Realignment and Macropartisanship
Author(s): Michael F. Meffert, Helmut Norpoth and Anirudh V. S. Ruhil
Source: The American Political Science Review, Vol. 95, No. 4 (Dec., 2001), pp. 953-962
Published by: American Political Science Association
Stable URL: http://www.jstor.org/stable/3117724
Accessed: 26-02-2015 17:21 UTC
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].
American Political Science Association is collaborating with JSTOR to digitize, preserve and extend access to The American
Political Science Review.
http://www.jstor.org
This content downloaded from 128.205.172.127 on Thu, 26 Feb 2015 17:21:53 UTC
All use subject to JSTOR Terms and Conditions
AmericanPoliticalScience Review
Vol. 95, No. 4 December2001
Macropartisanship
Realignment and
MICHAELF. MEFFERT Universityof Maryland, College Park
HELMUT NORPOTH State Universityof New York at Stony Brook
ANIRUDH V. S. RUHIL Universityof Illinois at Chicago
has exhibited
substantial
movement
in the U.S.
identification(macropartisanship)
Aggregate party
overthelast
electorate
halfcentury.Wecontendthata majorkeyto thatmovementis a rare,massive,
and enduringshift of the electoralequilibriumcommonlyknownas a partisanrealignment.The
research,whichis basedon time-seriesdata that employthe classicmeasurementof partyidentification,
showsthatthe 1980 electiontriggered
a systematicgrowthof Republicanidentificationthatcut deeplyinto
the overwhelmingDemocraticlead dating back to the New Deal realignment.Althoughshort-term
are responsiveto the elementsof everyday
fluctuationsin macropartisanship
politics,neitherpresidential
approvalnor consumersentimentis found responsible
for the 1980 shift.Realignmentsaside,macropartisanshipis guidedby a stable,not a continuouslymoving,equilibrium.
andIndependents
whoI
[A]llthosemillionsof Democrats
knoware lookingfor a causearoundwhichto rallyand
whichI believeI cangivethem.
-Ronald Reagan
he studyof partyidentification
has enteredthe
twilightzone. Gone seems to be the day when
partisanshipwas regardedas an immovableobject, impervious to the turbulence of politics and
guidingeverythingelse in the mannerof an irresistible
force. Instead,the view has dawnedthat this maybe an
object in perpetualmotion, buffetedby the prevailing
winds of politics and economics.There is no question
that macropartisanship-thebalance between Democrats and Republicansin the American electoratehas exhibitedsubstantialmovementover the last half
century.What is less certain is how seriouslyto take
such movementand where it originates.
We contendthat a rare,massive,and enduringshift
of the electoral equilibriumhas occurred,commonly
knownas a partisanrealignment.Althoughfamiliarto
studentsof elections,the conceptof realignmenthas so
far attractedlittle interest in effortsto come to grips
with macropartisanship.
This researchdemonstratesa
marked and lasting shift of the partisanequilibrium
after 1980.The election of RonaldReagantriggereda
systematicgrowthof Republicanidentification,which
cut deeplyinto the overwhelmingDemocraticlead that
dates back to the New Deal realignment.The new
GOP competitivenessenabled the RepublicanParty,
among other things, to crack the solid Democratic
control of the House of Representativesin the 1994
MichaelF. Meffertis AssistantProfessorof Communication,University of Maryland,College Park, MD 20742-7635(mmeffert@
wam.umd.edu).Helmut Norpoth is Professorof PoliticalScience,
State Universityof New York at Stony Brook, Stony Brook, NY
11794-4392([email protected]).
AnirudhV. S. Ruhil is
Assistant Professorof Political Science, Universityof Illinois at
Chicago,Chicago,IL 60607-7137([email protected]).
The authors thank Michael Kagay and his staff for making
availablethe partisanshipdata of the New YorkTimes/CBSNews
surveys.For comments,we are indebtedto George Krause,Brad
Lockerbie,Michael Lewis-Beck,Melissa J. Marschall,Marianne
Stewart,StephenWeatherford,the fouranonymousreferees,andthe
APSR editor.
midtermelection and to deliverunifiedcontrolin 2000
for the firsttime in nearlyfiftyyears.
The identificationof a realignmentdistinguishesour
model from recent efforts to come to grips with the
the moving equilibdynamics of macropartisanship:
rium (Erikson, MacKuen, and Stimson 1998; also
MacKuen,Erikson, and Stimson 1989), the constant
equilibrium(Green, Palmquist,and Schickler 1998),
and an intermediatetype of equilibriumprocessknown
as "fractional integration" (Box-Steffensmeierand
Smith 1998). Instead,we see a sharpbut rare regime
There is one equilibrium
changein macropartisanship.
before the Reaganpresidencyand anotherfollowingit.
Even thoughthe shift was not an overnightphenomenon but a gradualadjustment,it should not be mistaken for a moving equilibrium.In our model, the
equilibriumshift is a rare event, not a responseto the
flow of everydaypolitics.In large part,our accountof
macropartisanship
preservesthe AmericanVoterparaof
identification
digm party
(Campbellet al. 1960).
We begin by spelling out the case for a major
realignmentof partisanshipin the 1980sand then test
for its existencewith data from the National Election
Studies (NES). We next propose a dynamicmodel of
partisan realignmentto be tested with a version of
macropartisanshipthat matches the NES standard
(New YorkTimes/CBSNews surveys).We control for
the influence of everyday economic and political
forces, such as presidentialapprovalratingsand consumer confidence. Moreover, alternate specifications
of realignmentwill be considered.Finally,we discuss
how our findingsrelate to changingtheories of party
identification.
PARTISANREALIGNMENT
Studentsof Americanelectionscommonlydescribethe
relationshipbetweenthe majorpartiesas one of domination,with Republicansor Democratsenjoyingelectoral superiorityfor long stretches of time. One observer likens the majorityparty to the "sun"of the
electoral universe, orbited by the minority party
"moon"(Lubell 1952). Like planetaryconfigurations,
the majority-minorityalignment is not expected to
953
This content downloaded from 128.205.172.127 on Thu, 26 Feb 2015 17:21:53 UTC
All use subject to JSTOR Terms and Conditions
Realignmentand Macropartisanship
December2001
changeoften, and a shift is akin to a cosmicupheaval, took office as an avowed partisan and unvarnished
called a criticalor realigningelection (Burnham1970; ideologue"(Millerand Shanks1996, 168).
Fifth, the turmoil and discontent of the previous
Campbellet al. 1960;Key 1955). Sellers (1965) postufifteen years softened the electoralterrainfor a partilates an "equilibriumcycle"in Americanpolitics:Realignmentsthat lead to the ascendancyof one partyare san realignmentin 1980.Attachmentsto either major
followedby a returnto a more equal balanceuntil the
party had diminished through a process dubbed
arrivalof the next realignment.The elections of 1860, "dealignment"
(Converse1976;Nie, Verba,and Petrocik
1976; Norpoth and Rusk 1982). By 1980, a huge
1896, and 1932, accordingto most observers,can be
as
pool of voters with weak or no partisanattachments
designated realigning.
The causes and consequencesof realignmenthave waited to be tappedby a partywith a new agenda.
Hence, there was opportunity,motive, and weapon
received much scholarlyattention (Clubb, Flanigan,
in
1980 to fashion a partisanrealignmentout of an
and Zingale 1980; Sundquist 1973). A realignment
electoral
victory. Yet, Republicansin the past had
requires"the presenceof a great nationalcrisis,leadto win presidentialelectionsbut could not
found
ways
a
to
and
conflict
regardinggovernmentalpolicies
ing
the associationof the two majorpartieswith relatively crackthe solid lead of the DemocraticPartyin particlearlycontrastingprogramsfor its solution"(Camp- sanship.Would Reagan have more success?
bell 1966, 76). During these periods politics becomes
unusually ideological, and electoral defeats of the
CHANGEOF MACROPARTISANSHIP
majoritypartyare not simplydeviationsbroughton by As is common in
of macropartisanship,
the
the
temporary lapses in judgment or by the personal variable of interest isstudy
the
Democrats
of
percentage
appeal of a candidate. Instead, the vote begins to
reflecta majorreorientationof long-termpartyattach- amongthose with a majorpartyidentification(Demoments. Some argue that the reorientation occurs cratsand Republicans,regardlessof the strengthof the
identification).Independents,includingthose who prothrough conversionof partisanshipin the electorate fess
to lean towardeither party,are excluded.Macro(Eriksonand Tedin 1981;Sundquist1973), and others
simply registers the relative weight of
emphasize generational change or mobilization partisanship
Democratic
vis-a-visRepublicanidentification.For ex(Anderson 1979; Beck 1974; Campbell et al. 1960;
a
of 62 indicatesthat among 100 responample,
figure
Norpoth 1987). Moreover,the shift alters the compo- dents
a partisanidentification,there are 62
expressing
sition of partycoalitionsin the electorate,givingsome
Democrats
and
38
Republicans.The focus on partisans
groups a strongervoice and weakeningthat of others at the
is quite congenialfor
of
independents
expense
(Petrocik1981;Stanleyand Niemi 1995).
the examinationof a possiblerealignment,sincewe are
We contend that the 1980 presidentialelection, in
concernedwith shifts of the partisanbalance.
and of itself, exhibitedthe symptomsof a realignment
Figure 1 displays the percentage of Democrats
in five respects.First, for the firsttime since 1932, an
among party identifiersin all NES surveys,including
elected presidentwas defeated.Althoughthe majority midterm
elections,from 1952 to 1998. The imprintof
partyhadpreviouslylost controlof the WhiteHouse in the Reaganinnovationon the partisanbalancein the
"deviatingelections,"JimmyCarterwas repudiatedin
electorate is unmistakable.Between 1980 and 1988,
a landslide. He garnered even fewer votes in the
Democratic identificationfell from 64.5% to 56%,
Electoral College than did Hoover in 1932, whose whichis far the
by
singlelargestshift in half a century.
ousterusheredin the New Deal realignment(Burnham From 1952
to 1980, the DemocraticPartymaintained
1981;Pomper 1981).
the overwhelminglead in partisanshipforged in the
Second,the nationaleconomywas in severedistress, New Deal realignment,despite defeats in presidential
sufferingfrom a seeminglyincurablecombinationof
elections. If there exists a steady-state period for
high inflation and high unemployment(stagflation). macropartisanship,
it extendsfar beyond 1964, all the
This adversity,alongwithhumiliationin foreignpolicy, way to 1980. Although partisanidentificationunravcost Carterthe election (Markus1982). More impor- eled duringthe 1960s and 1970s, the relativeproportant, a dismal performanceof the economy under- tion of Democratsamong identifierswas virtuallythe
minedthe DemocraticParty'sNew Deal reputationas
same in the 1980 NES survey (64.5%) as in 1952
the partyof prosperity(Pomper 1981, 91).
(63.4%).Onlyin the aftermathof the 1980electiondid
Third,the winnerin 1980 promiseda radicaldepar- the Democraticshare fall sharply,and, while continuture from traditionalDemocraticideologyand a reso- ing to fluctuate,remainedat a lower level.
lutionof the nationalconditiondiagnosedas "malaise"
Table 1 providesa statisticalaccountof the change.
by the losing incumbent.Ideologicalzeal set Reagan For each presidential election for which there are
apart from the other victorious Republicansbefore sufficienttime points before and after in the NES
him. As the "most conservative president since ...
series,we determinedhow stronglymacropartisanship
1928"(Burnham1981, 99), Reagan would be able to
shifted after the election. The higher the correlation
count on a conservativetilt in the American public between the binarytime variableand macropartisan(Markus1982, 551).
ship, the strongeris the change;a negativesign indiFourth, Reagan was an intenselypartisanRepubli- cates a shift favoringthe RepublicanParty.Consider
can, determinedto make the GOP the majorityparty. the correlationsin the first column of Table 1, where
He was the "only presidentof the post-warera who no controlsare applied.It appearsthat after all elec954
This content downloaded from 128.205.172.127 on Thu, 26 Feb 2015 17:21:53 UTC
All use subject to JSTOR Terms and Conditions
AmericanPoliticalScience Review
Vol. 95, No. 4
FIGURE1. Macropartisanship, 1952-98 NES Surveys
70
65
60
55
50
'52
'60
'56
'54
'58
'64
'62
'72
'68
'66
'70
'76
'74
'80
'78
'88
'84
'82
'86
'92
'90
'96
'94
'98
Source: 1948-98 cumulativeNES File.
Note: The series representsthe percentageof respondentswith DemocraticPartyidentificationamong all respondentswitha majorpartyidentification
(excludingindependents).
tions from 1972 to 1988 the GOP made significant
but not all instances regains in macropartisanship,
main significantonce controls are applied. Only two
elections standout: 1980 and 1984.Clearly,something
after these
importanthappenedto macropartisanship
elections. The control for any of the other elections
that initiallyprovedsignificant(1972, 1976, and 1988)
does not make much differencefor the two Reagan
elections.At the same time, controllingfor either 1980
or 1984 rendersall others nonsignificant.It is safe to
TABLE1. Change in Macropartisanship after Presidential Elections
PresidentialElection
1956
1960
1964
1968
No Control
-.05
-.06
-.21
-.36
Correlationbetween Time and Macropartisanship
Controllingfor
1972
1976
1980
1984
.16
.16
.22
.17
.25
.24
.32
.25
.14
.12
.20
.11
.01
.00
.08
-.02
1972
-.44*
-
1976
-.50*
-.28
1980
-.68***
-.59**
-.55**
1984
1988
-.69***
-.58**
-.60**
-.48*
-.55**
-.42*
.02
-
.10
.17
-
-.31
-.24
-.05
1988
.08
.12
-.02
-.17
-.22
-.04
-.27
-.27
-.48*
-.46*
-.07
Source: 1948-98 cumulativeNES file.
Note: Entriesare correlationsbetween NES macropartisanship
and a binarytime variable.Foreach election,the time variablewas scored 0 for all time
points up to and includingthat election, 1 for all time points afterward.For example, the correlationof -.05 between the 1956 time variableand
indicatesthatmacropartisanship
was lower(less Democratic)afterthe 1956 electionthanbefore,althoughnot by muchor significantly
macropartisanship
so. *p
:-.05,
**p- .01, ***p- .001.
955
This content downloaded from 128.205.172.127 on Thu, 26 Feb 2015 17:21:53 UTC
All use subject to JSTOR Terms and Conditions
Realignmentand Macropartisanship
conclude that, in the nearly50 years of NES polling,
the only markedand lastingshift in macropartisanship
occurredin the wake of the Reagan elections.All the
others either made no dent or only appear to do so
because of their proximityto the Reagan elections.
These findingsare basedon an analysisthat includes
midtermas well as presidentialelections.Some of the
in Figure1, to
midtermreadingsfor macropartisanship
be sure, do not boost our claim. The first Reagan
midterm(1982) showsno declinewhatsoeverin Democraticpartisanshiprelativeto 1980.Had realignment
not yet begun? Or was this a temporarydelay in
Republicanconsolidationdue to an unusuallybad mix
of economic and political forces that obscured an
on-goingrealignment?The analysisbelow providesan
answer. Although some of the pro-Republicanshift
was left intact by the second Reagan midterm(1986)
and the Bush midterm(1990), they raise doubtsabout
realignmentclaims (Hurley 1991). Whateverthe economic and political fallout in those elections, the
between
pronouncedfluctuationin macropartisanship
on-yearand off-yearelections is a novelty.This is not
the placeto delveinto a resolutionof thiscomplication.
Sufficeit to saythatthe erraticbehaviorof partisanship
in midtermsnotwithstanding,
the Reaganyearsare the
only significantinstancein nearlyhalf a centurywhen
movedwith lastingconsequence.
macropartisanship
We agree that "between1980 and 1988, at least a
limited version of the long-heraldedpartisanrealignment took place" (Miller and Shanks 1996, 166; also
Miller 1991, 564). It was limited in that it did not
entirely sweep the Democratic Party from majority
status (amongidentifiers)in the Americanelectorate.
And it did not give the GOP immediatecontrolof both
houses of Congress,althoughthat finallymaterialized
in 1994. In some way, the 1980s realignmentis the
mirrorimageof the one in 1896,whichturneda highly
competitive balance between the two parties into a
decided advantage for the Republicans. Almost a
centurylater, a decided advantagefor the Democrats
that dated back to the New Deal was turned into a
highlycompetitivebalance.What type of realignment
model can explainsuch a markedshift of the partisan
equilibriumin the Americanelectorate?
A MODELOF REALIGNMENT
By definition,realignmentpresumesdurable,not transient, change. That does not mean that the new condition must prevail forever, but it must last a reasonably
long time. The concept of realignment is ambivalent,
however, about whether such change takes place at a
single critical moment or occurs more gradually. A
model of gradual change with permanent consequences has the form (minus the error term):
Yt = [o/(1 - 8)]Xt.
In the case at hand, Y denotes the macropartisanship
index and X the onset of realignment; o and 6 refer to
parameters that capture the dynamic of change (explained below). The gradual-permanent model is
December2001
rooted in transferfunctionanalysis(Box and Jenkins
1976;Box andTiao 1975;Clarke,Norpoth,and Whiteley 1998). Carmines and Stimson (1989) apply this
model with much success in their analysisof issueevolution realignments.In their "dynamicgrowth"
model, the centralfeaturesare as follows (p. 143):
The modelis dynamicbecauseit presumesthatat some
pointthe systemmovesfroma fairlystationary
steadystate periodto a fairlydramaticchange;the changeis
manifestedby a "criticalmoment"in the time series.
thechange-thedynamic
however,
Significantly,
growthdoesnotendwiththe criticalmoment.Instead,it continuesoveranextendedperiod,albeitata muchslowerplace.
In other words, a criticalshock graduallymoves a
systemfrom one steadystate to another.The effect is
enduring for the foreseeable future, but it is not
accomplishedin a single step. Two parametersgovern
this process:w capturesthe initial shock, and 8 summarizesthe aftershocks.Thinkof w as initiatingeither
a climbor a descentto a new equilibrium.Eitherway,
the initialeffectmaybe positive(a gain) or negative(a
loss), but the rate at which it continues to build
invariablyhas a positivesign. In the long run, the total
accumulationof change predicted by the gradualpermanentmodelwill add up to the ratiow/(1-6). It is
by such an amountthat the aftermathequilibriumwill
differfrom the one prevailingbefore the realignment.
Technically,the period of change is infinite, but in
practicaltermsmost of the changewill registerwithin
a foreseeabletime frame.
The requirementthat change is supposed to occur
over some longer period sets the gradual-permanent
model apartfromthe theoryof criticalelections,which
posits a relativelysudden conversion(Burnham1970;
Key 1955). Furthermore,the gradualmodel is not the
same as a secularrealignmentframework,whichviews
change as proceedingincrementally(Converse 1966;
Key 1959), on southern convergence).The gradualpermanentmodel capturesthe story of a rapidlyunfolding change during a "realigningelectoral era"
(Campbell1966,75). Even if the changeturnsout to be
abrupt,the model still proves useful. The aftershock
parameterwill dropto zero, and all the weightwill rest
on the initialshock.The differencebetween a gradual
and an abruptmodel is in the deltaparameter.We can
let the parameterestimatesettle this issue,butwe need
guidanceas to the onset of change. How certain is it
that the equilibriumshift commencedwith the 1980
election? Why not with the policy success of Reagan's
first year in office? Or with the recovery following the
1981-82 recession? Or with the "It's morning in America, again" reelection campaign of 1984?
To settle these questions, we need a more frequent
sounding of partisanship than the biennial NES but
one that nonetheless relies on the same instrument.
The NES party identification question, with its "generally" and "usually" qualifiers, aims to determine not
just any partisan but especially the "long-term identifier who is momentarily piqued at his own party, or
tempted to defect temporarily to vote for a charismatic
candidate of another party" (Converse 1976, 35).
956
This content downloaded from 128.205.172.127 on Thu, 26 Feb 2015 17:21:53 UTC
All use subject to JSTOR Terms and Conditions
AmericanPoliticalScience Review
Vol. 95, No. 4
TABLE2. AnnualShifts in
Macropartisanshipduringthe Reagan Era,
New York Times/CBS News Surveys
IndependentVariable
1980
1981
1982
1983
1984
1985
1986
1987
1988
Constant
Estimate
-6.72*** (1.15)
2.82 (1.48)
-1.35 (1.48)
-2.02 (1.48)
-4.22** (1.48)
1.07 (1.48)
.57 (1.48)
.20 (1.47)
-1.97 (1.09)
65.60***(.47)
Source:New YorkTimes/CBSNews polls, quarterlyaggregates, 1976:
01-1998:04 (N = 92).
Note: Entries are parameterestimates of an OLS regression (with
standarderrorsin parentheses)of New YorkTimes/CBSNews macropartisanshipon binarytime variables. Each of those variables was
scored 0 from the first quarterof 1976 to the fourthquarterof the
respectiveyearand 1 afterwardthroughthe fourthquarterof 1998. *p
.05, **p
***p-s .001.
<-.01,
Amongavailablesurveysthatprobethissentimentwith
some regularity,the New YorkTimes/CBSNews polls
offerby far the best alternative:"Generally
speaking,do
consider
a
yourself Republican,a Demoyou usually
crat, an Independent,or what" (emphasis added)?'
The New YorkTimes/CBSNews polls are conducted
often enoughduringa yearto permitquarterly,but not
monthly,aggregation.These polls are thus well suited
to constructa valid measureof traditionalpartyidentificationand to probe the dynamicpropertiesof the
partisanrealignment.In short,this is the best available
measure.Its advantagescomparedto the Gallupversion havenot escapedthe attentionof scholars(Green,
Palmquist,and Schickler 1998), although none have
probed for realignmenteffects.
Our effortsto identifythe onset of the realignment
focus on annual shifts in macropartisanshipduring
Reagan'spresidency,beginningwith the 1980election.
It wouldbe tedious to searchfor this type of needle in
a quarterlyhaystack;as a practicalshortcut,we limit
the searchto annualaggregates.Table 2 presentsthe
estimatesfor each of those years. Consider1982:The
estimate(-1.35) indicatesthat at the end of that year,
In contrast, the Gallup Poll measure ("In politics, as of today, do
you consider yourself a Republican, a Democrat, or an Independent?") probes a more short-term form of partisanship. The use of
this measure by MacKuen, Erikson, and Stimson (1989) has sparked
heated controversy (Abramson and Ostrom 1991, 1992, 1994;
Bishop, Tuchfarber, and Smith 1994; MacKuen, Erikson, and Stimson 1992). Considerable wording confusion reigns in the literature on
(macro)partisanship. MacKuen, Erikson, and Stimson (1989) did not
report the wording of the Gallup question in their original article and
later (1992) provided an incomplete version of the New York
Times/CBS News question, omitting the "usually" qualifier. Such
slippage, however, has an honored precedent. The American Voter
(Campbell et al. 1960, 122) itself left out "usually," although The
Voter Decides (Campbell, Gurin, and Miller 1954, 90, 217) got it
right.
fell by 1.35 percentagepoints (bemacropartisanship
coming less Democratic).Note that this is a partial
estimate that controls for the effects of all the other
annualshifts includedin the table. Althoughit points
in the right direction (change of macropartisanship
towardthe Republicanside), the estimatefor 1982falls
short of significance,as do many of the others in the
table.
Onlytwo entriesin Table 2 clearlypass the test: the
two presidentialyears. The 1980 election estimate
packsby far the largestwallop,cuttingthe Democratic
edge in macropartisanship
by 6.7 percentagepoints. If
there was a realignmentduringReagan'spresidency,it
began right after the election. Yet, althoughless impressive, the significant effect for 1984 suggests a
second shock.Is the one-shockrealignmentmodel too
simple?Did both Reagan elections shape the realignment, each perhapswith a rationaleand dynamicof its
own? In the next section,we test the rivalrealignment
specifications,but first we note the fundamentalsof
economics and politics introducedby the theory of
Those fundamentalspose serious
macropartisanship.
competitionfor any realignmentaccount of partisan
change.
THE DYNAMICSOF MACROPARTISANSHIP
Although many scholars concede that a partisanrealignmentof some form may have taken place in the
1980s,it is arguedthatthis onlyprovesthe sensitivityof
partisanshipto prevailingpoliticaland economic conditions. Contraryto the classicaltheory of partyidentification,whichleaveslittle, if any,roomfor systematic
short-termvariation in party identificationitself (as
opposed to short-termforces affectingthe vote choice
at any given election), the workof MacKuen,Erikson,
and Stimson(1989), in particular,documentsfluctuations in partisanshipthat are neither random nor
negligible.These variationsappearto go hand in hand
with political and economic circumstances(also Fiorina 1991;Weisbergand Kimball1995).
Withouta doubtthe incumbentin the WhiteHouse
is the most conspicuousstandard-bearerof his party.
His standingwith the mass public,among all possible
conditions,shouldbe able to translateinto supportfor
the party he represents. Few issues can match the
potentialof the economyto affectpartisanchoices, as
a vast domain of researchdocuments(e.g., Norpoth,
Lewis-Beck,and Lafay 1991). This is not the place to
sort out some of the numerousdisputesover the ways
in which the economyaffectspolitics,but the specifics
of the claims about macropartisanshipmerit close
attention.A criticalreplicationof those claims shows
that without numerouscontrol variablesto represent
the effectsof presidentialadministrationsand political
events macropartisanship
moves far less stronglywith
the economic and political winds (Green, Palmquist,
and Schickler 1998). Apparently,those factors exert
considerableinfluence, even though their role is not
explicatedin the theoryof macropartisanship.
In response to that challenge, Erikson, MacKuen,
and Stimson(1998)reformulatedtheirtheory.The new
957
This content downloaded from 128.205.172.127 on Thu, 26 Feb 2015 17:21:53 UTC
All use subject to JSTOR Terms and Conditions
Realignmentand Macropartisanship
December2001
TABLE3. Models of Macropartisanship
IndependentVariable
1980
Initialeffect
Dynamicrate
Model 1
Model 2
Model 3
Model4
-1.00***
(0.15)
0.91***
-1.03***
(0.15)
0.91***
-1.04***
(0.17)
0.90***
-1.05***
(0.15)
0.91***
(0.01)
(0.01)
(0.02)
(0.01)
1984
-1.78
Initialeffect
(1.86)
-
Dynamic rate
-0.76
(0.53)
PresidentialApprovalt-1
Initialeffect
0.11***
0.06***
0.06***
0.06***
(0.02)
(0.02)
(0.02)
(0.02)
0.61***
(0.12)
0.62***
(0.12)
0.65***
(0.12)
Dynamicrate
(PresidentialApprovalx Reagan)t_Initialeffect
(0.12)
(0.08)
Dynamic rate
-0.90***
(0.11)
ConsumerSentimentt_
-.02
Initialeffect
(0.02)
Dynamicrate
Constant
Adj.R2
Standarderrorof estimate
LBQ(20)
N
65.38***
(0.49)
0.82
2.17
15
92
65.60***
(0.46)
0.84
2.05
16
92
65.60***
(0.47)
0.84
2.06
14
92
65.60***
(0.46)
0.84
2.03
18
92
Source: New YorkTimes/CBSNews polls, quarterlyaggregates, 1976:01-1998:04(N = 92).
The variables
Note: Cell entriesare transferfunctionestimates (standarderrorsare in parentheses)for New YorkTimes/CBSNews macropartisanship.
"1980"and "1984"are binarytime variablesscored 0 from1976:01to 1980:04and 1984:04, respectively,and 1 afterwardthroughthe fourthquarterof
1998. Giventhe Democraticdirectionof the macropartisanshipmeasure, the observationsfor presidentialapprovaland consumer sentiment were
invertedaroundtheiroverallmeans for periodsof Republicancontrolof the presidency.***p- .001.
version argues that macropartisanshipis a moving
equilibrium,subjectto the accumulationof past political and economicshocks.The key factorsare still the
familiarvariablesof presidentialapprovalratingsand
consumersentiment,but the revisedhypothesisstates
that changesin approvaland sentiment,howeverslight
in the shortrun,cumulateto generatesubstantialshifts
in aggregate partisanship.At the same time, other
possibleinfluencesare dismissed:"Changesin macropartisanshipfrom other sourcesthat do not registerin
the approvaland sentiment series (such as election
campaignsand their aftermath)are mostly transient
and of little long-termconsequence"(Erikson,MacKuen,and Stimson1998,901). Seen fromthis point of
view, any realignmentin the 1980smusthave been the
work of an unusually popular president and good
economic times. Did short-termelements somehow
fuse into a long-termcompound?
Our firstattemptto answerthat questionfocuses on
a realignmentspecificationthat chooses as onset the
Reagan election in 1980. The variablesPresidential
Approvaland ConsumerConfidenceenter the macropartisanshipequationwith a lag of one quarterso as to
avoidproblemsof simultaneouscausation.2The results
in the firstcolumnof Table 3 providestrongevidence
for a dynamicmodel of realignment.Sparkedby the
1980 election, the decline of (Democratic)macropartisanshipproceedsat a rapidpace.The estimatefor the
initial effect (-1.0) may seem puny,3but with a dynamicrate of 91%,the aftershocksare nearlyas strong
as the initial tremor.Quarterafter quarter,they chip
awaythe Democraticedge and boost the Republican
Since the macropartisanship measure has a partisan direction
(percentage Democrat), both presidential approval and consumer
sentiment must be inverted when Republicans control the White
House. To do so we centered each series around its mean and
multiplied the observations for each of those variables by (-1)
whenever a Republican was in the White House.
3 This estimate should not be confused with the one for the 1980
election in Table 2. That was an estimate of the difference in average
macropartisanship before and after the election, without considering
a dynamic parameter.
2
958
This content downloaded from 128.205.172.127 on Thu, 26 Feb 2015 17:21:53 UTC
All use subject to JSTOR Terms and Conditions
AmericanPoliticalScience Review
Vol. 95, No. 4
FIGURE2. The Trajectory of Realignment
75
70
65
60
55
Observed Series
50 -
Predicted Series
45
)
i9
i9
Source: New YorkTimes/CBSNews polls.
index(1976:01-1998:04).The predictedseries is derivedfromthe
Note: The observed series representsquarterlyaggregates of the macropartisanship
parametersof the 1980 realignment(the initialeffect and dynamicrateof the "1980"variableof model 2 in Table3, withall other independentvariables
held constant). 5
ii
igi
i
iiiiiii3 iiiiiiiiiii,
&ii
idiii
iiiiiii
ii
11
side. It is reassuringto note that the prerealignment
in the New YorkTimes/CBS
level of macropartisanship
News series provides a constant (65.4%) that closely
matchesthe long-runaverageof Democraticidentifiers
in NES surveysup to that time (63.7%). Given the
virtually identical measurementinstruments,that is
how it shouldbe. The baselinefor testingthe dynamic
model of realignmentwith New YorkTimes/CBSNews
surveysis not a deviant case that would produce an
exaggeratedshift in the subsequentperiod.
Aside from the 1980 realignment,presidentialapproval ratings,as the estimate in the first column of
Table 3 indicates, also move macropartisanshipin a
highlysignificantfashion. Consumerconfidence,however, does not appear to have a direct effect. When
jointlyconsideredwith presidentialapproval,this economic measureloses out as a predictorof macropartisanship.Still, given the fairly strong correlation(.45)
between those two predictors,the economy has an
indirecteffecton macropartisanship,
mediatedby presidential approval.Economic assessments,not surprisingly, drive presidentialapprovalratingsand through
that vehicle manage to affect partisanship.Such an
indirectinfluencecould be parceledout of the overall
effect attributedto approval,but in the end that does
not give us any better purchaseon macropartisanship
than does presidentialapprovalalone. Therefore,the
consumerconfidencemeasurewas droppedfrom subsequent specifications.
To expandthe event horizonof the key short-term
variable,we added a dynamicfactor for presidential
approvalratings.Accordingto the resultspresentedin
the second column of Table 3, the dynamicrate for
approvalprovesstrongand significant.In otherwords,
the president'spartybenefits from high ratingsof his
job performance,and the effect cumulatesat an impressiverate.The question,however,is what that does
to the realignmentspecification.The answeris that the
1980 realignmentmodel continuesto performpowerfullyeven in the presenceof the economicandpolitical
fundamentals.Given the estimatefor the initial effect
(-1.03) and a dynamicrate of 91%, the cumulative
effect reaches 11.4 percentage points. That is the
long-term change in macropartisanshiptriggeredby
the Reaganvictoryin 1980. This shift in the electoral
equilibriumoccurred independentlyof the on-going
fall-out from presidentialperformanceas recordedin
job approvalratings.
Before we accept the conclusion that the 1980
election triggereda realignment,let us consider the
alternativeof a second shock in 1984, associatedwith
Reagan'sfourthyear in office, that culminatedin his
reelection.The resultsin the third column of Table 3
fail to supportthe two-shockmodel. The initial effect
for 1984 appearsquite strong and has the right sign
(-1.8), but it falls woefully short of significance,as
does the dynamicrate. At the same time, the specification for the 1980 election survivesunscathed.What
took the wind out of the apparenteffect for 1984?
We believe the 1980 election set in motion a shift in
the partisanequilibriumthat was well under way by
1984. Figure 2 maps the predicted trajectoryof the
shift, based on the parameterestimates for the 1980
election in Table 3 (secondcolumn).Accordingto that
959
This content downloaded from 128.205.172.127 on Thu, 26 Feb 2015 17:21:53 UTC
All use subject to JSTOR Terms and Conditions
Realignmentand Macropartisanship
trajectory,most of the realignmenteffectwas concentratedin Reagan'sfirstterm.The pace of changeslows
markedlyin his second term and flattens out afterward.4What obscures this long-termshift in the observedseriesof macropartisanship,
however,are shortterm setbacks for the Reagan administration,
especiallythe 1981-82 recession,which dropped the
president'sapprovalratingsto a record low in 1982.
We do not deny that presidential approval moves
macropartisanship,but that is short-term change
arounda baseline.As Reagan'sratingstumbledin his
first term, so did his party'sstanding,which led to an
above-equilibriumrise (given the Democratic direcindex. When Reagan's
tion) in the macropartisanship
popularityreboundedalong with the economyin late
inched
1983 and throughout1984, macropartisanship
back toward the equilibrium,even overshootingthe
mark.
By no means a trivialfactor for partisanship,presidentialapprovalratingsneverthelessdo not possessthe
leverage, even consideringtheir cumulativeeffect, to
generate realignment-sizeshifts in macropartisanship.
Given the parameterestimatesfor model 2 in Table 3
(an initial effect of .06 and a dynamicrate of .61), a
one-point change in presidentialapprovalproducesa
To
cumulativeshiftof just0.15%in macropartisanship.
engineer the realignmentof the 1980s-a long-term
shift of 11 pointsin macropartisanship-wouldrequire
a boost of nearly70 points in approvalratingsabove
the averagelevel, whichis a statisticalimpossibility.In
other words,the realignmentof the 1980swas not the
work of exceptional prosperityof a hugely popular
president, at least not in the way those conditions,
accordingto our estimates, normallyaffect partisanship.
But realignmentperiods are not ordinary times.
They are momentsof qualitativenot just quantitative
change.The upheavalsurroundingan electoralrealignment mayverywell be associatedwith mutationsin the
way factors such as presidentialpopularityand the
economy affect partisanship. Structural change in
model parameterscannot be ruled out (Maddalaand
Kim1998,chaps.13 and 15). In viewof the realignment
that unfoldedduringReagan'stenure, one could suppose thathis approvalratinghad a structurallydifferent
grip on partisanshipthanwas true underhis predecessors and successors.The same could be said for economic sentiment, but the failure of this variable to
prove significancein the presence of approvalprecludes its furtheruse.
A model that allows parametersto vary for the
This analysis relies on the New York Times/CBSNews measure for
frequent soundings of partisanship, but it should be noted that the
Gallup measure also registers an enduring shift in the wake of the
1980 election. An analysis of the Gallup series going back to 1953
and leading up to 1998 confirms a realignment effect that is similar in
nature to the one based on the New York Times/CBS News data for
1976-98. The magnitude of the immediate effect is smaller (-.49) in
the Gallup series, but the dynamic parameter (.94) is very similar,
and the cumulative shift still reaches -8.3 percentage points. At the
same time, the strong autocorrelation remaining in the Gallup series
testifies to the keener sensitivity of that survey item to short-term
disturbances.
4
December2001
Reaganyears,duringwhichmost of the realigningshift
accumulated,leads to the estimates shown in the
fourthcolumnof Table 3. To be sure, the initialeffect
estimate for Reagan's approval (.12) is impressive,
doublethe normaleffect (.06). The dynamicrate looks
formidableas well, but the negative sign hints at a
puzzlingtrajectoryof the cumulativeeffect.Thereis no
point in pondering this puzzle, however, since the
initialshocklackssignificance.It also can be noted that
the overall fit of the model including the Reagan
interactionis no betterthanthe modelwithout.Whichever way one looks at it, the resultsare disappointing
for the structural-changehypothesis.Presidentialapproval ratings did not move macropartisanshipany
more during the Reagan years than at other times
between 1976 and 1998.The realignmentin the 1980s
was not the work of especially strong presidential
approval.Of course, it did not hurt the GOP that, by
1984,economictimes were good and Reagan'sratings
were high, but those conditionsonly added a wiggle
arounda shiftingbaseline.The partisanequilibriumin
the American electorate had begun moving before
either prosperityor popularitytook hold.
THE NATUREOF PARTISANEQUILIBRIUM
The identificationof a realignmentmay go a long way
towardresolvinga knottypuzzleaboutmacropartisanship. On the one hand, it is undeniablethat aggregate
partisanshiphas not stood still duringthe half-century
span of observation.It has moved, and not just back
and fortharounda stablebaseline.On the other hand,
indexhas stayedwithina narrow
the macropartisanship
of
20
range barely percentagepoints.The latterpoints
to the existenceof equilibrium,but the formerraises
doubtsaboutit. Whatkindof equilibrium,if any,does
exhibit?
macropartisanship
The pioneersin this field make a strongcase for the
notion of a moving equilibrium(Erikson,MacKuen,
and Stimson 1998; MacKuen,Erikson, and Stimson
1989).The thrustof their argumentis that there is no
uniqueequilibriumto whichpartisanshipwill necessarily returnafter some departuredue to externalforces.
Instead,the equilibriumis whateveris producedby the
accumulationof politicalandeconomicconditions.In a
sense, politics and economicsconstantlymove macropartisanshipto a new equilibrium.
In contrast,Green, Palmquist,and Schickler(1998)
see greatmeritin the notion of a constantequilibrium.
They makea compellingargumentby askinghypothetically whether the fluctuationsin macropartisanship
would staywithinsuch narrowboundarieswithoutthe
gravitationalpull of a stableforce,whichtheybelieveis
very doubtful. Indeed, the bounded nature of the
movementis hardto reconcilewith
macropartisanship
the notion of an ever-changingequilibrium.We agree
that the familiarpoliticaland economicvariableslack
sufficientleverage to shift equilibrium,even with cumulativeeffectstaken into account.
To be sure, it is not easy to identifya constantlevel
over a half centuryeven though
for macropartisanship
its movementappearsto be bounded.This suggestsan
960
This content downloaded from 128.205.172.127 on Thu, 26 Feb 2015 17:21:53 UTC
All use subject to JSTOR Terms and Conditions
AmericanPoliticalScience Review
intermediatetype of equilibriumwhere shocks do not
dissipatequickly,yet theydo not remainat full strength
forever (Box-Steffensmeierand Smith 1996, 1998).
Such a dynamic fits uneasily into a zone between
autoregressiveadjustmentand randomdrift.If macropartisanshippossessesequilibrium,then it mustbe one
that can tolerate prolonged departuresor is of the
movingvariety.
We take the positionthatmacropartisanship
exhibits
a realignment-induced
At
any
equilibrium.
point there
may be movement,but the level of macropartisanship
is fairly stable until disruptedby conditionspowerful
enough to warrantthe label "realignment."Then the
baseline changes,and for a long time. The adjustment
is gradual,whichshouldnot be confusedwitha moving
equilibrium.The point of distinction is that in our
model an equilibriumshiftis rare,not a dailypossibility
dependingon the flow of everydaypolitics.If the 1980
realignmentis seen as a "structuralbreak"(Maddala
and Kim 1998),we can demonstratebeyonda reasonable doubtthatmacropartisanship
behavesas a stationtest,
aryprocess.Everyversionof the structural-break
by a wide margin,supportsthat conclusion."At the
same time, these test results cast doubt on the latest
technologyemployedby Erikson,MacKuen,and Stimson (1998) in defense of their macropartisanship
thea
if
The
of
structural
not
taken
break,
ory.
presence
into account, will lead to erroneous evidence for a
cointegratingrelationship(Maddalaand Kim 1998).
Ultimately,it is a backhandedconfirmationof the 1980
realignmentthat it makes such a powerfuldifference:
Ignore this intervention,and macropartisanshipis a
bundle of contradictions;take it into account,and we
have a well-behavedspecimenof equilibrium.Coming
to gripswith a realignmenteffectsolves a majorpuzzle
regardingthe natureof aggregatepartisanship.
CONCLUSION
Much of the movementin aggregatepartyidentification (macropartisanship)in the American electorate
duringthe last half centurycan be traced to a major
realignment.Our analysis,based on both the NES and
the New YorkTimes/CBSNews surveys,demonstratesa
substantialand lasting shift of the partisanbalance.
The 1980 election sparked a growth in Republican
identificationthat shrank the overwhelmingDemocratic lead dating from the New Deal realignment.
Although the GOP did not achieve outrightmajority
status, the new balance meant that the two major
parties would compete on nearly equal terms in na5 We examinedthree versionsof a modifiedunit-roottest with the
New YorkTimes/CBSNews macropartisanship
series, assuminga
structuralbreak in 1980. Version A (crash), which allows for a
one-timechangeof intercept,producesa test statisticof -5.26 (with
a criticalvalueof -3.77). VersionB (changinggrowth),whichallows
for a changeof slope,producesa test statisticof -5.81 (witha critical
valueof -3.80). VersionC (bothinterceptand slope changes)has a
test statisticof -5.23 (with a criticalvalue of -3.99). Thus, the
unit-root hypothesisis firmly rejected in all three instances. In
comparison,the standardunit-roottest withoutthe specificationof a
structuralbreakis a close call:a Dickey-Fullertest statisticof -2.79,
with a criticalvalue of -2.89.
Vol. 95, No. 4
tionalpolitics,whichpavedthe wayfor the Republican
victoryin the 1994congressionalelections.The success
of the Democratsin the presidentialcontests of 1992
and 1996 failed to reverse the Republicangains in
partisanship,let alonerestorethe pre-1980Democratic
lead.
The 1980 realignmentfits the patternof a dynamic
model with a gradual-permanent
set of parameters.A
criticalmomentinterruptsthe partisanequilibriumand
sets in motion an evolutionarychange that eventually
settles at a new steady state. This is no overnight
phenomenon.The processtriggeredby the 1980 election took most of Reagan'sfirst term before yielding
noticeablegains in Republicanpartisanship.However
much this suggests that the realignmentwas performance driven,our analysisshows otherwise.The key
factors-presidential approvalratings and consumer
sentiment-may generatewiggles aroundthe partisan
baseline, but they are not capable of triggeringan
avalanche.Reagan's popularitydid not generate the
1980s realignment.If popularitywere the key to such
shifts, the opportunitieswere far better under Eisenhower and even Nixon (before his disgrace).What is
more, given the relationshipbetween approvalratings
and partisanship,it would requireimpossible,not just
improbable,surgesin popularityto accountfor partisan realignments.
The key to the 1980realignmentlies in the ideological reorientationof nationalpolicyduringthe Reagan
years. Swept into office by an economic crisis of
uncommon severity, Reagan pursued a version of
conservatism (Reaganomics) that departed sharply
fromNew Deal liberalism.His partywas able to exploit
an ideological division in the electorate, in which
conservativesoutnumberedliberals. As in previous
realignments,the agenda of the federal government
underwenta change that to a large extent endured
under control by the opposition party. It would be
foolish, of course, to claim that the 1980 shift would
have taken hold even if the economy had not recovered, just as the New Deal realignmentmight have
been doomed had the Depression continued.Exactly
how the Reagan agendamay have helped bringabout
the 1980realignmentis a questionfor futureresearch.
REFERENCES
Abramson,PaulR., and CharlesW. Ostrom.1991."Macropartisanship: An EmpiricalReassessment."AmericanPolitical Science
Review85 (March):181-92.
Abramson,Paul R., and Charles W. Ostrom. 1992. "Question
Wordingand Macropartisanship:
Response."AmericanPolitical
ScienceReview86 (June):481-6.
Abramson,Paul R., and CharlesW. Ostrom,Jr. 1994. "Question
Formand ContextEffectsin the Measurementof Macropartisanship."AmericanPoliticalScienceReview88 (December):955-8.
Anderson, Kristi. 1979. The Creationof a DemocraticMajority:
1928-1936. Chicago:Universityof ChicagoPress.
Beck,PaulAllen. 1974."A SocializationTheoryof PartisanRealignment." In The Politics of Future Citizens:New Dimensionsin
Socialization,ed. RichardG. Niemi. San Francisco:Jossey-Bass.
Pp. 199-219.
Bishop, George F., Alfred J. Tuchfarber,and Andrew E. Smith.
1994."QuestionFormandContextEffectsin the Measurementof
961
This content downloaded from 128.205.172.127 on Thu, 26 Feb 2015 17:21:53 UTC
All use subject to JSTOR Terms and Conditions
Realignmentand Macropartisanship
December2001
A Republicanand Critique."
AmericanPolitPartisanship:ExperimentalTests of the Artifact Hypothesis."
"Macropartisanship:
ical ScienceReview92 (December):883-900.
AmericanPoliticalScienceReview88 (December):945-54.
Box, George E. P., and GwilymM. Jenkins. 1976. Time Series
Hurley, PatriciaA. 1991. "PartisanRepresentation,Realignment,
and the Senate in the 1980s."Journalof Politics53 (February):
and Control.San Francisco:Holden-Day.
Analysis:Forecasting
3-33.
Box,GeorgeE. P., andG. C. Tiao. 1975."Intervention
Analysiswith
Problems."Journal Key,V. 0. 1955."A Theoryof CriticalElections."Journalof Politics
Applicationsto Economicand Environmental
17 (February):3-18.
Association70 (March):70-9.
of theAmericanStatistical
Key, V. 0. 1959. "SecularRealignmentand the Party System."
Box-Steffensmeier,Janet M., and Ren6e M. Smith. 1996. "The
Journalof Politics21 (May):198-210.
AmericanPoliticalScience
Dynamicsof AggregatePartisanship."
Lubell,Samuel. 1952. The Futureof AmericanPolitics.New York:
Review90 (September):567-80.
Janet M., and Ren6e M. Smith. 1998."InvestiHarper& Brothers.
Box-Steffensmeier,
MichaelB., Robert S. Erikson,and JamesA. Stimson.
MacKuen,
Political
Fractional
Methods."
Integration
gating
DynamicsUsing
1989. "Macropartisanship."
AmericanPoliticalScienceReview83
AmericanJournalof PoliticalScience42 (April):661-89.
(December):1125-42.
Burnham,WalterDean. 1970.CriticalElectionsand theMainsprings
MacKuen,MichaelB., Robert S. Erikson,and JamesA. Stimson.
of AmericanPolitics.New York:W. W. Norton.
1992. "QuestionWordingand Macropartisanship."
AmericanPoBurnham,WalterDean. 1981."The1980Earthquake:Realignment,
liticalScienceReview86 (June):475-86.
Reaction,or What?"In TheHiddenElection:PoliticsandEconomMaddala,G. S., and In-Moo Kim. 1998. UnitRoots,Cointegration,
ics in the 1980 PresidentialCampaign,ed. ThomasFergusonand
and Structural
Change.Cambridge:CambridgeUniversityPress.
Joel Rogers.New York:PantheonBooks.Pp. 98-140.
Markus,GregoryB. 1982. "PoliticalAttitudesduringan Election
Campbell,Angus. 1966."A Classificationof the PresidentialElecYear:A Reporton the 1980NES PanelStudy."AmericanPolitical
tions."In Electionsand the PoliticalOrder,ed. Angus Campbell,
ScienceReview76 (September):538-60.
E.
E.
Warren
E.
and
Donald
Stokes.
New
Miller,
Philip Converse,
Miller, WarrenE. 1991. "PartyIdentification,Realignment,and
York:Wiley.Pp. 63-77.
PartyVoting:Backto Basics."AmericanPoliticalScienceReview85
Campbell,Angus,PhilipConverse,WarrenE. Miller,andDonaldE.
(June):557-68.
Stokes.1960.TheAmericanVoter.New York:JohnWiley& Sons.
WarrenE., and J. MerrillShanks.1996. TheNewAmerican
Miller,
Campbell,Angus, GeraldGurin,and WarrenE. Miller. 1954. The
Voter.Cambridge,MA: HarvardUniversityPress.
VoterDecides.Evanston:Row, Peterson.
Nie, NormanH., SidneyVerba, and John R. Petrocik.1976. The
Carmines,EdwardG., andJamesA. Stimson.1989.IssueEvolution:
ChangingAmericanVoter.Cambridge,MA: HarvardUniversity
Race and the Transformation
of AmericanPolitics.Princeton,NJ:
Press.
PrincetonUniversityPress.
Norpoth, Helmut. 1987. "Under Way and Here to Stay: Party
Clarke,HaroldD., HelmutNorpoth,and PaulWhiteley.1998."It's
Realignmentin the 1980s?"Public OpinionQuarterly51 (AuTime:
and
Political
Social
In
about
Modelling
Dynamics." Re376-91.
tumn):
in theSocialSciences:A Guideto NewApproaches,
searchStrategies
MichaelLewis-Beck,andJean-DominiqueLafay.
Helmut,
Norpoth,
ed. Elinor Scarbroughand Eric Tanenbaum.Oxford:Oxford
1991.EconomicsandPolitics:TheCalculusof Support.AnnArbor:
Press.
127-55.
University
Pp.
Universityof MichiganPress.
Clubb,JeromeM.,WilliamH. Flanigan,andNancyH. Zingale.1980.
Helmut,andJerroldRusk.1982."PartisanDealignmentin
Norpoth,
PartisanRealignments.
BeverlyHills, CA: Sage.
the AmericanElectorate:Itemizingthe Deductions."American
E.
of
1966.
"On
the
Political
Converse,Philip
Possibility Major
PoliticalScienceReview76 (September):522-37.
Realignmentsin the South."In Electionsand the PoliticalOrder, Petrocik,JohnR. 1981.PartyCoalitions:
andtheDecline
Realignments
ed. Angus Campbell,PhilipE. Converse,WarrenE. Miller,and
of theNew Deal PartySystem.Chicago:ChicagoUniversityPress.
Donald E. Stokes.New York:Wiley.Pp. 212-42.
Pomper, Gerald M. 1981. "The PresidentialElection." In The
Converse,Philip E. 1976. The Dynamicsof PartySupport.Beverly
Electionof 1980, ed. GeraldM. Pomper.Chatham,NJ: Chatham
Hills, CA: Sage.
House. Pp. 65-96.
Erikson,Robert S., MichaelB. MacKuen,and JamesA. Stimson.
Sellers, Charles.1965. "The EquilibriumCycle in Two-PartyPoliA Response to Green,
1998. "WhatMoves Macropartisanship?
tics."PublicOpinionQuarterly
29 (Spring):16-38.
Palmquist,and Schickler."AmericanPoliticalScienceReview92
Stanley,HaroldW., andRichardG. Niemi.1995."TheDemiseof the
New Deal Coalition:Partisanshipand Group Support, 1952(December):901-12.
Erikson, Robert S., and Kent L. Tedin. 1981. "The 1928-1936
1992."In Democracy's
Feast:Electionsin America,ed. HerbertF.
PartisanRealignment:The Casefor the ConversionHypothesis."
Weisberg.Chatham,NJ: ChathamHouse. Pp. 220-40.
AmericanPoliticalScienceReview75 (December):951-62.
Sundquist,JamesL. 1973.Dynamicsof thePartySystem.Washington,
DC: BrookingsInstitution.
Fiorina,MorrisP. 1991."Electionsand the Economyin the 1980s:
Short-and Long-TermEffects."In Politicsand Economicsin the
Weisberg,Herbert F., and David C. Kimball.1995. "Attitudinal
Correlatesof the 1992PresidentialVote: PartyIdentificationand
Eighties,ed. Alberto Alesina and Geoffrey Carliner.Chicago:
Feast:Electionsin America,ed. Herbert
ChicagoUniversityPress.Pp. 17-38.
Beyond."In Democracy's
F. Weisberg.Chatham,NJ: ChathamHouse. Pp. 72-111.
Green, Donald P., Bradley Palmquist,and Eric Schickler.1998.
962
This content downloaded from 128.205.172.127 on Thu, 26 Feb 2015 17:21:53 UTC
All use subject to JSTOR Terms and Conditions