is the

discount

actor.

These

are

characterized,

s

in

theHibbs

model,

by

6)

above,

where,

n

order

o obtain

he

cycles

n

the

rational-partisan

odel,

t leastone of

the

nequalities

n

6)

must e strict.

The other rucial

hange,

relative o

both the Nordhaus

nd

Hibbs

models,

s

thatAlesina

eplaces

he

ssumption

f

daptive

xpectations

byrationalxpectations,o nstead f 3),expectednflation7r sgiven y

rt

=

Etl(Tt).

(8)

In

determining

he

evolution

f

nflation

nd

unemployment

uring

term

f

office,

ay

t

and

t+ 1,

the

key

variable

n

the

model s

expected

inflation

n those

periods,

his

xpectation

eing

formed

efore

he

lec-

tion

n

period

.

Conditional

n

expected

nflation

n each

half

erm,

he

partynpowerchooses tsoptimal olicy, ymaximizing7) subject o

(1).

We retain

he

assumption

rom

arliermodels

hat

he

government

has

perfect

ontrol

ver nflation.

n

turn,

xpectations

f

nflation

e-

pend

on

the

expectation

fwho

willwin the

upcoming

lection.

fout-

comes

were

fully

nown,

herewouldbe

no

cycle,

ince

party's

olicy

wouldbe

fully

nticipated

nd hence

have no

effect

n real

activity.

The existence

f

a

cycle

hus

depends

on

uncertainty

bout

election

outcomes.

xpected

nflation

or he half

erm fter

he elections

the

weighted

um

of

the

two

parties'policies,weighted y

the

probabilitythat achwillwin theelection,

amely,

7;

=

qLTr

+

(1

-

qL)rt,

(9)

where

L

s the

probability

hat

he

eft-wingarty

will

win the

election,

and

where

IrL

and

irr'

are the

optimal

policies

of the

two

parties

n

the




18/44

The Political usiness

ycle

fter

5 Years

?

91

first

alf

fthe

term,

which

depend

not

only

n

their

olicy references

(6),

but

also on

the

election

robability

L

itself,

s

optimal olicy

de-

pendson7rr.ince he eft-wingarty ollows morenflationaryolicy

once

in

office

han

the

right-wingarty,

xpected

nflations

between

these wo

values.

Hence,

there

s a

positive

nflation

urprise

f

he

eft-

wingparty

wins the

election,

mplying nemployment

elow the

natu-

ral

rate,

nd a

negative

nflation

urprise

f

he

eft-wing

arty

wins the

election,

mplying nemployment

bove thenatural ate.

n the

econd

half fa

president's

erm,

here

re

no

fluctuations

n

economic

ctivity,

as

the

dentity

f he

party

n

power

s knownwhencontracts

re

signed

(in

the first

art

of the

term).

n

contrast,

ibbs's

partisan

model

sug-

gests

higher

conomic

ctivity

n

eft-wing

dministrationshan n

right-

wing

dministrationsver he

ife fthe

erm.12

4.3

A

CONCEPTUAL

SSESSMENT F

THE

RATIONAL-PARTISANODEL

The

theoretical

tructure

f he

ational-partisan

odel aises number f

questions

bout

the

underlying

riving

orces.

irst,

nd

most

difficult,

theres

the

uestion

f

whetherhe

underlying

icroeconomic

tructure,

namely ominalwagecontractsigned efore lections,makes ense n

the

context f the

model. The

question

of

microfoundationss often

raised about

models

in

which

policymakersxploit

n

expectations-

augmented

hillips

urve,

utthe

mportance

f

lectoral ffects

ives

t

special

mportance

ere.

A

standard

rgument,

sed also

by

Alesina,

s

that

ominal

wage

contractsre

igned

t

discrete

ntervals,

here omi-

nal

wage

increases eflect

ationallynticipated

nflationt

thetime

he

contracts

signed,

o that

urprise

nflation

etween ontract

ates can

have realeffectsvenwhenagents rerational. he basicproblem,s

Rogoff

1988)

oints

ut,

s

that,

n theone

hand,

lectionsre

an

mpor-

tant

ourceof

fluctuations

ue

to

their utcomes

eing

ess than

fully

anticipated,

ut,

on the

other,

he

lection ate s

fully

nown.

The

mag-

nitude f

he

hanges

n

nflation

nd

unemployment

he

model s

meant

to

explain

re

sufficientlyarge

hat

here houldbe

a

large

tility ayoff

to

eliminating

he

uncertainty

hat

eads to these

luctuations.ut hat

s

easy

to do. To

theextent

here

s

a

significant

ffect

n

unemployment,

12. Hibbs (1994)

presents

uch a

theory

f

adjustment

of

partisan

objectives

contingent

n

economic

outcomes

and

learning,

which

predicts

that

unemployment

nd

inflation

outcomes across the

two

parties

may

diverge

more

in

the first

art

of their erms

han

in the

second,

though

not

because

of

uncertainty

bout

electoral

outcomes. The

key

to

Hibbs's model of

changing

objectives

and

to the result

on

time-varying

utcomes)

is

that

policymakers

re

uncertain about the

structure f

the

economy

and

the effects f

policies.

They

use

outcomes

to

refine heir eliefs about

attainable

argets,

eading

to a

feedback

from

utcomes to

partisan objectives

and

thus

policies.




19/44

92

-

DRAZEN

old

contracts

hould

be timed to

expire

and

the

signing

of

new

contracts

postponed

until

ust

after n

election,

o that

they

an

reflect he

election

results.Hence, themain drivingforceofthe model would seem to de-

pend

on behavior of

workers

nd

unions that

s less

than

rational,

not

in

the formation ftheir

xpectations er

se,

but n their

abor-supply

ehav-

ior.

A

simple

change

in

the

timing

f contract ehavior would eliminate

the

political ycle.

Garfinkel

nd

Glazer

(1994)

present

mpirical

vidence

that for abor contracts

f

less

than

two

years

signed

in a

presidential

election

year,

there s a clear

tendency

to

delay

the

signing

of

abor con-

tracts ntil after he election.13

A

second crucial

question

concerns the electoral

uncertainty

hat

drives the model. The

magnitude

of the

cycle

depends

on the

degree

of

electoral

uncertainty,

s

well

as

on the

difference

n

the

parties'

desired

inflation ates.

One

problem

s that

these

key

driving

forces

re

exoge-

nous.

Far more troublesome is the

predicted positive

correlationbe-

tween the

extent of

the

electoral

surprise

and the

size of

postelectoral

movements

in

real economic

activity.

As the

key probability

qL ap-

proaches

zero or

one,

the

magnitude

of the fluctuations

will

approach

zero,

with fluctuations

eing

maximal

(all

else

equal)

for

q'

=

1.

Hibbs

(1992), among others,has argued that thisprediction s not consistent

with

the

empirical

evidence for he United States. Consideration

of

ndi-

vidual elections reveals

the

problem.

For

example,

the outcome

of

the

1964

presidential

lection

s

probably

the closest

we have seen to

a

sure

thing

n the

postwar

era,

with

Lyndon

Johnson's

victory

widely

antici-

pated.

Yet the rate

of

real

GNP

growth

n the

first wo

years

of the

Johnson

administration

veraged

5.8%

per

year,

the

highest figure

of

any

Democratic

administration.

n

contrast,

mong postwar

Republican

victories hroughRegan's first lection,Nixon's victoryn 1968was the

closest

and least

certain,

but

corresponds

to the smallest

drop

in

real

output

n the critical

econd

year

of

the administration.

Alesina,

Roubini,

and Cohen

(1997,

Chapter

5)

construct

n

index

of

electoral

surprise

for

the

U.S.

presidential

elections

from

1948 to

1992,

with

Republican

victories

ntering

s

negative surprises.

They

use differ-

ent

variants

s an

explanatory

ariable in

a

real-GDP-growth

egression

of the

form

4)

and

find

that the coefficient

n the

surprise

variable

is

significantlyositive,meaning

that

arger

Democratic

Republican)

sur-

prises

imply

higher

lower)

postelection

eal

growth

rates.The construc-

13. Garfinkel

nd Glazer's results

may

be

interpreted

n two

ways.

One

is that

postpone-

ment

of contract

igning

indicates that electoral

uncertainty

s

important

n

forming

inflation

xpectations,

onsistent

with

the basic

thrust f

the

rational-partisan

model.

The

other s

that

n industrieswhere

this s

true,

ontract

igning

s

postponed,

under-

cutting

he

empirical

relevance

of

the main

driving

force

f the model.




20/44


ycle

fter

5 Years

93

tion of the variable is

complicated,

so that

t

is not

easy

to

see

why

the

results of the

regression

nd

of

the

simple

case

study

do not

agree.

The

relation ofpre-electoral ncertaintynd postelectoralfluctuations s an

important uestion

deserving

further esearch.

A final

question,

which can be

applied

to

all

the models discussed

so

far,

s

the central

role

assigned

to

moving along

the

Phillips

curve

to

reduce

unemployment

via inflation

urprises.

That

is,

even

though

real

effects f

monetary

policy

are

consistent

n this

approach

with

rational

expectations,

the

reliance on

monetary policy

as

the

driving

force

of

cycles

is

inconsistent

with

the

evidence

on the

important

ole of

fiscal

policy

n

PBCs.

We

return

o this

point

n

Section

6.

5.

Empirical

ests

f

artisan odels

The

partisan

PBC

has been

tested less

than

the

opportunistic

model.

There is

general

agreement

on the

existence of

partisan

effects

er

se,

especially

on

economic

activity.

owever,

there

s far

ess

consensus on

the

mechanism at

work.

5.1 ECONOMIC ACTIVITY

Perhaps

the

strongest

egularity

n

the U.S. data

was

first

ointed

out

by

Alesina

(1988),

with

Faust and Irons

(1999)

confirming

he

effect ver

a

longer

time

period

using

more

sophisticated

econometric

techniques:

For

the

United

States,

real

GDP

growth

s

substantially

higher

under

Democrats than

Republicans

in

years

2

and

3

of their

administrations.

Alesina,

Roubini,

and

Cohen

(1997)

report

hat

over the

period

from

he

first

uarter

of

1949

through

the second

quarter

of

1994,

growth

rates

duringDemocraticand during Republican administrationssharplydi-

verge

starting

bout

the third

quarter

after

he

election.

The

quarterly

growth

rate

averaged

over

Democratic

administrations

ises from

bout

3%

per

annum in

quarter

3 to about

6%

per

annum

by quarter

6 or

7 in

the

administration's

erm

of

office,

nd

falls from

he same

level

to

zero

by quarter

6

or

7

in

the

administration's

erm

averaged

over

Republican

administrations.

eal GDP

growth

ates

then

mprove

under

Republican

and

worsen

under

Democratic

administrations,

o that n

the

fourth

ear

of

the

administration,

he

growthperformance

nder

the

two

parties

s

identical.

Unemployment

hows

analogous

partisan

patterns

n

the

ex-

pected

direction.

Alesina,

Roubini,

and

Cohen

(1997)

present

more

for-

mal

econometric

ests for he

United

States to

confirm

his

result,

using

autoregressive

quations

like

(4)

in

quarterly

data from

1947:I

through

1993:IV

with a

political

dummy

that

equals

+1 in

the

first

part

of a

Republican

administration,

1 in

the

first

art

of a

Democratic

adminis-




21/44

94

*

DRAZEN

tration,

nd

0 otherwise.

hey report

esults avorable

o

the

rational

partisan

heory

or

ealGDP

growth

nd

for

nemployment.

hey

find

significantolitical ummy ver hewhole ife f n administration,ut

by dividing

he

variable

ntofirst

nd

second

halves of

the

administra-

tion,

hey

eject

Hibbs's

version f

the

partisan

heory.

hey

run imilar

tests

n

a

sample

of 18 OECD countries

ver

he

period

1960-1993,

lso

finding

upport

or he

rational-partisan

odel

and

lack

of

support

or

both

heHibbs

and theNordhausmodel.

Faust

nd Irons

1999)

find

imilar

artisan

ifferences

n

both

utput

growth

nd

unemployment,

hich

re

strongest

n thefirst

alf

f

the

term.

However, hey

find his

partisan

differenceemains

ven

after

controlling

or bservableconomic ariables nd for

olitical

ffectss

in

partisan

models,

uggesting

hat

hedata do

not

give support

o

any

partisan

model.

Graphs

of

quarter-after-inauguration

ffects

imilar

o

those

presented

ere

may

be

found

for

large

group

of variables.

he

key mpirical

egularity

n which

here

s wide

agreement

s

REGULARITY

There

s a

clear

partisanffect

n economic

ctivity

n

the

United

tates,

with

conomic

ctivity

eing

ignificantly

igher

nder emo-

cratshan epublicansnthe irst alff heirerms.

5.2 INFLATION

AND MONETARY

OLICY

There

re

partisan

ifferences

n inflation

as

measured

y

the

rate

of

change

n theConsumer

rice

ndex),

hough

hey

o notconform

im-

ply

to

the

partisan

heory,specially

he

rational-partisan

heory.

emo-

cratic

dministrations

ave

lower

verage

nflation

han

Republican

d-

ministrations

n

the first

alfof their

erms,

ut

that nflation

s

rising

underDemocratsndfallingnderRepublicans,findingeportedoth

by

Alesina,

Roubini,

nd Cohen

1997)

and

by

Faust

and Irons

1999).

Hence,

the

basic

inflation

ata

for

he United

tates

do not

support

monetaryartisan

model,

whereby

he

evel

f

nflation

hould

be

higher

under

Democrats

han

Republicans.

In

interpreting

hese

results,

lesina,Roubini,

nd Cohen

argue

hat

the

differences

ound

n

changes

n

inflation

ates re

consistent

ith

their

heory,

hough

he

rgument

s

only

partially

onvincing,

ince

he

rational-partisanheory

ased

on

the

expectations-augmentedhillipscurve s builton the rateof

nflation,

oton

changes

n thatrate.The

econometric

ests

for

nflation

ycles

n the UnitedStates

re

far

ess

favorable

o

partisan

models,

paralleling

he

nonparametric

ests

dis-

cussed

above.

Alesina,

Roubini,

nd Cohen

(1997)

find

hat fter

973

(and

the

move

to

floating

ates

fter he

collapse

ofBretton

Woods),

he

difference

n

average

nflation

ates etween

emocratic

nd

Republican




22/44

ThePolitical usiness

ycle fter

5 Years

?

95

administration

The Political Business Cycle After 25 Years_allan Drazen

Documents