The Political Consequences of Electoral Laws, 1945-85 Author(s): Arend Lijphart Source: The American Political Science Review,

Vol. 84, No. 2 (Jun., 1990), pp. 481-496 Published by: American Political Science Association Stable URL: http://www.jstor.org/stable/1963530 Accessed: 04/11/2009 22:03
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=apsa. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. 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 support@jstor.org.

American Political Science Association is collaborating with JSTOR to digitize, preserve and extend access to The American Political Science Review.

http://www.jstor.org

CONSEQUENCES THEPOLITICAL OF ELECTORAL LAWS,1945-85

AREND LIJPHART
Universityof California San Diego
A systematicanalysis of the relationshipsbetween the main electoralsystem variables (electoralformula, districtmagnitude,and ballot structure) in the 20 and multipartism) and electoraloutcomes (the degreesof disproportionality Westerndemocracies from 1945 to 1985-representing 32 distinctelectoralsystems (an electoralsystem beingdefinedas a set of electionsheld underbasicallythe same rules)shows that the effects of both formula and magnitude on proportionalityare very havesuggested; strong, muchstrongerthan Douglas W. Rae and subsequentresearchers in electionsis that on the other hand, theireffectson the numberof partiesparticipating only in surprisinglyweak; and that ballot structureaffects the degree of multipartism districtsystems. Thesefindingssuggestthat strategicbehaviorby politisingle-member cians and voters plays a less importantrole in reducingmultipartismthan is usually assumed.

W henthe first ficientcriticalattention.In theprefaceto editionof Douglas W. Rae'sThePolitical thesecondedition,Rae(1971,vii) himself colleagueswho Consequences of ElectoralLawswas pub- chides "the over-gentle in 1967,it wasalmostinstantly lished rec- reviewedthe originaledition."'In fact, as a great froma variety of concepognized scholarly breakthrough. thebooksuffers A secondeditionwas published in 1971, tual, methodological, and empirical andit hasmaintained itsstellar reputation weaknesses,which are so grave and that they cast doubton all of bookin thefieldof numerous as themostimportant comparative electoral conclusions. systemseversince. Rae's Moreover, someof his Morerecent are so surprising-especially major innovative bookshave conclusions usedit bothas a modelandas a thefinding thatdifferences in proportionexplicitly pointof departure (Katz 1980,20;Nohlen al representation haveonly (PR)formulas 1978, 14), and contemporary authors minor effects and that neithertwo-tier on thesubject stillunfailingly cite districting northeballotstructure systems writing it. It received the 1989George H. Hallett have any effectat all-that they cry out Awardof the American new examination. PoliticalScience for a critical Association's My purpose Representation hereis to reanalyze andElectorRae's section. classic study with more accuratedata, al Systems As thefirstsystematic broadly compar- stronger andbettermethods. hypotheses, ativestudyof electoral systemsand as a I shallalso use more data:to Rae'selecpowerful stimulus to subsequent research, tions (of the loweror only housesof the it clearly deservesits reputationas a legislature) in the20 Western democracies classicin the field.On the otherhand,it duringthe 20-yearperiodfrom 1945 to hasbeenaccorded thisstatuswithoutsuf- 1964, I shall add the few electionsin
AMERICAN POLITICAL SCIENCE REVIEW VOLUME 84 NO. 2 JUNE 1990

American Political Science Review Vol. 84

1945-64 missing from Rae's book and, more importantly, all elections in these countries during the subsequent21-year period from 1965 through1985.2 In order to save space and becauseI found no major differences between the patterns in Rae's 1945-64 period and in the entire 1945-85 period, I shall presentthe results for the latter only. My reanalysis and update will yield conclusionsthat are significantlydifferent from the relationshipsthat Rae finds between his variables. Rae defines three aspectsof the electoralsystem as his independent variables: (1) the electoral formula, such as plurality, various forms of PR, and so on; (2) the average district magnitude,that is, the averagenumberof legislatorselected per district;and (3) the ballot structure,which may or may not allow the voters to split their votes between two or more parties (pp. 15-46). His two dependentvariables are (1) the proportionalityof the electoral outcome, which is the immediateor (in Rae'stermieffect of the electoral nology) "proximal" system's translation of votes into seats; and (2) the degree of multipartismthat is-or is presumedto be-the indirector effect of the electoralaystem (pp. "distal" 47-64, 67-68, 84-85). Rae examines five major relationships between his independentand dependent variables.With regardto all five of these, my conclusionswill divergefrom Rae'sin crucial respects: (1) the relationshipbetween the electoral formula and the proportionalityof the election result is much stronger than Rae finds; (2) the link between districtmagnitudeand proportionality is also much stronger, especially as far as concernsthe effects of two-tier districtingsystems(whichRaeclaimsto have betweenthe no effect);(3) the relationship electoralformulaand the degreeof multipartism is much weaker than in Rae's finding; (4) similarly, district magnitude and multipartismare much less strongly and perfectlyrelatedthan Rae claims; (5)
482

the link between ballot structure and multipartism hypothesized and rejected by Rae is in fact valid for one majorcategory of electoralsystems-single-member districtsystems. In otherwords, the proximal effects of electoralsystems are much stronger than their distal effects. The theoretical significance of this finding, which I shall discuss in greater detail in the concludingsection, is that strategic(or insincere)voting and strategicelite behavior play a considerablyless importantrole in helping the larger, and hurting the smaller, parties than is usually supposed. How do my methods differ from, and how are they superiorto, Rae's?Five differencesare of importancethroughoutthe analysis. Firstin contrastwith Rae'streatment of each election as a differentcase, my unit of analysisis the electoralsystem, defined as one or more elections held underbasicallythe same or closely similar rules. For instance, instead of counting the 11 Israeli elections as 11 separate cases, I have only two cases: the PR system that used the d'Hondtformula(the 1949 election and the four elections from 1973 to 1984) and the system under the rule (the six elections largest-remainders in the 1951-69 period). My approach avoids the artificial and deceptive inflation of the weight of the available evidence; elections under the same rules are not really independentcases but merely repeatedoperationsof the same electoral system. It also avoids the relatedproblem of artificiallyweighting the evidence. For instance, France held only 3 elections under the pure d'Hondt rule while the Netherlandshad 12; but this is not a good reason to give the Frenchevidence only one-fourththe weight of the Dutch. This approach yields 32 cases (based on 255 elections) in the 1945-85 period. Second, I use the evidence of all of the available cases as much as possible. For instance, in contrastwith Rae, who simpthe difficultcases of two-tier ly disregards systems that use different formulas at

ElectoralLaws
theirtwo levels, I determinewhich level is of predominantimportance and classify the electoralsystem accordingto the formula at this level. This procedure increasesthe numberof casesin my analysis from 20 to 31; the-onlycase that I cannot use for the purposeof studyingthe effects of electoralformulasis France(1951-56), because it used majority, d'Hondt, and largest remainders formulas in such a complex way (in different parts of the country, not at differentlevels) that they cannot be disentangled. Third, since Rae finds that both formula and magnitudestrongly affect proportionality and multipartism, it is a curious omission in his analysis that he fails to examinethe possibleinteractionof his two explanatory variables. I correct this omission by systematically controlling for the influenceof the otherindependent variable. Fourth,in order to measurethe degree of disproportionality of differentelectoral systems, I use JohnLoosemoreand Victor J. Hanby's (1971) index D. It has become the most widely used index of disproportionality (see, e.g., Mackie and Rose 1982, 411-12; Rose 1984; Taageperaand Shugart1989, 104-11). D is the total percentage by which the overrepresented parties are overrepresented-which is, of course,the same as the total percentageof underrepresentation. In orderto calculate D, the absolute values of all differences between the parties'vote and seat shares are added and then divided by 2. Rae's measureI appears to be similar. It, too, entails the addition of all vote-seat share but this sum is dividednot by differences; 2 but by the number of parties. As RichardS. Katz (1980, 140) has pointed out, Rae'sindex I has the grave defect of "givingtoo much weight to small parties; at the extreme, if the infinite number of (hypothetical) parties that receive no votes and obtain no seats is included, every electoral system would appear perfectly proportional."Rae (p. 84) tries to avoid this problem by disregarding parties with less than .5 percent of the vote, but this arbitrarycutoff point is still quite low; the presenceof several parties with just over .5% of the votes will depressI even if these parties fail to win any seats.3 Finally, a relatively minor difference between Rae's methods and mine concerns the measure of multipartism.Rae uses a large number of indicators of multipartism-such as the total number of partiescontestingan election, the total numberelectedto the legislature,the vote and seat shares of the largest party, and the vote and seat sharesof the two largest parties combined-but his most important and comprehensivemeasure is the fractionalization of the party system, which is sensitive to both the numbersof partiesand their relativesizes (pp. 47-64, 67-68, 84-85). I shall use an adaptationof Rae'sfractionalization index-Laakso and Taagepera's(1979) "effective number of parties"-which carries exactly the same information as Rae's index but is more meaningful. For instance, in a party system with two equally strong parties, the effective numberof parties is exactly 2.0; for three equal parties it is 3.0; for two strongpartiesand one weakerparty, it will be somewherein the neighborhood of 2.5. Rae'sfractionalization indexescan be easily converted into the LaaksoTaagepera measure. The effective numberof parties (as well as the index of fractionalization) can be calculatedon the basis of either the parties' vote or seat shares. The former is the better indicator of the long-term nature of the party system. It is affected by the operation of the electoral system in previous elections but not yet by the translationof votes into seats in the currentelection. I shall therefore use the effective number of elective

(insteadof legislative)parties as my indicator of multipartism.

483

PoliticalScienceReview Vol. 84 American

following scalefromthemostto theleast modiproportional: pure Sainte-Lague, (seeBalinfiedSainte-Lague, andd'Hondt 1986, 1982,60-66;Lijphart Rae finds, as expected,that plurality skiandYoung and majorityformulasare considerably 172-75). differences occurwithinlargest Similar lessproportional thanPRbut, rather sur(LR)systems.The oldestand prisingly, thatthereis not a greatdealof remainders difference within the family of PR best known of these simplyuses as its of validvotescast systems.The mainreasonfor the latter, quotathetotalnumber divided magniby thedistrict conclusion is Rae'simproper in a district unexpected classification of thedifferent PRformulas. tude (m, the total numberof seats availThis quota, usually He hypothesizes thathighest averages (or able in the district). to as theHarequota,is impartial divisor)formulas yield less proportional referred results than largestremainders (quota) as betweensmall and large partiesand formulas.The distinction results. betweenthese tendsto yieldcloselyproportional two typesof list PRformulas is indeed an Lessproportional outcomes areproduced important one-but mainlyas regards the by the Droop quota, which dividesthe practical procedureselection officials votes by m + 1 insteadof m, and the Imhaveto useto allocate seatsto partylists. perialiquota,which uses m + 2 as the However, the two groups of methods denominatorsThe use of these lower should not be expectedto differ with quotasmeansthattherewill be fewerreand hence regardto the proportionality they pro- mainingseats to be allocated votes, duce,sincethe degreeof proportionality also more wastageof remaining harmful to thesmaller depends on the particularquota that whichis especially in a decrease in proporandresults quota methods use and the particular parties divisor used by divisor methods.This tionality.It can be shownthatwhenthe to the exmeansthatdifferences withregard to pro- quotais loweredeven further, portionalityshould occur within each tent thattherewill not be any remaining seats, the outcomebecomesexactlythe them. groupinsteadof between formula (Van Amongthe highestaverages methods, sameas thatof thed'Hondt the d'Hondtformula (which uses the den Bergh1955, 68-72). In preferential to as singletransferdivisorseries1, 2, 3, 4, etc.) is the least PR, usuallyreferred and systematically proportional favors able vote (STV)systems,the quotais as for the proportionality of the the largerparties.It contrasts with the important Sainte-Lagud formula, which, in the electoraloutcomeas in list-PRlargestTheone caseof STV systems. originalform proposedby its inventor remainders divisorseries1, 3, among the Western democracies-Ire(usingthe odd-integer 5, 7, etc.), approximates proportionality land-uses the Droopquota. This discussion suggests a much very closely and treatslarge and small in a perfectly evenhanded parties way. In strongerand more plausiblehypothesis is more aboutthe effectsof PR formulas on the method theSainte-LaguE practice, resultthan of theelection often used in a modifiedform that uses proportionality which is basedon the 1.4 instead of 1 as the first divisor, Rae'shypothesis, for smallparties twofold distinctionbetween largest-reit harder thereby making methods. and highest-averages to gaintheirfirstseatsandhencereducing mainders classificaof the electionresult Mine is based on a threefold the proportionality and LR-Hare to theirlogical tion: (1) pureSainte-Lagud to someextent.According (2) properties,therefore,the three highest arelikelyto be themostproportional; methodscan be placedon the modified Sainte-LaguE, LR-Droop,LRaverages

ElectoralFormulasand ElectoralDisproportionality

484

ElectoralLaws
Table 1. AverageDegrees of Disproportionalityof ElectoralSystems Classifiedby ElectoralFormulaand AdjustedDistrict Magnitude,1945-85 (%)
LR-Droop,

Adjusted District Magnitude 1-1.1 1.1-5

5-10 10-25 100-150 Alla

LR-Hare and Pure Saint-LaguE

-

LR-Imperiali, Modified and Sainte-LaguE, STV-Droop

-

d'Hondt
-

Plurality and Majority

12.93 (6)
-

Alla
12.93 7.53 5.61 3.69 3.32 6.45 (6) (4) (9) (5) (7) (31)

_~
-

4.60 (1) 5.18 (3)

-

2.81 (2) 2.46 (3) 2.60 (5)

3.53 (2) 4.53 (6)

8.51 5.83 4.28 4.39 5.87

(3) (6) (3) (2) (14)

12.93 (6)

Source:Basedon data in Mackieand Rose1982, 1983, 1984;Mackie1985, 1986;Ornstein,Mann,and Malbin 1984, 120-21. ServiceCentralde la Statistique 1987, 47-48; and Luxembourg, are basedare in parentheses. Note: The numbersof cases on which the percentages Franceof 1951-56. aExcept

Imperiali, and STV-Droop shouldbe less proportional; and (3) d'Hondt is likelyto be the least proportional of the PR formulas. Thebottomrow of Table1 showsthat is stronglysupported this hypothesis by The plurality the evidence. and majority aremorethantwiceas disproporsystems tional as d'HondtPR, and the d'Hondt aremorethantwiceas disproporsystems and pure Saintetional as the LR-Hare thein-between PRsystems Lague systems; are also, as hypothesized, roughlyinterin terms of theirdegree of dispromediate portionality. Thesefiguresare basedon 31 of the 32 cases,as explained earlier.6 My findingspartly supportand partly contrastwith Rae's. Rae (pp. 96, 106) findsa similar difference between pluralandtheleastproportional PR ity-majority systems:3.96% and 1.78%respectively in terms (measured of Rae's I). Butthedifferencebetweenhis two types of PR is quitesmall:1.22%and1.78%(alsomeasuredin termsof I). Giventheweakness of Rae's hypothesis aboutPR systems,how can we account for the fact that he finds any difference
485

betweenhis two types?After carefully checking all of the likelyexplanations, I foundthatneither Rae's different methodology nor his empirical errors(of the 56 list PR electionsthat according to Rae's own criteria, shouldhave been included in his analysis,28 are misclassified or omitted) providethe answer.The explanationis thatd'Hondt to be the happens mostfrequently usedhighest foraverage mula:of the 18 highest averages cases,14 are d'Hondt.Since d'Hondtis also the of all PRmethods, leastproportional it is now clearwhy Rae'scategory of highest turnsout to be lessproportional averages than that of his largestremainders. His finding,modestas it is, is entirelyfortuitous. Table1 also showsthe indicesof disproportionality of the PRformulas when district magnitude is held constant. Withineachcategory of adjusted magnitude (to be definedin the next section), the differences betweenthe formulas are reducedto some extent; but they still andtheyareall in theexshowup clearly pecteddirection.

American Political Science Review Vol. 84 District Magnitudeand ElectoralDisproportionality

The strongest relationship found by Rae is that between district magnitude and proportionality: as magnitude ingoes down. He creases,disproportionality comparesthe effects of electoral systems classified according to five categories of averagedistrictmagnitude,m: (1) singlememberdistrictswhere m = 1, (2) m between 2 and 6, (3) m between6 and 10, (4) m between 10 and 20, and (5) m between 100 and 150, which occurs in the two countries with a single nationwide district, Israel and the Netherlands. Rae never explainswhy he selected these particularfive categories;but with one exception the dividing lines make good sense, since they coincidewith naturalbreaksin the distribution of the magnitudes. The exceptionis that the naturaldiscontinuity in the range between single-memberdistricts and a district magnitude of 10 occurs at about m = 5 instead of m = 6. Another small improvementis to make 1.1 the dividing point between the two lowest categories. This takes care of the few deviations from single-memberdistricts in the pluralityand majority countries, which are all essentially singlememberdistrict systems; and it removes Rae'sawkwardgap betweenm = 1 and m = 2. A final helpful adjustmentis to expand the 10-to-20-seat category to 10-to-25 so that it can accommodatethe Austriancase, which since 1971 has had a magnitudeof slightly higher than 20.7 Averagedistrictmagnitudeis definedas the averagenumberof seats (or representatives) per district; and it can be calculated easily by dividing the total number of seatsin the legislatureby the numberof districts.Only with regardto two-tierdistrictingsystems do we run into complications. The solution proposed by Rae (p. 21) is that "this problem is tentatively resolved by counting the total numberof districtsat both levels in the computation
486

This is an unsatof averagemagnitudes." isfactory solution because its effect is to make the averagemagnitudelower than it would be on the basis of either the lowtier or the high-tier districts by themselves. Forinstance,from 1945to 1970the Austrian 165-member legislature was elected in 25 low-tier and 4 high-tierdistricts. Counting the lower tier only, the average magnitudewould be 6.6; counting the highertier only, it would be 41.2. The most reasonable assumption is that the "true" magnitude should be somewherebetween6.6 and 41.2. Accordingto Rae's definition, however, it is 5.7-well below 6.6. Rae states that his solution is it. only tentative,but he neverreconsiders Since I do not want to prejudgethe effect of a higher tier of of the superimposition districts (which is often a single national district),my own tentativeand temporary solution is to calculateaveragemagnitude solely on the basis of the lower tier. Rae's (p. 124) hypothesis concerning the impact of two-tier or "complex"districting is eminently plausible: "One would expectcomplexdistricting-the use of two tiers of districts-to increaseproportionality at any given level of magnitude." However, Rae immediatelyrejects this hypothesis. Without supplying precise figures, he states that "the four systems based on complex districtingIcelandic, German, Austrian, and Danish-do not behave in accordance with this expectation. For two of them, the German and the Danish, disproportions are below average for their respective levels of district magnitude. But for the Austrian and Icelandicsystems, complex districting associates itself with greaterthan-averagedisproportions,given their levels of district magnitude."It is worth noting that Rae'sconclusion, for once, is based on a controlled test. He does not simply compare all complex-districting cases but cases with all simple-districting controls for districtmagnitude. Nevertheless, Rae's negative finding is

ElectoralLaws
plex districtingor simple districting.The one case of near-nationwidedistricts is Austria since 1971, which has used only two higher-tierdistricts. For the reasons just outlined, the 1946-59 Icelandicelectoral system should not be included among the strong complex-districting cases. When the disproportionalitiesof the two types of districtingsystems are contrasted,the strongcomplex-districting systems do indeed turn out to be more plausible explanation is available. . . , it proportional within each category of must be concludedthat complex district- magnitudethan the other systems. I can now proposea modificationin my ing does not producegreaterproportionality than simple districtingat any given tentative definition of average district level of district magnitude."In fact, the magnitude: strong complex-districting systems should be classifiedone category deviant Austrian and Icelandiccases do have some special features that may ex- higher than what their lower-tiermagniplain theirunexpecteddisproportionality. tude would entitlethem to. I shall call this In the Icelandic electoral system from their adjusteddistrict magnitude. For in1946 to 1959, more than half of the seats stance, since 1970 Sweden has had a at the lower tier were in single-member lower-tier magnitude of 12.5 seats but districtsand almost a third in two-mem- also a single, nationwide, higher tier; its ber districts. The few adjustment seats adjusted magnitude is therefore in the 100-to-150-seat category together with available at the national level could not caused Israel and the Netherlands.The relationneutralizethe disproportionalities by these low magnitudes. The Austrian ship between adjustedmagnitudeand diselectoral system from 1945 to 1970, proportionality is considerably stronger than that between magnitude,as I tentadespitethe country'srelativelysmall size, tively defined it, and disproportionality. had four higher-tierdistricts instead of The last column of Table 1 presents the one national district. The Austriancase suggestsa generalex- average degrees of disproportionalityin planation. Complex-districtingsystems the five classes of adjusted magnitude. typically try to restrict the benefits The greatest differenceagain occurs beobtainableat the highertier to partiesthat tween the plurality-majority systems (which are also the single-member district have received a minimum number of catevotes or seats at the lower tier. Since this systems) and the least-proportional entailsdiscrimination againstthe smallest gory of PR systems. Within the family of amongthe difparties, it also places a limit on the pro- PR systems, the differences effect of two-tierdistricting ferent categoriesof magnitudeare roughportionalizing systems;and it means that the higher-tier ly similar to the differencesfound earlier districtsmust be large, preferablynation- among the several formulas. These findwide, in orderto make the election result ings are in sharp contrastwith Rae's (pp. 114-24) conclusion that district magniappreciably more proportional. A stronger hypothesis than Rae's would tude is a more potent force than the electhereforebe that complex districtingwith toral formula,which has become the conventional wisdom (Sartori 1986, 53, 66; nationwide-or almost nationwidehigher-tierdistricts (which can be called see also Taagepera and Shugart 1989, 112-25). strong complex districting) yields greater What is the effect of each of the indeproportionalitythan other forms of com487

for severalreasons. One is unsatisfactory that he fails to recognize and include Belgiumand Italy as complex-districting cases. Second, he suddenly shifts from electionsto countriesas his unitsof analysis and therebyignoresthe majorelectoral law changesin Icelandin 1959 and in Germany after the 1953 election. Third, his conclusion is so surprisingthat a closer look at the deviantcases is calledfor. Rae (p. 124) gives up too quickly: "Sinceno

American Political Science Review Vol. 84

pendentvariables-electoral formulaand district magnitude-on the dependent variable of proportionality when the other independentvariable is controlled? This question is relevant only for the PR systems since the non-PR systems do not vary with regard to magnitude. Table 1 also shows the indices of disproportionality of all PR cases classifiedaccordingto both effectivedistrictmagnitudeand electoral formula.When districtmagnitudeis betweenthe held constant, the differences formulasare reducedbut not at all eliminated. When we control for electoralformula, sizable differences between the magnitudeclasses remain in the d'Hondt systems; but the differencesare less impressive for the other formulas. The reason why these relationshipsnow appear less strong is that effective district magnitude and electoral formula are themselvescorrelated.As the numbersin parentheses in Table 1 show, the most and pure proportionalformulas(LR-Hare occur only in the two most Sainte-Lague) proportional classes of adjusted magnitude (those largerthan 10 seats); and the majority of the least-proportional formulas (d'Hondt)are used in the least-proportional magnitudegroups (less than 10 seats). BecauseTable 1 has severalempty cells, its evidence is somewhat ambiguous. Moreover,thereare a couple of cells with values that are out of monotonic order. Nevertheless,the overallpatternis one of interactionbetween formula and magniimpactof tude but also of the independent each of them on the proportionalityof the electoral outcome. Further evidence of this patternis providedby Figure1, which shows the changes in disproportionality due to major electorallaw changesin the same country. For instance, Sweden first shifted from d'Hondt to modified SainteLagueand subsequentlyto a much higher adjusted magnitude. Because countryare now also conspecific characteristics trolled, we have a rigoroustest of the ef488

fect of magnitudeand formulaon disproportionality.The nine majorelectorallaw changesall produce the expectedchanges in the degree of disproportionality.With the exception of Israel'sreturn from LRHare to d'Hondt after the 1969 election, all of these major changes entailed shifts to more proportionalrules. To sum up, district magnitude was found to have a very strongeffect on proportionality, confirming Rae's conclusion. This effect remains strong-albeit not as strong-when the electoralformula is held constant, a control that Rae fails to apply. And contraryto Rae, complexdistricting was found to make the election result considerablymore proportional.

Formula,Magnitude, and Multipartism

of the effectof the elecThediscussion
toral formula and district magnitude on proportionalityhas paved the way for a more expeditioustreatmentof their effect on multipartism.The reason is that proportionality is the logical link that connects these electoral system variables to the number of parties: the more proportional the electoral system is, the more favorable it is for small and new parties and hence the more it will allow-some would say encourage-a large numberof parties.Of course, thereare also otherimportant causes of multipartism,particularly the numberand depth of the cleavand Grofman ages in a society (Taagepera 1985). We should thereforenot expect a very strong correlation between magnitude and formula on the one hand and multipartismon the other. Rae (p. 141) correctly points out that "party systems are influencedby many variables-social, economic, legal, and political. [The]electoral law ... is to be counted only one of many determining forces. And it is, secondly, impossible to sort out all the contributingfactors, or to assigneven ap-

Electoral Laws
Figure1. Average Disproportionalityand Average Number of ElectiveParties of ElectoralSystems in Change, 1945-85

ElectoralFormula
LR-Hare and Pure S.-L. LR-Droop and Mod. S.-L. d'Hondt
Germany 1953 D= 7.40 N= 3.31

1.1-5

Iceland
1946-59 D =9.72 N = 3.66 Norway 1953-85 D=7.06 Norway Germany 1945-49 1957-83 D= 10.80 D=3.31 N=3.87 N=2.56 Sweden 1948 D = 5.25 N = 3.35

V3

oD 5-10 .+CJ) cm

N=3.77
Austria 1945-70 D = 5.04 N = 2.49 Sweden 1952-68 D = 3.42 N = 3.30

10-25

Denmark 19361598 = 3.21 D

N =4.01

Iceland D = 3.94 N = 3.84

100-150

Denmark Austria 1971-84 1971-83 D=2.86 D=2.00 N = 5.49 N = 2.31 Israel 1951-69 D=2.53 N=5.10

Sweden 1970-85 D=2.39 N = 3.52

Israel 1949 D = 5.80 N=5.40

Israel 1973-84 D=5.88

---

489

American Political Science Review Vol. 84

Table 2. Average EffectiveNumbersof ElectivePartiesin ElectoralSystems Classifiedby ElectoralFormulaand AdjustedDistrictMagnitude,1945-85
LR-Droop, LR-Imperiali, Modified and Sainte-Lague, STV-Droop 3.04 (1) 3.19 (3)
-

Adjusted District Magnitude

1-1.1

LR-Hare and Pure Sainte-LaguE

--

d'Hondt
-

Plurality and Majority

2.95 (6)

Alla
2.95 (6)

1.1-5 5-10
10-25

4.04 (2)

3.96 (3) 4.10 (6)

4.27 (3)
-

3.73 (4) 3.80 (9)

4.18 (5)

100-150 Alla

4.30 (3) 4.20 (5)

3.75 (2) 3.35 (6)

4.76 (2) 4.20 (14)

2.95 (6)

4.28 (7) 3.79 (31)

Source:Basedon datain Mackieand Rose1982, 1983, 1984;Mackie1985, 1986;Ornstein,Mann,andMalbin 1987, 47-48; and Luxembourg, ServiceCentralde la Statistique 1984, 120-21. Note: The numbersof cases on which the percentages are basedare in parentheses. Franceof 1951-56. aExcept

proximate weights to them. Worse yet, electoral laws are themselves shaped by party systems." Notwithstanding these cautionary words, Rae (p. 98) reportsa big difference for pluralityand majority in multipartism systems (which are also the singlememberdistrictsystems)on the one hand and PR multimemberdistrict systems on the other-2.17 and 3.70 parties respectively. My figuresshow more modest differences-2.95 and 4.07 parties.The main reason why Rae finds such a small effective number of parties in plurality and majority systems appears to be that he counts in terms of elections:the 2 French Fifth Republicmultipartyelections (5.56 parties) are overwhelmed by the 10 almost purely two-party U.S. elections (2.04 parties). On the other hand, if we focus on the plurality systems-Canada, New Zealand, the United Kingdom, and the UnitedStates-and excludethe French and Australian majority systems, the degree of multipartism is considerably lower-only 2.54 parties. Since the differencesin disproportionality among PR systems are smallerthan those between PR and plurality-majority, we should also expectmore modest differ490

Rae (p. enceswith regardto multipartism. 106) reports that in accordancewith his hypothesis, highest-averageformulas are associated with lower multipartismthan largest-remainder formulas-3.57 and 4.00 partiesrespectively.However, when the missing1945-64 data and the 1965-85 data are addedand the numerousmisclassifications are corrected, the difference disappears-4.39 and 4.40 parties. My own findings are hardly more impressive, however. They are summarized in Table 2. The d'Hondtsystemshave exactly as many partiesas the most proportional LR-Hare and pure Sainte-Lague systems (see the bottom row of the table). Also unexpectedly,the intermediatesystems in terms of proportionality (LRDroop, etc.) are considerablylower than the d'Hondtsystemswith regardto multipartism. We must reject the hypothesis that within the PR family, formula and multipartismare related. The findingsconcerningdistrictmagnitude are somewhatmore positive (see the last column of Table 2) but not as strong as Rae (p. 121) suggests. Rae reports a monotonic relationship between magnitude and the numberof partiesfor the PR systems-from 3.18 parties in the lowest

Electoral Laws
category of magnitudeto 4.65 parties in the highest, an increaseof approximately 1.5 parties. My reanalysisshows a much smallerincrease-only about .5 parties. When the two independent variables are cross-tabulated, the weak relationship between magnitude and multipartismis not weakenedfurther.The monotonic increasein multipartism as magnitudegoes up is even maintainedwithin each of the three groups of formulas, but the differences are small. Controlling for magnitude makes the influenceof the electoral formulaon multipartism look worse. For each magnitude class, multipartism is lower insteadof higherin the intermediate categorythan in d'Hondtsystems. And in two of the four classes of magnitudethat allow a comparisonof d'Hondtwith LRHareand pureSainte-Lague, multipartism is higher, not lower, in the d'Hondt systems-again contraryto our expectations based on the formulas'tendenciesto disproportionality. Similar, mainly negative, evidence is provided by Figure1, which applies the within-countrytest of the effect of major electoral law changes-applied earlier to changes in disproportionality-to changes in multipartism. Whereas the electoral law changes produced the expectedchangesin disproportionality in all nine cases, thereis no commensurate pattern for multipartism:the evidence runs five to four against the hypothesis.Of the five changes in electoral formula, only one yields the expected change in the degree of multipartism;and of the five shifts to larger magnitudes (one case, Austria, entails a change in both formula and magnitude)threeyield more, and two yield less, multipartism. These findings suggest that the impact of disproportionalityon the number of parties is even weaker than our modest expectationshad led us to believe. When we take a direct look at this relationship, we find that it is in the expecteddirection but that the correlationcoefficient is an
491

almost negligible -.10.

shallpointout shortly,thisnegative finding needs to be qualified in several respects.

However, as I

Ballot Structureand Multipartism

With regardto the impactof ballot structure on multipartism, Raeformulates hismostoriginal hypothesis. Based on the distinction between ordinal ballots,which "allowthe voterto favormorethanone partywith his mandate," and categorical ballots,which"require thatthevotergive his mandate to a singleparty,"Rae (p. 126) hypothesizesthat ordinal ballots "alloweach voter'smandateto be dispersed among several parties, thereby producinga sort of microfractionalization"and that electionsheld with such ballots,and hencewith repeated microfractionalizations, "produce more fractionalizedelective party systems than would be found underother elections." he findsthattherelationship However, is theotherway around: hisordinal systems havean average effective number of only 2.94parties compared with3.23parties in categoricalsystems. His conclusionis, 'My theoryis absolutely wrong" (p. 127). Hereagain,we musttakea critical look at Rae's methods and procedures.He makestwo seriousmistakes of classification. His ordinal ballot systems are Australia, Ireland, Luxembourg,and Switzerland. Thefirsttwo usepreferential ballotsthatallow-in theAustralian case, require-the votersto rank-order candidatesandparties to the voters' according preferences. Luxembourg andSwitzerland uselist PRsystems butgiveeachvoteras manyvotes as thereare seatsin the district and allow the voter to distribute these votes over two or more parties, equallyor preferentially. Thesefourelectoral systems are correctly classified. However,Rae misclassifies the German andFrench Fifth Republic as catesystems

American Political Science Review Vol. 84

Table 3. Average EffectiveNumbersof ElectivePartiesClassifiedby AdjustedDistrict Magnitudeand Ballot Structure,1945-85
AdjustedDistrictMagnitude Ballot Structure Categorical Ordinal All 1-1.1 2.54 (4) 3.78 (2) 2.95 (6) 1.1-5 4.28 (2) 3.18 (2) 3.73 (4) 5-10 3.99 (8) 4.00 (2) 3.99 (10) 10-25 4.34 (4) 3.53 (1) 4.18 (5) 100-150 4.28 (7)
-

All 3.92 (25) 3.64 (7) 3.85 (32)

4.28 (7)

Source:Basedon data in Mackieand Rose1982, 1983, 1984;Mackie1985, 1986;Ornstein,Mann,and Malbin 1987, 47-48; and Luxembourg, ServiceCentralde la Statistique1984, 120-21. Note: The numbersof cases on which the percentages are based are in parentheses.

gorical; they are clearly ordinal. The French two-ballot system allows the voters to vote for differentparties on the first and second ballots and quite oftenwhen candidatesare eliminatedor withdraw after the first ballot-require that they do so. In Germany, each voter has two votes, which may be cast for different parties;this ordinal ballot system has existed since the 1953 election.8 When these misclassificationsare correctedand also, as usual, the missingdata are added and electoral systems rather than elections are used as the cases for analysis, the findingsare still negative(see the last column of Table 3). The ordinal systems have less, insteadof more, multipartism than categorical systems; although the differences are relatively small. In trying to explain his negative result, Rae (p. 129) speculatesthat "ballot structureconstitutes a rather weak variable. It may be much less importantthan electoral formulae, district magnitude, and, more likely still, the sociopolitical forces which underlieparty competition." He continues, "Unfortunately,the available data allow no definitive test of this notion, and it must be left unverified." This is an unnecessary defeatist stance, since we can test the influence of ballot structureon multipartismwhile controlling for formula and magnitude. Table 3 shows the average effective
492

numbersof parties under categoricaland ordinal ballots when magnitude is controlled. In the multimemberdistrict systems, my findings are still mainly negative: two of the three pairs of values still run counterto the hypothesis,and the one positive instanceis only marginallypositive. However, in single-memberdistrict systems, ordinal ballots result in a dramatically higher degree of multipartism than categoricalballots. The differenceis roughly that between a two-and-a-halfparty system and an almost-four-partysystem. The explanationis not difficultto find. The categoricalsystemsare plurality systems in which the incentive to vote or insincerely,in ordernot to strategically waste one's vote, is strong. This resultsin greater support for the large parties by voters and by politiciansand a reduction in the effectivenumberof parties.The ordinal ballot removes most of the risk of wasting one's vote; hence it encourages both sincerevoting and multipartism. When we control for the electoral formula, we find the same importantdifference among the plurality-majorityformulas, since these coincide with the single-memberdistrict systems, and the same mainly negative result for the PR systems.9 The link between ballot structure and multipartism applies only to districtsystems, but it is a single-member strong link.

Electoral Laws
Conclusion The two electoralsystemvariables,formula and magnitude,have a strong effect on electoral disproportionality(considerably stronger than reportedby Rae) but only a weak effect on multipartism (much weaker than Rae suggests).The reason is that disproportionality-the intervening variablebetween the electoralsystem and multipartism-is only a weak predictorof the numberof parties.However, this conclusion has to be qualifiedin threeimportant respects. In the first place, the hypothesis that disproportionalityshould reduce multipartism is based on the assumption of strategicbehavior. When smaller parties are expected to be discriminatedagainst, voters, as well as politicians, political activists, and money givers will favor the largerparties (Gunther1989; Riker1986, 33-41). The findingon the effect of ballot structure suggests that the majority systems should be excluded from the hypothesizedrelationship.They are characterizedby high disproportionality; but becauseof their"ordinal" ballot structure, they do not encouragestrategicbehavior and hence do not discourage multipartism.10 The case of Franceof 1951-56 should also be excluded because its disproportionalitywas designedto favor the smaller parties-those in the ideological center-and thereforeencouragedinstead of discouraged multipartism (Campbell 1965, 113-26). Forthe remaining 29 cases, the coefficientof correlationbetween disproportionalityand the effective number of partiesis now -.29 (compared with an r of -.10 for all 32 cases). The second qualification is that this somewhat stronger relationship mainly reflects the differencesbetween the four plurality systems with high average disproportionality (11.08%) and low average multipartism (2.54 parties)on the one hand and the 25 PR systems with lower disproportionality (4.89%) and higher
493

multipartism(4.00 parties) on the other. However, within the PR category,thereis virtually no relationshipbetween the two variables;the correlationis .02. This finding casts doubt on Sartori's (1986, 54) assertionthat within the family of PR systems, a distinctionmust be made between highly proportional, hence "feeble," systems and less proportional "strongfeeble"systems, with the latterin an intermediateposition betweenthe "feeble" systems and "strong" pluralitysystems. The third qualificationis that since the focus has been on gauging the electoral system's long-range effect on the party has been consistentsystem, multipartism ly measured in terms of the effective numberof elective parties. It is logical to expecta strongerimpactof disproportionality on the effectivenumberof legislative parties because the latter is affected not only by strategicresponsesto the expectation of discrimination againstsmallerparties-what Rae (pp. 67-68) calls the eleceffect and Maurice toral system's "distal" Duverger (1963, 226) its "psychological" effect-but also by the immediate,"proximal" (Rae) or "mechanical" (Duverger) effect of the translationof votes into seats in the particularelection under observation. For the 29 cases, the correlationbeand the effective tween disproportionality numberof legislativepartiesis a stronger -.45. This means that disproportionality explainsonly about 8 % of the variancein elective multipartismbut about 20% of that in legislativemultipartism.The latter percentage must be interpreted as the combined effect of the psychological and mechanical factors, but it still indicates that the mechanical factors are the strongerinfluences." To sum up, electoral systems display wide and predictable differences in disproportionality, smaller differences in multipartismthat are the direct effect of disproportionality,and even smaller differences in elective multipartismdue to strategiccalculationsby elites and voters,

AmericanPolitical Science Review Vol. 84

Plurality systems are an exception: their accounts, via the high disproportionality effect of strategicbehavior, for theirrelatively small effective number of elective parties. Strategicvoting is neutralizedin majority systems, which can therefore combine high disproportionality with high elective multipartism.And the substantial differencesin disproportionality among PR systems are apparently not large enough to produce either commenin strategicbehavioror, suratedifferences as a result, commensuratedifferencesin elective multipartism.
RichardG. Niemi, Matthew S. Shugart,and Rein for helpfulcomments.And I should like Taagepera for the financialsupport to expressmy appreciation of the GuggenheimFoundationand the German MarshallFundof the UnitedStates. 1. My referenceswill be to the 1971 edition of Rae'sbook. 2. Raewas so kind as to makehis originaldatafor 1945-64 availableto me, but I decidedthat it was preferableto use the election data in Thomas T. Mackie and RichardRose's (1982) standardhandbook in order to facilitate replication by other andbecauseI also usedthe latterdatafor researchers handbookinthe 1965-85 period.The Mackie-Rose dudes the elections held through 1981. For the 1982-85 electionsI turnedto the annualupdatesin the European Journal of Political Research by Mackie and Rose (1983, 1984) and Mackie (1985, 1986). The resultsof U.S. House of Representatives elections, not reportedby Mackie and Rose, were takenfromOrnstein,Mann,andMalbin1987. Since voting figuresprovidedby Mackie the Luxembourg and Rose fail to adjustfor the unequalnumbersof votes that voters have in differentdistricts,I used 1984. ServiceCentralde la Statistique Luxembourg, 3. It may be arguedthat comparedwith 1, D errs in the oppositedirection:it has a slight tendencyto of systems with exaggeratethe disproportionality index manyparties.However,the two-major-parties (definedas the averagevote-seatsharedeviationof the two largest parties), which was expresslydesignedto steera middlecoursebetweenD and1, is in 1985, practicemuch closer to D than to I (Lijphart with both 10-12). A greatadvantageof D compared indexis that it does not I and the two-major-parties decisionseitheron cutoff points (for entailarbitrary I, should the cutoff point be .5%, which is Rae's criterion,or shouldit be 1, 2, 5, or 10%?)or on the of the two numberof partiesto be counted.(Instead largestparties,why not the three largestpartiesor perhaps just the largest party?) Mackie and Rose (1982)and Rose (1984)use a slightlydifferentform indexof disproof D. Insteadof usingthe "negative" portionality,they preferto thinkmorepositivelyin that is, instead termsof an indexof proportionality; of D, they subtractD from 100%. is represented 4. If the index of fractionalization by F and the effectivenumberof partiesby N, they are relatedas follows: N 1/(1 - F). The F values to 2.0, 2.5, and 3.0 partiesare .50, corresponding .60, and .67. 5. In ContinentalEurope, the Droop quota is quota. Strictly often called the Hagenbach-Bischoff speaking, there is a slight differencebetween the two. If v standsfor the total numberof districtvotes and m for the districtmagnitude,the HagenbachBischoffquota is v/(m + 1), usually roundedup, whereasthe Droop quota is definedas v/(m + 1) + and Natkiel 1, roundeddown if necessary(Leonard 1987, 3). Therefore,the two quotasdifferif v/(m +

Notes
This is a revisedand muchshortenedversionof a of WorldCongress at the fourteenth paperpresented the International Political Science Association 1988.1 shouldlike to acknowl(IPSA),Washington, edge the assistance and advice, especially with regard to electoral laws and election data, that I in the comparhave receivedfrom my collaborators ative and collaborativeproject of which this is a product: Don Aitkin (AustralianResearchCouncil, Canberra),Asher Arian (Tel Aviv University, Israel), Thomas C. Bruneau(Naval Postgraduate School, Monterey), Ivor Crewe (University of Essex),WilfriedDewachter(CatholicUniversityof (CatholicUniverLeuven,Belgium),A.-P. Frognier sity of Louvain, Belgium), William P. Irvine (Queen's University, Kingston, Canada), W. K. Jackson (Universityof Canterbury,Christchurch, New Zealand), Gary C. Jacobson (University of California,San Diego), MarkkuLaakso(University of Kuopio, Finland),RafaelL6pezPintor (Autonomous University, Madrid), Thomas T. Mackie (Universityof Strathclyde,Glasgow), George Th. Mavrogordatos (University of Athens), Sten S. of Oslo), DieterNohlen (UniverNilson (University Uni(Queen's O'Leary Cornelius sity of Heidelberg), versity, Belfast), Jea-LucParodi (Fondation Nationale des Sciences Politiques, Paris), Mogens Pedersen (Odense University, Denmark), Anton Pelinka (University of Innsbruck), Bo SArlvik (GothenburgUniversity, Sweden), Yasunori Sone (Keio University,Tokyo), Alberto Spreafico(University of Florence),Bjorn S. Stefansson(Agricultural ResearchInstitute, Reykjavik),Jfirg Steiner of North Carolina,ChapelHill), andJan (University Verhoef (Ministryof Housing, Physical Planning, The Hague).I am also grateful andthe Environment, and to Andre Blais, to the IPSA panel participants Keith Bybee, David Collier, Richard Gunther, 494

ElectoralLaws
1957 on as two separatecases because they differ 1) happensto be an integer;but in mass elections with regardto districtmagnitude. wherev is measuredin thousandsof votes, this dif9. In spite of these negative findings for PR ferenceis so tiny that it can be safely ignored. To systems, Rae's hypothesis still has a good deal of makematterseven more complicatedand (unfortunately) confusing, the term Hagenbach-Bischoff plausibility,and perhapsit should not be rejected completely.Forinstance,it seemsquiteclearthatthe methodis sometimesalso used to describea divisor smallbut cruciallyimportant FreeDemocratic party system, as in Switzerland. This method merely (FDP)in West Germanyhas benefitedsubstantially representsa shortcut in the application of the from its appeal to the voters to give it theirsecond d'Hondtformula,and Itsresultsareexactlythe same votes (whichare actuallythe moreimportantof the as those of d'Hondt. For all practical purposes, two votes); the FDP has in fact been called therefore, the two can be regardedas identical. to my descrip- (somewhat sarcastically) the "party of second Finally,I have to add a qualification 1987). It is virtuallycertainthat quota, whichis usedexclusively choice"(Dittberner tion of the Imperiali withoutthe ordinalballot, the partywould not have in Italy. It has been definedas v/(m + 2) since the faredso well and thatmultipartism wouldhave been 1958election;but it was evenlower-v/(m + 3)-in reduced. the 1948and 1953electionsand, on average,slightly 10. Thisis especiallyclearin the case of the French higher in 1946, when v/(m + 2) applied to the Fifth Republic. The Australian ordinal majority districtsand v/(m + 1) (in effect, higher-magnitude the sameas the Droop quota)in the smallerdistricts system does not behave very differentlyfrom the pluralitysystems. (Carstairs1980, 157-59; Spreafico 1983, 188-90, 11. For the 25 PR systems (excludingFranceof 194). betweendisproportionali6. In orderto classifythe difficulttwo-tiersystems 1951-56), the correlation ty and legislativemultipartism is still only a very between to electoralformula,I distinguish according weak -.17. two basictypes. In one type the districtsat the lower level are used for the initial allocationof the seats; but the final allocation takes place at the higher References level, often the nationallevel, on the basis of all of the votes cast in all of the lower-tierdistrictsthat Balinski, Michel L., and H. Peyton Young. 1982. district.Thus, the togethermakeup the higher-level Meeting the Ideal of One Fair Representation: formulaat the higherlevel is clearlythe moreimporMan, One Vote. New Haven: Yale University tantone. Belgium,Denmark,Germany,and Iceland Press. (1946-59) fit this type. (Sweden from 1970 on is anotherexample;but since it uses the same formula Campbell, Peter. 1965. French Electoral Systems since 1789. Hamden,CT: Archon. and Elections at both levels, it does not representa classification 1980. A Short History Carstairs,AndrewMcLaren. problem.Thisalso appliesto Icelandsince1959, but Europe.London: Systemsin Western of Electoral not to the earliercase of Iceland).The second type Allen & Unwin. foruses one of the varietiesof largest-remainders Dittberner,Jurgen.1987. FDP-Partei der zweiten mulasin the lower-tierdistricts;but insteadof alloWahl: Ein Beitragzur Geschichteder liberalen votes in seats to the remaining catingthe remaining Partei und ihrer Funktionenim Parteiensystem these districts, all remainingvotes and seats are Opladen:Westdeutscher. der Bundesrepublik. districts. to and allocatedin higher-tier transferred Duverger, Maurice. 1963. Political Parties: Their Here the formulaat the lower level is decisive. No and Activity in the ModernState. Organization formulais able to favor the largerover higher-level New York:Wiley. the smallerpartiessystematically,since the parties Gunther, Richard. 1989. "ElectoralLaws, Party with the highest totals of remainingvotes are not Systems, and Elites: The Case of Spain." the largestparties.Rather,the lower-tier necessarily AmericanPoliticalScienceReview 83:835-58. formuladetermineshow many seats will be availS. 1980.A Theoryof PartiesandElecable at the higherlevel, which is of crucialimpor- Katz,Richard JohnsHopkinsUniver-. toralSystems.Baltimore: tance to the smallerparties. Only LR-Hareat the of the secsity Press. Examples lower tier is fully proportional. Laakso, Markku,and Rein Taagepera.1979. "The ond type are Austria and Italy. Becausesome of Number of Parties:A Measurewith 'Effective' countries underwent these six complex-districting PolitiApplicationto WestEurope.Comparative changes in formula, district magnitude, or ballot cal Studies12:3-27. structure,there are 11 additionalcases of electoral Leonard,Dick, and RichardNatkiel. 1987. World systemswith which the hypothesiscan be tested. Atlas of Elections:VotingPatternsin 39 Democbut welcomebyproductof these 7. An unintended racies.London:Hodder& Stoughton. adjustmentsis that the few erroneous magnitude valuessupplied by Raeno longerentailmisclassifica- Lijphart,Arend. 1985. "TheFieldof ElectoralSys-, tems Research: A Critical Survey." Electoral tions. Studies4:3-14, 8. I treatthe 1953electionand the electionsfrom

495

American Political Science Review Vol. 84

Lijphart,Arend. 1986. 'Degrees of Proportionality of ProportionalRepresentationFormulas."In Electoral Lawsand TheirPoliticalConsequences, New ed. Bernard Grofmanand Arend Lijphart. York:Agathon. Loosemore,John, and VictorJ. Hanby. 1971. "The Theoretical Limits of Maximum Distortion: Some Analytic Expressions for Electoral Systems." British Journal of Political Science 1:467-77. 1984. Luxembourg. ServiceCentralde la Statistique. "Les Elections lgislatives de 1945 A1984."Bulletin du Statec 30:119-31. Rae, Douglas W. 1971. The Political Consequences

Mackie, Thomas T. 1985. "GeneralElections In Western Nationsduring1984."European Journal

of Political Research 13:335-39.

Mackie, Thomas T. 1986. "GeneralElections In

Western Nations during 1985." European Journal of Political Research 14:695-97.

Mackie, Thomas T., and RichardRose. 1982. The

International Almanac of Electoral History. 2d

ed. New York:Factson File.

Mackie, Thomas T., and Richard Rose. 1983. "Gen-

eral Electionsin WesternNations during1982."

Research European Journal of Political 11:345-49. Mackie, Thomas T., and Richard Rose. 1984." Gen-

eral Electionsin WesternNations during1983."

Research European Journal of Political 12:335-42. Nohlen, Dieter. 1978. Wahtsysteme der WeltDaten und Analysen: Ein Handbuch. Munich:

Piper. Ornstein, Norman J., Thomas E. Mann, and

Michael J. Malbin. 1987. Vital Statistics on Congress, 1987-1988. Washington: Congressional

of Electoral Laws.2d ed. New Haven:Yale University Press. Riker,WilliamH. 1986. "Duverger's LawRevisited.' In Electoral Laws and Their Political Consequences, ed. Bernard Grofman and Arend Lijphart. New York:Agathon. Rose, Richard.1984. "Electoral Systems: A Question of Degreeor of Principle?" In Choosingan Electoral System: Issues and Alternatives, ed. Arend Lijphart and Bernard Grofman. New York:Praeger. Sartori,Giovanni.1986. The Influence of Electoral Systems: Faulty Laws or Faulty Methods7"In Electoral Lawsand TheirPolitical Consequences, ed. BernardGrofmanand Arend Lijphart. New York:Agathon. Spreafico, Alberto. 1983. "Systznes klectorauxet system politiqueen Italie depuis 1945."In Les modes de scrutin des dix-huit pays libres de l'Europeoccidentale-leurs resultatset leurs effets compares: elections nationales et europ6enes, ed. JacquesCadart. Paris: PressesUniversitaires de France. Taagepera,Rein, and Bernard Grofman.1985. "RethinkingDuverger's Law:Predicting the Effective Numberof Partiesin Pluralityand PRSystemsParties Minus Issues Equals One." European Journalof PoliticalResearch13:341-52. Taagepera,Rein, and Matthew S. Shugart. 1989. Seatsand Votes:TheEffectsandDeterminants of ElectoralSystems. New Haven: Yale University Press. Van den Bergh,G. 1955. Unity in Diversity:A Systematic Critical Analysis of All Electoral Systems. London:Batsford.

Quarterly.

Arend Lijphartis Professorof Political Science, Universityof CaliforniaSan Diego, La Jolla, CA 92093.

496