KlTAKYUSHU SHIRITSU DAIGAKU HOU-SEIRQNSHU Journal ofLaw and Political Science.Vol.XLIV No.3・4March， 2017
E
x
p
l
a
i
n
i
n
g
t
h
e
F
a
t
e
o
f
E
t
h
n
i
c
M
i
n
o
r
i
t
y
P
a
r
t
i
e
s
:
An An
alysisacross Seventeen Post-Communist Countries.*Nakai
Ryo
Abstrac
t
.
Why do ethnically minority parties succeed or fail to win seatsワ To solve this puzzle， this article explores cross'national and inter-temporal differences in the post-communist new democracies This study argues thatgovernmental policypositions and thestrategic behaviourofethnicminorities determine the successand failure ofethnicparties and account for thevariancein theirstandings A spatialmodel demonstrates how ethnic minoritiesreact to other parties'policychanges， and statisticalanalysisusing seventeen Central and East European countries' data verifiesthishypothesis.
Keywords. Ethnopolitics; Voting Behaviour; Party Politics * An earlier versionofthis paper was presented atthelnternationalPoliticalSαenceAssociation World Cong問ss(held in 2009， Santiago， Chile). This岡 田archwas financially supported by JSPS Grant-in-Aid forYoungScientIst Start-up #21830123 -1 (203)ー
Explainingthe Fate ofEthnic MinorityParties{Nakai}
Introduction and method
Why do ethnicminority parties win seatsinsorne countries but not inothers?Why do ethnicparties rise or fall in thesame country? Sincemany Central and EasternEurope (CEE) countries democratised inthelate1980s and early 1990s， several ethnicminority groupshave had the chance tovoice theirown intereststhrough theelectoral system and party politics. These processesled to the emergence ofethnic parties inseveral countries.For example， the
Turkish minority's party has won a vastnumber ofseats inthe Bulgarianparliament. The Russian minority in Latvia formed political partiesand holdscapital citylocalgovernment. Contrastingly， many Gagauzians living in Moldova have failed toelecttheirown ethnic partiesto theMoldovan Parliament.The Russian minority in Estonia rarelysupport the ethnicRussian partiesnowadays despitetheir popularityamong Russian votersinthe 1990s.How should we interpret thesevariationsin political partysystemsbetween countriesand overtime?
Such questions have typicallybeen accounted forby classically institutionalor socio"economicexplanations， asMeguid (2005)pointed out. However， as many new studieshave observedthese typesof classicexplanationstrend limitations tosome extentand rarely solve a puzzle likethe one mentioned above.When institutional
and socio-economic explanationshow some limitations， actor-centric perspectivewillbe useful.Buildingon theoretical works that are more actor-oriented which assuming strategicbehaviour of political partiesand ethnicminorityvoters， this studyarguesthat setting policy relatedtonational integration， such aslanguagelaw and
Explaining the Fate ofEthnic Minority Parties{Nakai} franchising， areimportantvariablesinexplainingthe varianceof ethnic parties' representation. Inaddition， the rationality of ethnic minorities is a very important assumption tosupportthis hypothesis and explainthevariances acrosspartysystems This article consists of threeparts. The firstsection represents a theoretical reviewof theliterature， show limitationsofinstitutional and socio-economic explanation， and summarizerecent researches The second section offers an original hypothesis， deduced theoretically from a simply formalized model.In the third section， multi-regression analysesfor seventeen CEE countries verify thehypothesis and inspect the causal effects of several variables.Throughout， this paper use the term 'ethnic minority' frequently asa term thatincludes the broader conceptof ethnic minorities， national minoritiesand minority races， and define 'ethnic party' as 'the political party which insists on a particularethnicminoritygroup's interests as first among others'(Chandra 2009)
L
i
mitations ofInstitutionaland Societal ExplanationsIn a broader meaning， thesituation of ethnic parties could be characterized as a questionofpolitical partysystems and a question of politicalrepresentationby smallgroups.When we try to examine an ethnicparty'ssituation， theories ofpoliticalparty systems shouldbe mentioned. Classically， thevariationsin party systems have been observedmainly from two perspectives. 80me scholarshave emphasised theimpactof institutional effects， and other scholars have focused on socio・economicfactors.However， after the Third Wave of Democratization， these two classical typesof explanationshave showed theirlimitations insome extent
Explaining the Fate of Ethnic Minority Parties{Nakai}
SinceDuverger's(1954) seminal work， political scientists have been eager to answer why some states have many politicalparties intheirparliamentwhile othershave only a few.Institutionalists oftenhave stressed the significance of electoralrule， majoritarian or proportional， to explain thevarianceof partysystems.Other institutionalistswho have paid attention to the magnitude ofelectoral districts assume thatsmall minority partiestend togetseatsin a district wherever or whenever the electoralmagnitude ishigh However， several scholars (Moser 1999;Kitschelt 2000; Moser and Scheiner 2004， 2009)have argued that the electoral systems in post-communist democraciesdo notaffect in these areas as Duverger said， but rather institutional factors will affectitin a reverse way
(Meleshevich 2007)
Table 1 clearly exh尚北sthe limitationof institutional explanation
inCEE countries. Severalethnicpartieshave won seatseven their countries'electoral magnitudes have been low. Contrastingly， several ethnicminority groups have failed to represent itsethnic parties with highmagnitude electoral system. Even itis alsotrue that several ethnicpartieshave won seatsin high magnitude electoral system and vice versa. Insum， it is appropriate to argue that institutionalfactor (electoral magnitude intable 1) could notbe a single significantfactor to explainminority representations， or have weak effectatleast.
Explaining the Fate of Ethnic Minority Parties {Nakai}
Table 1:Ethnic Darties'reDresentations and institutionallsocietal factors electoral mag川tude(institutional factor)
ethnic party wm seats ethnic pa目y does not 、;vmseats ethnicpa代y 、;vmseats cthnic paはy does not 、;vinseats under 7日1 Russian (Estonia; bcfo問99)， Russian (Latvia)， Gennan (Poland)， Albanian削acedonia; after02)， Se出ian(Macedonia; after02)， Roma (Macedonia; after02)， Turkish (Bulgaria)， Hungarian(Romania) Russ凶n(Es旬nia;afteru3)， Slovak (Czech)， Roma (Czech)， Hungarian (Roma刷a)，ltalianand Hungarian (Sloveniai， Serbian (Croatiai， Roma (Croatia)， Albanian (Macedonia;before98)， Serbian (Ma田donia;bcforc98)， Roma (Macedonia; be伽 e98)，Roma 他 ulgaria)，Gennan (Romaniai， 竺旦金旦記込 over 701 Polish (Lithuania)， Hungarian (Slovakia)， Hungarian (Serbia)， Bosniak (Serbia)， Serbian(Montenegro)， Greek (Albania) Russian(Lithuania)， Roma (Slovakia)， Croat (Serbia)， Roma (Serbia)， Gagauzian ( 恥101dova)，Slavic peop1e (Moldova)， Russ叩 1 (Ukraine)， Crimean Tatar (Ukraine)
demographic composition (societal factor)
under 5 percent Gennan (Poland)， Hungarian (Scrbia)， Bosniak (Scrbia)， Grcck (Albania)， Serbian (Albania; aftc凶2) Roma (Czech)， Slovak (Czech)， Italian and Hungarian(Slovenia)2， Roma (Croatia)， Croat (Serbia)， Roma (Serbia)， Serbian (Macedonia; before98)， Gem聞 1 (Romania) 2， Gagauzian(Moldova)， SlavicPeople(Moldova)， Russian (Ukraine)， Crimean Tatar (Ukraine) Over 5 percent Russian(Estonia;bcforc99)， Russian (Latvia)， Polish (Lithuania)， Hungarian (剖ovakia)，Serbian ゆ1accd叩陥)，Albanian (Macedonia; after02)， Roma (Macedonia; afte同2)， Turkish(Bulgaria)， Hungarian (Romania)， Russian (Lithuania)， Russian(Estonia;after03)， Roma (Slovakia)， Roma(Hung町 )，Serbリan (Croatia)2， Albanian (Macedonia; befo陀 98)， Roma (Macedonia; before98)， Roma (Bulgaria)， 主 旦笠旦 出 弘 1. an average of electoralmag印刷deinCEE countriesis around 70 2.takinga seat only in ethnicminority specialmandate seats isnotsorted out as “winning seats" here Source:Cordelland 、，volffeds. 2口04;Bugajski 2005; and each state centra1 electoralcommittee's webpage -5 (207)ー
Explaining the Fate of Ethnic Minority Parties{Nakai}
Other scholars have treatedsocio・economicvarianceas themost
important factor. The brilliant work ofLipsetand Rokkan (1967)， thegodfathers of the socio-economicperspectivein party-system formation， has suggestedthat social cleavagesdefine the situation inparliament.
I
f
a countryhas many social cleavages， economic， orclass struggles willyielda polarisedparliament. Works that emphasisethe special role ofethnicvoting and see an ethnic line as an essentialcleavage are also categorisedhere. Essentialists have emphasisedthatethnic minoritiesfeelthe greatest pleasure invotingitself， notby receivingsome political and material benefits through an election， because ethnic minoritiesplace a special emphasis on theeffusionand the manifestation of their identity Socio-economic explanations also have not accounted for post-communist Europe countries and show limitations.We easily understandthat demographic intensities ofethnicminorities have nevermattered forethnicparties' representation asshown in table1.Moreover， eventhough all CEE countries have some differences
intheirhistoriesand international relations， butmostof them have facedsimilar politicaltasks (Geddes 1995; Millard 2004). Most ofthem have faced economic privatization， democratization， the need to rebuild the nation， pensionreformand debates concerning whether ornotto join the EU. Whitefield(2002) asserted that CEE countries generally had dealtwithsimilarpolitical tasksdespitetheir unique historicaland socio-economic contexts. He inspected 12 CEE countries' 'political cleavages' and presented10of them thathad ethnic problems ascleavages. However， their partysystems as related to ethnic partieshavedisplayed differences.
Explainingthe Fate ofEthnic MinorityParties{Nakai}
Applying Spatial
Th
eory for Ethnic VotingRecent literatureshavetried toaccount for thevariationof partysystems from a more actor'orientedapproach. Some studies
explain how those variedcircumstancescan be accounted forby
differencesthatresult from interest coordinationbetween political parties and voters. Horowitzand Brown (2005)verifiedthrough regression analysisthatinstitutionalvariablesand socio-economic variables were invalid inexplaining political partysystems inCEE countries， including thenumber of significant political parties. They argued instead that the ideologicalconcentration'polarization in politicalparty systems is a very significant variable.Ideological positioning shouldbe treated asa set ofpolicypositions foreach party. Inshort， toexplainthevarianceinethnicparties'situations， we must pay attention to policy positions. Mahr and Nagle (1995) emphasisedpolicy positionsinCEE countriesand noteddifferences between westernand eastern partysystems
Every party seekspower resourcesinelectoral markets， and
ethnic minority voters， likeallother voters， try toregistertheir preferencesforofficeholdersin parliament through elections.Inthe
post'communist emerging democracies， generally speaking， organizationalbases for each politicalpartyareweak and highly vulnerable.Inaddition， voters supportforeach partyand linkage between parties and societiesaresofluidand unconsolidated(Kitchelt
et al 1999; Lewis 2003).Thissituationleteach partyhave strong
incentives to appeal various voters for support maximization. As
Kitschelt(2000)pointed out， ethnic groups remained assolidsocietal
groups after the atomization in the communist era.Under the newly
Explaining the Fate ofEthnic MinorityParties{Nakai}
democratized situations， voters does not have enough political knowledge about political parties (]ike policy orientation， credibility，
intereststructure， and so on)， hence they tend torely on more visible factorsof candidateliketheir ethnicity.Consequently， several politicalparties， evenorganized by ethnicmajority， try to retain ethnicminorities'voteby credit claimingorby porkbarrelmeasures Birnir (2007)argued that ethnicrepresentationsareneveralways "intransigent"and it is less possible thatethnicminoritiescast a ballot forethnicminority parties when saliency of ethniccleavages defined by governmental policy orethnicattractor is low. Kostadinova's regression analysisshowed that， even though itwas not the main pointofher argument， ethno-linguisticfractionalization indexvariables
have never had unique and static effectsto account for partysystem polarization (Kostadinova 2002).Although some scholarshave argued that ethnicminorityvotingbehaviour have its own specialty (Cox 1997;Horowitz 1985)， Cox'sown regressionanalysisimplied that societal ethnicfactors neveraffected theresultsof votingbehaviour It is possible that ethnic votershave voted strategically， using their rationality asevery voterdoes. This research assumes rationality asvoters could voteretrospectively evaluatingpoliticiansactions (Fiorina 1981).Chandra (2004， 2009) theorisesaboutthemechanism that generatesdifferencesof support forethnic parties in each local district in India.
I
f
margins ofwin by ethnicparties inelectoral districtatlast time is larger than the percentage of minorities in that district， leverageof votingits partygonegative. Itmeans thatethnic minorityvoters will abandon their votesforthe ethnic partiesand seek otherpartiesthattendto contribute to theirpolitical and material interests. Inshort， ethnicminorityvotersbehave very strategicallyOike asevery votersdo)and not ideologically且xlng -8(210)ーExplainingthe FateofEthnic MinorityParties{Nakai}
They wilJreactto a political elites' policyperformance and itschanges，
and will try to prevent their votefrom beingwasted. Sincetheir behaviourresembles strategic votingW and react appropriatelyfor policy change， we are able to treatethnic voting behaviour asusual voting behaviouris. Even though some theoristsofelectoralstudies and strategic voting have hedged， saying thattheirtheoretical frames would be inapplicableforethnic voters， 1 argue thatit is more appropriateto treat ethnicvoters inthe same way as'normal' voters Formalizing thesearguments with thebasicHotelling (1929) model，ω this articlerepresents a simple frame toshow the voting behaviourofethnic minoritiesand the situationofethnic parties in national parliaments
E
型旦ユD
Q
PJ v，
P
2 P，
(1)There areseveral definitions of'strategic voting'.This article defines it here as'casting a votefor a second-orthird-best partytoescape wastinga vote' 121 This modelhas beenfamous as an originalwork following Downs (19571model The argument aboutthe reasonabilityaround this model willbeomittedfrom thepresent article becausethere are a vastnumber of worksfrom thepast 80 years， and itis outofdomain to this work -9(211
l
ー
Explainingthe Fate ofEthnic Minority Parties {Nakai} Fi郡lre1 denotes a highly hypothetical situationinwhich there are only three political parties in parliamen
t
.
Party A shallbe a conservativepolitical party and mainly supportedby titular majority people. Party B shallbe a liberal political partyand alsomainly supportedby titular majority people. PartyC shallbe a small ethnic minorityparty.We can assume that everypolicy， including integration policy， willbe determined by bargaining and coordination among the big parties， A and B， because party C is relativelysmalland unable to mobilisewelltoexertits presence. Every party has alinear function representingsubjective costs thatthevotersfeel in votingforeach party. I
f
t
he policypositions of party j， represented asPj， and the preferencesof voteri (Ui) are quite thesame， voterwill not feelany cost (other than thecostofvotingitseli) ， but
thebroader the distance between the party j stanceand the voter preference， the more discountcostDυ= I pr u;Ithe voter will feel A voter will votefora party that they perceivecoststhem the least. Moreover， when voters perceivea veryhigh costofvoting for any party， theywillnotcasta vote.Then， there shouldbe theconceptof a cost permissible limit forevery voteri， and it is representedasa function Qi.
I
f
and only if Dυ>Qi forany j， voter does notvotefor any party.However， most ethnic smallpartiessufferfrom an inability
to win an election， and we should also take this into accountin thecalculation， along with policy positions. Every voterfeels some
discountcosttovotefora smallparty because hislhervote could be wasted. Thisextradiscountcostfeelingsfor voting for a small
partycould be represented as an increaseon the discountcost
function， likeparty C in Figure2
Explaining the Fate of Ethnic MinorityParties{Nakai} 堕 型 堕
2
D
Q
V2_.FD3 _ V • IP
2 P，
In this situation， distancebetween policyposition of party P3 and preference of voter U1 is closer than distancebetween p2 and U1. However， voter U1 cast their ballot for party B in this world， because that choicecostsless. In other words， D1.2is smaller than D1，3 Let us consider about policy change. When political party B shifts their policy stancefrom p2 to P2'， it is denoted as Figure 3 below， and the voter U2 is likely to changehislhervoting party Becausethe Dd became smallerthan D2.3E
型 旦2
D
Q
VzP; P; v，
P
，
-11(213)ーExplainingthe Fate ofEthnic MinorityParties{Nakai}
Now， the policychange of Party B from p2 toP2' means that everypotentialsupporterfor partyC couldchange theirvoting
preference from Party C to Party B. Ifwe defineand expressthis
situationin naturallanguage， many members of theelectorateGn
thisarticle， thismeans minorityvoters) casttheir ballot for a large， second-bestpartyeven though theyknow thereisa party whose
policy is quite close to their preferences， inorder toavoidwasting a vote.Itis reasonable to assurne that a policychange should be treated assignificant variablestoexplainthevariation inethnic parties' fates based on theoretical deduction. Itleads to thehypotheses below Hypothesis 1 The vote share of ethnic partiesisaffected by other parties' (or governmentaJ)policychange In addition， a rival hypothesiswillbe representedasbelow Hypothesis2. Ifan institutional designis advantageous (high proportionality，
low thresholdand large magnitude ofa district) for a small
party， the ethnicminority partieswill win seats. Hypothesis3. If thesociety ismore fractionalised in terms ofethnicity， or has much ethnic minority population， ethnicminority parties could win seats.
一
12(214)ーExplaining the Fate of Ethnic MinorityParties{Nakai} Statistical
An
alysis To evaluate the hypothesis presented above， this study empirically analyzes the relationship between electoral outcomes and the government positionon national integration policies towardethnic minority groups. The integration policyrefers to several factors， such as politics， economicsand social concerns， hencethesefactors are integrated into a set of policyinitiatives. Even if it is difficult to estimate and scorethe individual parties' policy stanceoverseveral years and in several different countries， we can estimate the governmental policyorstancethat has been generated by coordination among significant political parties This analysis takesinto account several other factors.Institutional factors consist of several aspects， asmentioned in literature review section， like the electoral system， magnitude of electoral districts and so on. Societal factors consist of mainly the language.group or ethnic.group situation. External diplomatic factorwill also be tested The unit of analysis is each ethnic group ineach election ineach CEE country. For example， Lithuania has two ethnic groups. Russians and Poles. and they have experienced elections 5 times， this his generates10 casesinthe dataset.The ethnic party's seats and policy variablesare coded separately to correspond with each ethnicgroup. A country.year， which is coded as 'democracy' at PolityIV and as'free' atFreedom House， shallonly be used inthis data sample because this researchtreats electoral issues
(See Appendix). For example， Croatia in the '90s shall be excluded from this dataset becauseitwas coded as 'NotFree' by Freedom
Explainingthe Fateof Ethnic MinorityParties{Nakai}
Houseatthat time.
Dependent
v
c
αnαbleItmeasuresproportion (permillage) of seatshare by ethnic minority parties calculated from each country's election result Determination ofethnicparties hasbeen made in linewith Bugajski's
work (2002)， and several expert literatures(seeappendix). Each
ethnicgroup iscoded separately， asis each ethnic party Independent viα口αbles Political
D
i
scriminationand Political Rightsarethe scoresof governmental integration policy about politicalofficialdiscrimination and empowerment for minorities. Thesearefirst independentvariables used to assessthis article'shypothesis.It is codedfrom theMinoritiesAt Risk (MAR) project dataset ωThesevariablesaredummy variables
Ifthere is officialpolitical discrimination forethnic minority (if MAR score takes，)1 Political Discriminationtakes1 and theothers
takeO. Ifthereis official empowerment policy for ethnicminority，
Political Rights takes 1 and the otherstake 0
Language Discrimination and Language Rights are thesecond
independent variables usedtoevaluate argument.These variables
indicatethe status of ethnic minorities' languages. These variables
are also on dummy scales.Ifofficiallanguage policyis reslはlated to use ethnic language even incivil societyand outlinessome type
(3) Detalied definitions are:1 -officialpolitical discriminationandintensive discrimination incivilsociety;2 -no official discriminationbuthistorical mter】引vediscrimination incivil society; 3 -noofficial discriminationhut weak discriminationincivil society; 4 'official protective policybutweak discriminationincivil society; 5 -there is noethnicdiscrimination 14 (216)
Explaining the Fate of Ethnic Minority Parties{Nakai}
of punishment， Language Discriminationtakes 1 and the others take O.
I
f
minoritylanguage has or couldhave official statusin thelocalor nationalunit， Language Righttakes1， and the others take O.1 refer to Spolsky (2004) tomake and code these variables Control Variables PR-SMD controlsfortheinstitutionaleffectas an explanation of the situation ofethnicparties.Itis theproportionof seats (0-1) which is determined by Proportional Representation (PR) system Therefore， if a country adopts a full Single-member district(SMD) system， this variable is coded as 0， and ifa country adopts Mixed-systems and 120of 200 seatsareelectedby a PR system，it is coded as 0.6
Magnitude is the averagenumber of districtmagnitude.When a country'smagnitude has some extentof range between tiers， and tiersundera mixed-method system， this variable is coded with a largenumber in cording this variable， becausesmall parties such as ethnic parties generally tend torun forin thedistrictwith largermagnitude Threshold is simply coded as thepercentagethatis needed to getseatsineach country'snationalelection.
I
f
some countries give ethnic groupssome privileges asdischarging of threshold， the simple Hare quota， which is needed forwinning a seat， would be appliedforcodingEthnicSeats isdummy variables whether there arespecialmandate seatsfor ethnic minority ornot. Ifthere are， it is coded as1.
Fractionalization is the primary variable forcontrollingfor socio-economic factors. The famous work by Alesinaetal.(1997) is usedhere.This variableindicatesthepolarizationofethnicity
Explainingthe Fate ofEthnic MinorityParties{Nakai} and language ineach state. Composition isa variable that represents theratio ofeach ethnic group compositionin the national population. This is also a control variableforsocio-economicfactors. 1 codedthis data from the CIA World LobbyActor Factbook'sweb resources. IS international factor control variable entertaining external or Jenne (2004， 2007)demonstrate thatethnic
minoritygroup enhance demands when theycouldtake a support
from externallobby actor (their national "Homeland" ortheother
organizations). Each minoritygroup who has "national homeland" takes1， and the others take 0 in this test. Pre-8eats measure the proportion ofseat shareby ethnicparties in pre-election. Thisvariableiscodedbased on electoral result at timet-1 1n regression analysis， there iscorrelationbetween political integration policy factorsand language integration policy factors 80， this analysis calculates separatelyfrom eachother. Moreover，
this study employs generalizedlinear model (GLM) method assuming two types ofdistribution， negative binominal distribution(NBD) and log-normal distribution (LND)， because many dependent variables
areslanted to zero.
Explaining the Fate ofEthnic Minority Parties{Nakai} Table 2:The resul!s 01 s!a!is!ical analysis Dependent Variable:ProportionofSeats Shareby Ethnic MinorityParties Statistical恥1odel:GeneralizedLinear恥10del Distributionassumption NBD LND
modell model2 modcl 3 model4 Composition 0.038 0.079ホ 0.778事 申 0.547 (0.032) (0.040) (0.372) (0.394) Fractionalization 1.210 1.093 24.959 8.273 (2.107) ( 1.978) (26.048) (24.455) Threshold 0.414 0.231 0.060 0.016 (0.158) (0.163) (0.421 ) (0.409) EthnicSeats 1.046寧 事 0.774 5.691 1.457 (0.522) (0.568) (7.799) (7.588) PR-SMD 1.596 2.785*事 $ 27.316 18.776 (1.217) (0.983) (16.780) (14.079) Magnitude 0.003 0.001 0.028 0.230 (0.002) (0.003) (0.030) (0.030) LobbyActor 1.845事 事 事 2.632事 事 事 7.949 14.576ホ 申 (0.542) (0.427) (7.388) (5.856) Pre-Seats 23.925*ホ * 21β20申 *ホ 837.680ホ * 寧 832.643*牢 * (6.114) (4.934) (71.923) (68.212) PoliticalDiscrimination 2.403*事 19.770
・
(0.835) (10.551) PoliticalRights 0.747 14.470 (0.606) (8.833) LanguageDiscrimination -1.043 31.989事 (1.315) (16.699) LanguageRights 0.035 15.318輩 申 (0.615) (6.578) Constant ー0.042 2.210 -42.198事 $ -24.347 (1.837) (1.458) (20.378) ( 15.578) p-value: omnibus test 0.000 0.000 0.000 0.000 log likelihood -356.6 -359.2 -591.4 -589.8 AfC 737.3 742.5 1206.8 1203.6 N 124 124 124 124 ホ 申 事 p< 0.01 * 牟 p< 0.05 車 p< 0.10 17(219)Explaining the Fate ofEthnic MinorityParties{Nakai}
We couldobserve severalpoints from thisresult. First， political discrimination as policy output have significanteffects toexplain the proportion ofseats won by ethnic parties. We could interpret that it is important whether official political acts of discrimination are implemented ornot， as model land 3 show.When and where ethnicminoritiesobtainpolitical rights， ethnicpartieslose support.
This iscounterintuitive， but reasonable and line up withtheoretical hypothesisofthisstudy.Second， Language policies alsoappropriately
have significancein model 4 as hypothesis predict. However， these lack significant effects in negative binominalregression. Third， Institutional factorscouldhave some weak effectonlyinmodel
1 and 2， but notconsistently.There are ineffectiveness ofelectoral
institutions' factors in model 3 and model 4. Third， Socio-Economic
factorshavenever consistentlyaffected theperformanceof ethnic
minority parties well. Only demographic compositionsofethnic
minoritiesshow significancetoexplaindependent variablesin model 2 and 3.At last， presencesof ethnic minoritygroups'national homeland
slgnl日cantlypredictthe winning seatsby them. This may paritially
explain why most ofallRoma groups'parties have never been ableto represent
The credibilityofwinning seatsby ethnic parties have been determined by its performance(pre-seαtsvariable)， even any other significant variablesarecontrolled.Inother words， supportersof ethnic parties (mostofthem belong toan ethnic minority) did not beexclusivelyaffected by variation of electoralinstitutions. Slhe votesfor an ethnicparty if its performance have been considerably good， butthey avoidswastinga voteifsuch ethnicparties have failed to win seats;and allof a voter'sbehaviours are defined by thesestrategies.Moreover， this result also demonstrates that 18(220)
Explaining the Fate ofEthnic Minority Parties{Nakai}
thefate of ethnic parties is notsolely determined by electorateぜ ideologicalprinciples.
I
f
ethnic voting had unusual aspect compared with general voting behaviour， and the expression of identity would be important for members ofan ethnicminority， then they would vote without considerationof a party's performancein thelastelection However， theyhave changed their supportforethnicparties in responsetothe parties'capability.This infers the rationality of ethnic voters Regression analysis revealed the significance of the policy factor and of voters'responsive attitude.Itrepresented the ineffectiveness and limitationofa perspectivetreating electoralinstitution as an sole determinant factor to explain therise and fall of ethnic minority parties. This finding verifiesthe theoretical deduction in the above section， and has some relevance to the applicationof thattheory to CEE post-communist democraciesConclusion
This paper has argued about the divergent situationsof ethnic minority political parties in newly democratised countries， and concludesthat the fate ofethnicparties has been explained by a more actor-centric approach， which focusesespecially on the changes in integration policyand the strategic voting behaviour of minority voters. Institutional orsocio-economical explanations show some limitations
Assuming voters' rationality and theresponsibilityforpolicy change， ourtheoreticaland formalised deduction explains well the reason why ethnicparties could or could not win seats.Moreover，
through an analysis of CEE democracies， the statistical verifications
Explainingthe Fate ofEthnic MinorityParties{Nakai}
support the theoretical argument empiricaJly. It has been shown that ethnicpartysupports seldom strongly depend on theneither electoral institutional factors nor socio-economic factors. Ethnic composition seems tobe somewhat effectiveinaccountingforethnic parties' fates， but it is inconsistent. Policy does in
f
1
uence whether ethnicminorities can enter thepolitical arena ornot Thisresearch showed the importanceofthe policy factor and ofvoters'incentives toexplain ethnicparties' fates. The question aboutethnicpartiesshouldbe incJuded in questions about the partysystem， butcJassical works about partysystems rarely definethe varianceamong ethnic partiesinCEE countries.My argument proposed a more plausibleexplanationfor this question， and builds a bridge for furtherempiricalresearch about ethnopolitics Appendix. Statistical analys回 units Countl陪Sα凡d election yeαrs Estonia (1992， 1995， 1999， 2003， 2007)， Latvia (1993， 1995， 1998， 2002， 2006)， Lithuania(1992， 1996， 2000， 2004， 2008)，
Poland [Lower Chamberl (1991， 1993， 1997， 2001， 2005， 2007)，
Czech [Lower Chamberl (1996， 1998， 2002， 2006)， Slovakia(1994， 1998， 2002， 2006)， Hungary (1990， 1994， 1998， 2002， 2006)， Slovenia(1992， 1996， 2000， 2004， 2008)， Croatia(2000[LowerChamber
，
J
2003， 2007)， Serbia (incJuding SerbiaCrna Gola) (2000， 2003， 2007， 2008)， Montenegro (2006)， 20 (222)Explaining the Fate of Ethnic Minority Parties{Nakai} Albania (2005)， Macedonia (1994， 1998， 2002， 2006， 2008)， Bulgaria (1990， 1991， 1994， 1997， 2001， 2005)， Romania (1996， 2000， 2004， 2008)， Moldova (1994， 1998， 2001， 2005)， Ukraine (1994， 1998， 2002， 2006， 2007)。 Ethnic Minoritypαr， tws
Estonia: Our Home isEstonial United People'sPartylConstitutional
PartylRussian Partyin Estonia，
Latvia: Equal Rights， Russian Citizen Partyl National Harmony
PartylFor Human Rights and United Latvial Concord Center， Lithuania: Lithuanian Russian Union/ Pole's Electoral Union of
LithuanialPole's Electoral Alliance ofLithuanial MinorityPeople's Alliance，
Poland: German Minorityl Cultural Society of the Germans inthe Opole District，
Czech: Union of Slovak/ Community of Slovak/ Democratic Alliance of SlovakslHSMS， Romany Civil InitiativelRomany National councillPolish Councill U nionof German culturel Council of Jewish Communities (This analysisexcludesmolavia'oriented partiesbecauseit is appropriate to treat those as regional parties)，
Slovakia: Hungarian Coalition/ Hungarian Christian Democratic
Movementl Coexistence，
Hungary: National Minority Roundtablel Council of Gypsies in Hungaryl Roma Parliamentl Democratic Alliance Hungarian Gypsiesl Independent Gypsy Associationl Association of Germansl Democratic Union of Slovak in Hungaryl Jewish Cultural Alliancel Democratic Federation of Serbl Democratic Association of
Explainingthe Fate of Ethnic MinorityParties{Nakai}
Romaruan，
81ovenia: ltalian UnionJInterestCommunity of Hungarian Minority，
Croatia:8erb People'sPartyl Independent Democratic8erbianPartyl Party ofDemocratic ActionofCroatial Democratic Union of Hungarian (This analysisexcludes "Istoria deomcraticcouncil" because itseems to be regional parties)， Montenegro: SerbianRadicalPartyl 8erbianNational Renewall8erbian Fatherlandl8erbia NationalParty(8NP)1 8NP-NP-D881 Together for Yugoslavial 8erbListl Radical PartyofMontenegrol 8erbian
UnionJ Bosniak CoalitionJ Albanian Alternativel Albanians
Togetherl Democratic Union ofAlbanianslLiberals and Bosniak Party-Correct in the Past， Right for the Future，
Albania:Unity for Human RightParty，
Bulgaria:Movement for Rights and Freedoms，
Romania: Association of Italians of Romanial Bulgarian Union of
theBanatl Cultural Union ofRuthenians ofRomania Uniunea/
democraticForum of Germans of Romanial DemocraticUnion
of 810vaks and Czechs in Romania/ Democratic U nion of
Turco-Islamic Tatars of Romania/ Federation of Jewish Communities ofRomanial Greek Union of Romania/ League
ofAlbanians of Romanial Lipovan RussianCommunity ofRomania/
Party oftheRomal Turkish DemocraticUnion ofRomanial Union ofArmenians ofRomanial Union of Croatiansof Romania/Union
of Polesof Romania Dom Polskil Union of8erbs ofRomanial Union of 81avicMacedonians of Romanial Union ofUkrainians
of Romania，
Moldova: Unity-Transnistria， Union ofPatriotMovementl Gagauz PeoplelDemocratic Party of the Gagauz，
Macedonia:DemocraticUnion for IntegrationJParty for Democratic
Explaining the Fate of Ethnic Minority Parties{Nakai}
Prosperityl Democratic League ofBosniakslDemocratic Party of AlbanianslDemocratic Party of Serbsl United Party of Romas in Macedonia! Democratic Party ofTurkslPartyfor Roma IntegrationJParty of the Democratic Forces of Roma in Macedonial Democratic Party of the Bosnmiaksl Union of Roma in Macedonial Party for the Movement of Turks in Macedonia，
Serbia: Hungarian CoalitionJBosniac Listfora European Sanjakl Albanian CoalitionJRoma Partyl Roma forRomal Roma Union of Serbial Montenegrin Partyl Alliance of Vojvodina Hungariansl Together for Tolerance ?anak， Kasza， Ljaji?1 Alliance of Vojvodina Magyarsl Democratic Fellowship of Vojvodina Hungariansl DemocraticReform Party of Muslims (This analysis defines Vojvodina orientedpartiesas regionalparties exceptabove)， Ukraine: Crimia Russian Partyl Democratic Movement of Donbas
References
Alesina， A.， Devleschawuer， A.， Eastealy， E.Kurlat， S.， Wacziarg， R.， 2003. nFractionalization.l1 Journα1 of Economic Growth， 8，
155-194
Birnir， J. K.， 2007. Ethnicityαnd Electorα1 Pol山cs.Cambridge University Press， New York.
Bugajski， J.， 2002.PoliticαI Rαrties of E
，
αstern Europe: A Guide to Pol山 田 川 thePost-Communist ErαM.， E.Sharpe， New York and LondonChandra， K.， 2004. Jf古:yEthnicpα，rtiesSucceed:pα，tronα:geαnd
Ethnic Heαd Coun白 山 lndiαCambridge University Press，
New York
Explaining the Fate of Ethnic Minority Parties{Nakai}
Chandra， K.， 2009. "Why VotersinPatronage DemocraciesSpJit TheirTickets:Strategic VotingforEthnic Parties." Electorαl
Studies， 28 (1)， 21-32
Cordell， K.， Wolff， S. (Eds.)， 2004. The Ethnopoliticα1 Encyclopediα
of Eu/'ope.Palgrave Macmillan， Hampshire and New York. Cox， G. W.， 1997. Mαking Votes Count: Strategic Coo/'dinαtwn
in the Wo，/ld's Electo/'α1 Systems.Cambridge University Press，
New York
Duverger， M.， 1954. PoliticαI Rα/'ties: Orgαmzαtwnαnd Actiuity in the Modern Stαte. Cambridge University Press， New York Fiorina， M. P. 1981. Ret/'ospective Voting in Ame/'icαn Nαμonαl
Elections.Yale University Press， New Haven
Geddes， B.， 1995. "A Comparative Perspective on the Leninist Legacy
inEastern Europe."Compαrαtive Political Stud同s，28 (2)， 239・274.
Horowitz， D. L.， 1985. Ethnic G/'oups in Conflict. University of California Press， Berkeley and Los Angels
Horowitz， S.， Browne， E.C.， 2005. "Source of Post-Communist Party System ConsoJidation."Pa/'ty Politics， 11(6)， 689-706
Hotelling， H.， 1929."StabilityinCompetition."The Economic Journal，
39(153)， 41.-57
Jenne， E.， 2004. "A Bargaining Theory ofMinority Demands: Explaining the Dog thatDid not Bite in1990s Yugoslavia."Internαtwnαl
Studies Quα/'te/'ly.48 (4)， 729-754.
Jenne
，
E. 2009. Ethnic Bα/'gαining: The pα，
pαdox of Mino/'ity Empowerment. Cornell University Press， Ithacaand LondonKitschelt， H.， Mansfeldova， Z.， Markowski， R.， eta.l(Eds.)， 1999
Post-Communit Rαrty Systems: Competition， Representαtion，αnd Inte/'-Pα
，
rty Coope/'αtion. Cambridge University Press， New YorkKitschelt， H.， 2000. "Linkagesbetween Citizens and Politicianin 24 (226)
Explaining the Fate ofEthnic Minority Parties {Nakai}
Democratic Politics.n Compa1αtive Politicα1 Studies， 33 (6)，
845・879.
Kostadinova， T.， 2002. "DoMixed Electoral Systems Matter?: A
Cross-national analysis ofTheir Effect in Eastern Europe."
Elector.α1 Studies， 21
u
l
， 23-34Lewis， P.G.， 2003.PoliticalParties.In:White， S.， Batt， J.， Lewis，
P.G.(Edsl.， Developmentsin Centralαnd ，Eαst Europeαn Politics
3. PalgraveMacmillan， Hampshire and New York
LipsetS.， Rokkan， S.(Eds.l， 1967.pα，rtySystemsαnd Voter A lignment:
The Cross-nαtionalPerspect山 田 The Free Press， New York
Mahr， A.， Nagle，
J
.
D.， 1995. "Resurrection of the SuccessorParties and De町mocr阻atlzatl回onin East-CentralEurope凪 " CommunistαndPost-Communist Studies， 28 (4)， 393-409
Meguid， B. M.， 2005. "Competition between Unequals: The Role
ofMainstream Party Strategy inNiche Party Success，" Americαn
Pol山cα1ScienceR即日山， ，99 (3)， 347-359
Millard， F.， 2004. Elections， pα，rtlAs，αnd Representαtwn ln
Post-Communist Europe.Palgrave Macmillan， Hampshire and New York
Minority atRisk Project， 2005. MinoritiesatRisk Dataset.College
Park: Center for International Development and Con
f
1
ictManagement. Retrieved April 20， 2010， from http://www.cidcm umd.eduJmarl
Meleshevich， A. A.， 2007.PartySystems in Post-Soviet Countries
A Compαyαtive Study ofPoliticα1 Institutionαlizαtioninthe
BαlticStαtes， Russiα，αnd Ukrαine.Palgrave Macmillan， New
York
Moser， R.G.， 1999."ElectoralSystems and the Number of Parties inPostcommunist States." World Politics， 51(3)， 359-384
Explaining the Fate ofEthnic Minority Parties {Nakai}
Moser， R. G.， Scheiner， E.， 2004. "Mixed ElectoralSystems and
Electoral System Effects: Controlled Comparison and Cross-NationalAnalysis." Electorα.1 Studies， 23 (4)， 575-599 Moser， R.G.， Scheiner， E.， 2009. "StrategicVoting inEstablished and New Democracies:TicketSplitting inMixed-member electoral Systems."Electorα1 Studies， 28 (1)， 51-61. Muller-Rommel， F.， Fettelschoss， K.， Harfst， P.， 2004. "Party Government inCentral EasternEuropean Democracies: A Data
Collection (1990-2003)."Europeαn JournalofPoliticαIR目 的 陀h，
43， 869・893
Rose， R.， Munro， N.， 2009.p，αrtl8Sαnd ElectionsinNew Europeαn
DemOClαcies.ECPR Press， Colchester
Spolsky， B.， 2004.LαnguαgePolicy: Key Topicsin Sociolinguistics Cambridge UniversityPress， New York.
Whitefield， S.， 2002. "PoliticalCleavages and Post-Communist Politics." Annual Review ofPoliticalScience， 5， 181-200
(本学法学部准教授)
