19

ELECTORAL SYSTEMS IN
COMPARATIVE PERSPECTIVE

WILLIAM M. DOWNS
Georgia State University

E lections are central to the functioning of democ-

ratic systems, and as such they have been the
focus of extensive political science research for
centuries. Scholars and practitioners seek to understand
the variation in choices of different electoral systems
politics, such understanding is an example of the crucial
connection between theory and practice in political science.
A first level of comparison identifies the different
types of elections designed to determine national execu-
tive power. Presidents and other chief political execu-
cross-nationally. They also try to isolate the impact of tives may be elected through direct or indirect means. In
those choices on a range of individual-, institutional-, and direct elections of presidents and presidential-type exec-
system-level outcomes. Those outcomes include the qual- utives, voters cast ballots for one of the eligible candi-
ity and breadth of representation; size and polarization of dates. A candidate can win outright in a single round of
political party systems; citizen participation and voting voting by garnering an absolute majority of the ballots
behavior; and government, as well as system, stability. cast; however, when no one candidate captures 50%
Much of the research conducted on electoral systems and plus one of the eligible votes, a runoff round is held at a
elections has evolved from theoretical and empirical work subsequent date, with the top two finishers from the first
on the United States and other established Western round squaring off. This model of direct executive elec-
democracies (especially those in Europe), but consider- tion (exemplified by countries such as France, Russia,
able effort in recent decades has been devoted to under- Poland, and Argentina) has the putative advantage of
standing elections in transitioning and new democracies providing a wide range of candidates from which voters
globally. Although elections do take place in nondemoc- may choose in the first round. If a second round proves
ratic polities, they usually fail to be free, fair, or competi- necessary, the system then yields a winner supported by
tive and therefore typically fall outside the domain of an absolute majority of those turning out for the runoff.
comparative research. What is clear is that the increasing Indirect election of a country’s president or other execu-
sophistication of theoretical and statistical tools available tive leader, alternatively, entails voters’ selecting other
to political scientists (along with an expanding universe of persons (electors), who will then determine the winner.
cases against which to test expectations) has resulted in The United States, for example, uses an indirect mecha-
important advances in our understanding electoral sys- nism whereby voters choose presidential electors, who
tems and elections. Because electoral processes and out- then comprise the Electoral College, which then votes
comes exert such profound effects on the real world of on who will become the next president. That process

159
160 • COMPARATIVE POLITICS

leaves open the possibility that the person chosen to be ian governance should, according to this logic, produce
president by the Electoral College is not the same person more stable and effective polities. Detractors, however,
who secured the greatest number of votes among the find that aggregating district-level winner-take-all elec-
general population. Elsewhere, directly elected parlia- tions into a national whole can produce skewed represen-
ments (either an upper house or both houses in joint ses- tation in the legislature. For example, a party that runs a
sion) constitute the arena in which presidents are consistent and respectable second place throughout the
indirectly selected; this occurs in Germany and Italy, for country but that fails to win any single district would be
example. When Westminster style parliamentary systems excluded from taking seats in the legislature. Such a sys-
(i.e., those modeled after the British House of tem, then, has the potential to underrepresent small parties
Commons) use votes by legislators in plenary sessions to in a democracy.
approve (or remove) prime ministers as heads of gov- The alternative to single-member, winner-take-all sys-
ernment, they are engaging in indirect elections of polit- tems of electing representative assemblies is one based on
ical executives. Taking the selection of national leaders proportional representation (PR) in multimember districts.
out of the direct control of voters represents the skepti- In PR systems, the goal is to have the percentage of a
cism that constitutional architects have for the general party’s seats in the legislature reflect the percentage vote
population, and it can provide an apparent elite-level share captured by that party in the general election. The
check on the sentiments of mass electorates. party securing 25% of the vote would, accordingly, be
The second major dimension along which political sci- rewarded with 25% (either exactly or approximately)
entists compare elections is the method of voting for of the legislative seats. Here ballot structure, which shapes
legislative assemblies. Indeed, examining legislative elec- how voters cast their votes, becomes critically important.
tions across countries reveals considerable variation in Ballots can be categorical or ordinal. The categorical bal-
such key dimensions as district magnitude, electoral for- lot structure allows a single either–or choice of one candi-
mulae, ballot structure, and the use of electoral thresholds. date. By contrast, the ordinal ballot structure gives voters
District magnitude refers to the number of candidates who the opportunity to vote for more than one candidate. In
will be elected to a legislature from any given con- some ordinal ballots, political parties devise rank ordered
stituency, and the basic distinction here is between sys- lists of candidates to determine which persons ultimately
tems that rely on single-member districts and those that claim those seats. In this closed party list system, citizens
employ multimember districts. District magnitude is usu- vote only for a party in a multimember constituency (often
ally studied in tandem with the system’s chosen electoral the whole country), whereas in an open party list system,
formula, which represents the particular mechanism for voters can choose from a published list or select an indi-
translating votes into legislative seats. Such mechanisms vidual candidate. The closed party list mechanism clearly
are most frequently of the plurality, majoritarian, and pro- vests considerable power in the hands of party leadership.
portional varieties. In the single-member district system, a Often, PR systems will set a minimum threshold (5% in
country is divided into discrete electoral districts from Germany, for example) that parties must clear in order to
which one individual will emerge as the elected represen- win seats. Electoral thresholds are an increasingly com-
tative. This system normally relies on a plurality rule, mon way for PR systems to limit the entry of minor (and
meaning that the candidate with the most votes wins sometimes extremist) parties into legislatures. Thresholds
(regardless of whether that candidate has captured an normally require a minimum percentage of votes or a min-
absolute majority). As such, single-member district sys- imum number of seats in order for a party to gain seats in
tems are often deemed first past the post systems and a legislature. Thresholds vary, with some countries opting
also constitute a winner-take-all approach that provides to set the bar low (Israel, for example, at 2%) and others
no electoral prize for coming in second. The United States raising it to high levels (e.g., Turkey, at 10%). Numerous
and the United Kingdom are among the countries where varieties of proportional representation exist, each with
the single-member plurality system has a long-standing different counting and procedural mechanisms. One such
history; however, a range of countries elsewhere—includ- variety is the single transferable vote method. By this
ing Canada, Ghana, and India—have adopted the same method, voters rank candidates preferentially, and if a
method. Others, most notably France, employ a single- voter’s first-choice candidate has already cleared a set
member district system with two rounds of voting. In such threshold and does not need additional support to win, then
cases, individual candidates can win outright in the first that vote is transferred to a second choice. This process,
round with an absolute majority of votes cast, or they can exemplified most clearly by Ireland, is designed to avoid
secure the plurality of votes cast among eligible candi- “wasting” votes.
dates in the second-round runoff. Single-member district Although there is a tendency among political scientists
systems are defended by their advocates as those that can to classify electoral systems in democratic countries into
enhance clarity of responsibility and democratic account- either the majoritarian or proportional camp, the reality is
ability by giving citizens in each district one individual to that many hybrid or mixed systems exist in between those
whom credit or blame can be assigned. The clarity and two types. The additional member system, for example,
accountability that are supposed to accompany majoritar- combines elements of conventional first-past-the-post
 Electoral Systems in Comparative Perspective • 161

systems with some characteristics of party-list propor- Indeed, in this view, elections constitute the minimum
tional systems. In this combination, voters get two votes: necessary requirement for democracy. For Schumpeter
The first helps allocate seats to single-member con- (1942), democracy is “that institutional arrangement for
stituencies, and the second goes to a party list. The per- arriving at political decisions in which individuals acquire
centage of second or party-list votes won by a party the power to decide by means of a competitive struggle
determines the party’s overall number of representatives, for the people’s vote” (p. 269). Likewise for Huntington
and the number of seats won in single-member districts (1993), democracy is defined most essentially by the fair
is “topped off” to match that overall percentage. This and periodic voting procedures that select a country’s
method finds use in elections to Germany’s Bundestag, leaders. Przeworski, Alvarez, Cheibub, and Limongi
New Zealand’s House of Representatives, and the (2000) also view contested elections—that is, those in
Scottish and Welsh Assemblies in the United Kingdom. which there is ex ante uncertainty and ex post irreversibil-
The presumed advantage of this mixed member system ity—as the litmus test for democracy. Others, such as
approach is that proportionality is ensured, and at the Dahl (1971), counter that such a thin, minimalist, or pro-
same time, a directly accountable representative for each cedural definition of democracy-as-elections fails to
constituency is also identified. It is also said to allow account for other necessary conditions, such as the pro-
strategic voters to express support for an individual tection of civil liberties and the actual responsiveness of
politician while not necessarily endorsing that candi- government policies to voter preferences. Whether suffi-
date’s political party. Disadvantages are said to include cient or not, elections typically figure as necessary condi-
the creation of two (potentially unequal) classes of tions for the existence of democracy.
politicians, with those elected under the second-ballot Theoretical work on elections and comparative elec-
topping off beholden not to the voters but to party lead- toral systems has largely focused on (a) the relationship
ers instead. between electoral rules and the size and polarization of
While elections in democratic settings constitute the political party systems, (b) the tendency of certain elec-
overwhelming preponderance of all voting processes toral systems to impact voter turnout and citizen participa-
studied by political scientists, it is important to note that tion, (c) the congruence between electoral verdicts and
nondemocratic systems (e.g., authoritarian and semiau- government policy, and (d) the potential for electoral sys-
thoritarian systems) can also employ electoral mecha- tems to predispose new or transitioning systems to success
nisms. Such regimes may organize controlled and or failure. Political scientists developing theory in each of
uncontested elections as a means of mobilizing mass these areas represent some of the main ontological camps
endorsement of a national leader or a single-party legisla- in the discipline, such as structuralists, rationalists, and
ture. Doing so can provide symbolic legitimacy for the culturalists. As such, attention has been devoted to formal
ruling elite, and it may neutralize popular discontent by rules, voter preferences and behavior, and the contextual
creating the false appearance of citizens having a say in influence on system choice and outcomes.
the affairs of their country. For example, while the The causal relationship between electoral rules and the
Communist Party monopolizes power and controls politi- nature of a country’s political party system has animated
cal processes in China, direct elections of village-level scholarly interest for decades. Perhaps the most famous
offices do take place, as do indirect elections for people’s proposition, tested repeatedly since its early assertion by
congresses above the local level. The one-party Soviet Duverger in 1954, is that plurality elections using one-
Union held its own brand of uncontested elections, as did ballot single-member districts will favor the creation of
Iraq under Saddam Hussein. Brazil under military rule two-party systems whereas proportional representation
orchestrated compulsory voting in tightly controlled elec- rules with multimember districts will lead to multiparty
tions, even though the frequency of blank and spoiled bal- systems. Duverger went further to posit that a majority
lots often suggested popular rejection of the process. vote on two ballots increases the likelihood of a multi-
Semicompetitive, hegemonic party systems such as party system as well as the necessity of postelection
Egypt’s hold elections in which there is little a priori coalition formation. It is rare indeed that causal relation-
uncertainty of the outcomes; there is, in such cases, ships in political science theory elevate to lawlike status,
some element of choice and voter expression. Although but in this case, “Duverger’s Law” has achieved consid-
nondemocratic variants of the electoral process illustrate erable staying power. The logic guiding Duverger’s
more about a regime’s methods of system control than assertions depends on what are conventionally deemed
they do about representation, responsiveness, and mechanical effects and psychological effects. The
accountability, they clearly merit attention. mechanical effects highlight the underrepresentation of
third (and fourth, and fifth, etc.) parties, which is likely
to occur over time in a single-seat legislative district
Theory requiring an outright plurality or majority vote. Given the
mechanical impediments to minor party success, voters
Some democratic theorists view elections as a central— who support minor parties then have psychological
if not the central—component of liberal democracy. incentives not to “waste” their votes and may often cast
162 • COMPARATIVE POLITICS

ballots against their preferred candidate in a strategic advantage” (p. 18) over the majoritarian alternative because
effort to exercise some influence over the most likely they encourage broad cross-party bargaining to form a gov-
winner in the two-party competition. Such claims have ernment and to pass legislation. Such bargaining should
spawned much subsequent work, and not a little dissent. produce governments that include the median legislator,
Sartori (1968) extended Duverger’s assertion of a link who is, in turn, close to the median voter. The median voter
between proportionality and party system size, specify- is located at the middle of a political system along most
ing that district magnitude (i.e., the number of seats in a issue dimensions, such that one half of the electorate is
district) is the single best predictor of the effective num- positioned to the political left and the other half is posi-
ber of political parties in a district. Riker (1982) chal- tioned to the political right. The median legislator is like-
lenged Duverger’s hypothesis about PR and multipartism wise the elected representative located such that half of the
by contending that, if true, we should see a recurring other legislators are to the left and the other half are to the
increase in the number of parties over time rather than right, politically. Electoral systems that produce govern-
party system stability or modest decreases in the effective ments proximate to the median voter should, therefore, be
number of parties (as most frequently occurs in practice). more responsive to policy preferences. Proportional elec-
Debate over the relationship between choice of electoral toral systems should also give greater policy influence to
system and party system size is important, given the opposition parties, making for a more inclusive process
propensity to view two-party majoritarian countries as of policy making.
more stable than those with polarized multipartism. Theories underpinning our understanding of electoral
If electoral rules biased in favor of two-party systems rules and their consequences can have extremely impor-
are theorized to bring gains in terms of system stability, tant practical applications. While much effort is devoted
then those rules favoring proportionality figure promi- to understanding how and why established democracies
nently in political science theories that attempt to explain tinker with their electoral systems to enact reforms or
citizen engagement, voter turnout, and representativeness alter a range of political outcomes, even greater attention
of legislatures. According to Lijphart (1994, 1999), majori- has been directed in recent decades to the role of elec-
tarian and plurality electoral systems dilute citizen enthu- tions in facilitating regime change. Indeed, one of the
siasm and voter turnout because so many supporters of growth areas in political science literature addresses the
minor parties conclude that casting their ballots will have prospects for successful electoral engineering. Given
little to no impact on electoral outcomes, government for- that the last decades of the 20th century witnessed transi-
mation, or policy choices. Conversely, proportional sys- tions from communism, apartheid, and other forms of
tems with low thresholds for representation and large autocratic rule, alternative theories about the prospects
district magnitudes should increase the chances that for implanting democracy through institutional engineer-
smaller parties from across the ideological spectrum will ing have become increasingly salient. Likewise, nascent
be able to secure voice and seats in the legislature. With postauthoritarian systems in early 21st-century hot spots
that greater likelihood of electoral success for minor and such as Iraq and Afghanistan have emerged as testing
even fringe parties, voter efficacy and incentives to cast grounds for the discipline’s theoretical assertions. Norris
ballots should be increased. Voter turnout is “an excellent (2004) identifies two theoretical traditions—rational
indicator of democratic quality” (Lijphart, 1999, p. 284), choice institutionalism and cultural modernization—that
and PR systems are theorized to be superior to their purport to explain the possibilities for electoral engineer-
majoritarian counterparts in generating democratic gains ing on human behavior. In the rational choice institution-
in this area. As part of an overall inclusive and consensual alism approach, political parties adopt discernibly
approach to democratic governance, proportional electoral different strategies based on the nature of electoral
systems should also improve citizen satisfaction with the thresholds and ballot structures. Preference-maximizing
political system, ceteris paribus. citizens likewise should be expected to respond differ-
A third major area of theoretical work on comparative ently to alternative electoral rules. If correct, this logic
electoral systems has evolved around the presumed cor- would predict that rule-based incentives will shape con-
respondence between voting outcomes and public policy. sistent patterns of behavior; therefore, changing those
If democracies are to be responsive to the preferences of incentives through electoral engineering “should have the
the public, then periodic voting should work to translate capacity to generate important consequences for political
the “will of the electorate” into identifiable policy choices. representation and for voting behavior” (Norris, 2004,
Scholarship in this area builds on the majoritarian– p. 15). By contrast, the cultural modernization approach
proportional dichotomy to examine citizen control over— suggests that deep-rooted cultural habits arising from
and influence on—government policy making. Powell processes of social modernization place real limits on the
(2000) explores elections as “instruments of democracy” potential of formal rules to alter behavior in systemati-
and distinguishes a proportional vision of “citizen influ- cally meaningful ways. This culturalist argument is often
ence” from a majoritarian vision of “control.” He contends employed to explain why wholesale introduction of elec-
that “proportional influence designs enjoy a surprising toral rules into culturally divided, postconflict settings so
 Electoral Systems in Comparative Perspective • 163

frequently fails to produce short-term transformations of a proportional representation system with compulsory
individual behavior. voting and a 5% threshold for representation in the federal
Chamber of Representatives produced enough support to
grant parliamentary seats to 11 parties in the June 10,
Applications and Empirical Evidence 2007, general election. The largest among those, the
Christian Democratic and Flemish Party, claimed only
Political scientists have endeavored to assemble an abun- 18% of the 150 seats in parliament and could therefore not
dance of empirical evidence in support of their theoretical form a government by itself. Protracted negotiations com-
claims. Perhaps nowhere has greater effort been extended menced after the election, and 196 days later, the best the
than in tests of propositions about the linkages among Belgian parties could do was constitute an interim care-
electoral laws, party systems, and coalitional incentives. taker government. That interim government lurched
Countering an alternative hypothesis that underlying soci- along, with further negotiations taking another 79 days
etal cleavages are the primary agents determining size and before the parties could agree on a full-fledged new gov-
polarization of party systems, a literature has evolved (cf. ernment. That government, in turn, failed to finish out the
Cox, 1997; Lijphart, 1994; Rae, 1967; Riker, 1982; year. Although electoral rules biased in favor of majori-
Sartori, 1968; Taagepera & Shugart, 1989) to contend that tarianism typically yield governments that combine cer-
electoral laws have their own independent effects. tainty with disproportional representation, those rules
Duverger’s notions about first-past-the-post, single-ballot favoring multiparty outcomes tend to better reflect the
elections tending to produce two-party majoritarian sys- dispersion of political preferences throughout the country
tems find extensive application in the United States, as but may also add considerable uncertainty to the govern-
well as the United Kingdom. In elections for the U.K. ment formation process.
House of Commons and the U.S. Congress, the evidence Evidence also exists on the relationship between elec-
seems to suggest a compelling link between electoral rules toral systems and the production of such democratic
and strong, stable, two-party government. Electoral struc- goods as high voter turnout and citizen satisfaction.
tures in the United States, for example, help explain the Where the electoral rules reduce the costs (e.g., time and
consistent failure of third parties to mount successful cam- effort) to citizens of registering and voting, we should find
paigns. This winner-take-all system has, though, placed greater turnout. Similarly, where party choices available
such significant importance on the drawing of district to voters are more extensive we should expect to see ele-
boundaries that the pernicious practice of gerrymander vated turnout. Finally, voter efficacy—the belief that cast-
ing—consciously redrawing the lines to ensure a majority ing a ballot can actually impact the government formed
for one party—emerged as part of American politics. and the ultimate policy direction taken—should be
Although smaller parties have been able to win parliamen- directly related to turnout at elections. According to
tary seats in the United Kingdom, their ultimate represen- Norris (2004), “Institutional rules do indeed matter: vot-
tation in the House of Commons is highly disproportionate ing participation is maximized in elections using PR, with
to their overall support in the electorate, and they have lit- small electoral districts, regular but relatively infrequent
tle chance at becoming the party of government or forcing national contests, and competitive party systems, and in
a coalition. To illustrate, the perennial third-party Liberal presidential contests” (pp. 257–258). There is also evi-
Democrats won 22.1% of the vote in Britain’s 2005 gen- dence to support theoretical contentions that the type of
eral election but secured only 9.6% of the 646 seats in the electoral system can impact the opportunities for women
House of Commons. Tony Blair’s Labour Party, having and minorities seeking to earn a legislative seat or execu-
won only 35.3% of the votes nationwide, nevertheless cap- tive office. Among established democracies, the countries
tured 55.2% of the seats in parliament and 100% of the that consistently sit atop comparative rankings of the pro-
cabinet positions in government. portion of women winning seats in national parliaments
Single-member-district plurality systems normally are Sweden, Iceland, Finland, and the Netherlands. Each
provide rapid certainty after an election about who will country employs some form of proportional electoral
govern and who will constitute the opposition. However, rules with low thresholds, and in each it is routine for
systems that introduce even a modicum of proportionality women to constitute more than 40% of national parlia-
likewise introduce an element of uncertainty into the gov- mentary representation. Findings such as this, it should be
ernment formation process. Proportionality (especially noted, must also take regional political culture and other
when combined with low thresholds in multimember dis- potentially intervening factors into consideration.
tricts) does increase the number of effective parties in the In his study of democratic performance in 36 countries
political system. When no single political party secures an from 1945 to 1996, Lijphart establishes empirically that
outright legislative majority, the postelection period electoral systems favoring consensus-oriented governance
becomes one marked by formal negotiations as well as yield gains in citizen satisfaction. When the rules of the
backroom deals between parties jockeying to join a gov- electoral process encourage multipartism and coalition
erning coalition. The case of Belgium is illustrative. There building, the policy preferences of the median voter have a
164 • COMPARATIVE POLITICS

greater chance to be represented in the government of the the two-round system for the 1988 parliamentary election,
day. Lijphart’s data show that the distance between gov- the National Front’s 9.7% of the first-round vote translated
ernments and median voters is highest in majoritarian sys- into only one seat!
tems (with the United Kingdom representing the high end In Japan, major reforms occurred in 1994, when the old
of the scale) and lowest in more proportional systems system of single nontransferable votes (allowing one
(with Ireland and its single transferable vote system pro- choice per voter in elections for three to five district rep-
ducing the narrowest gap). Because in PR systems elec- resentatives) was scrapped and replaced by a mixed-mem-
toral “losers” often have a chance to join postelection ber system. The new Japanese system for electing the
coalitions—and due to the frequent proportional represen- House of Representatives combines first-past-the-post
tation of opposition parties on legislative committees— single-member districts (for 300 seats) with PR party-list
Lijphart is able to find a statistically significant difference seats (200) in an “attempt to craft a competitive two-party,
between citizen satisfaction in countries with alternative issue-oriented politics and a cleaner, more efficient gov-
electoral systems. Lijphart’s study corroborates earlier ernment” (Norris, 2004, p. 5). Whereas Japanese politics
work by Klingemann (1999), who found that Danes and prior to the reform consisted mainly of one dominant
Norwegians—each with highly proportional systems— party (Liberal Democrats) regularly overwhelming a
scored the highest levels of democratic satisfaction. handful of opposition, the new hybrid of majoritarian and
Given the volume of empirical applications of existing proportional approaches (most analogous to the system in
political science theoretical work on elections, it is not sur- Russia) aims to create a polity with alternating parties in
prising that there is a foundation of cases demonstrating power. In New Zealand, at roughly the same time, reforms
how changes in electoral rules actually impact voter to replace the long-standing first-past-the-post system
behavior and system characteristics. Indeed, the lessons of came to fruition. There, a mixed-member proportional
major 20th-century electoral reforms in three countries— system now allows 70 of the 120 national parliamentary
France, Japan, and New Zealand—are instructive. The seats to be elected directly in single-member districts,
French case illustrates how constitutional architects can try with the remainder coming from party lists in a style sim-
to contain what are perceived to be the excesses of pro- ilar to Germany’s. The addition of proportionality to New
portional representation. Those designing the 1958 Fifth Zealand’s electoral system—endorsed by a majority of
Republic sought to use electoral rules to avoid reproducing citizens in a binding 1993 referendum—has had a quick
the fleeting and weak multiparty coalition governments and dramatic impact. Whereas the average number of
that had plagued the Fourth Republic from 1946 to 1958 political parties gaining seats in New Zealand’s national
and brought the system to the brink of collapse. The new parliament was just two during the 1946-to-1993 period,
two-round, single-member district system established in in the five elections since introducing the mixed-member
1958 encouraged broad political party competition in a system, an average of seven parties has secured represen-
first round and awarded National Assembly seats to all tation. Electoral engineering, at least in this case, seems to
candidates winning an outright majority. Absent a major- have achieved the end envisioned for it.
ity, all candidates receiving at least 12.5% of first-round Perhaps nowhere is political science research into com-
votes could then contest in the runoff election, in which a parative electoral systems more salient than in countries
plurality would suffice for victory. In practice, this runoff attempting to transition away from authoritarianism. The
mechanism encourages the weakest candidates to volun- cross-national lessons available to architects of new sys-
tarily stand down in favor of a better positioned candidate tems are always imperfect, as transporting a model from
closest to them on the left–right ideological spectrum and one country to another without sensitivity to local condi-
to have their supporters cast their second-round ballots for tions and histories is a formula for failure. However, such
that person. This system has effectively preserved France’s comparative learning does take place, and most new
multiparty system while simultaneously creating a stable electoral systems today are adaptations and amalgams of
two-bloc system of parties on the moderate left and right. those found elsewhere. When elections were held in
The runoff system often means that parties with meaning- December 2005 to constitute a post-Saddam Iraqi Council
ful support nationwide may still fail to secure national leg- of Representatives, a proportional party-list system deter-
islative seats, as has been the case with the far-right mined 230 of the total 275 seats in 18 multimember dis-
National Front party. Indeed, when the French tinkered tricts (governorates). An additional 45 compensatory seats
with their electoral laws in the 1980s, it became apparent were then allocated to political entities that did not win any
how decisive the rules can be for representation. In 1986, seats outright in the governorates but that did clear a mini-
the Socialist government of President François Mitterrand mum national threshold. Also worthy of note is that Iraq’s
opted to change from the two-round system to a single- electoral law requires at least 25% of the members of the
round proportional one in hopes of dividing the right wing parliament to be women. In Afghanistan, post-Taliban elec-
opposition parties. As a result, the National Front’s 9.6% tions have struggled to secure domestic and international
of the vote earned it 35 of the 577 national legislative legitimacy. The 2005 elections for Afghanistan’s lower
seats. When party strategy changed and France reverted to house of parliament employed the single nontransferable
 Electoral Systems in Comparative Perspective • 165

vote method in 34 multimember constituencies. Candidates, translation of electoral verdicts into governmental policy
however, ran independently because parties and lists were becomes significantly more indirect. Moreover, the dis-
not recognized by the governing law. As in Iraq, the Afghan tribution of policy portfolios across multiple parties blurs
system reserved a number of seats (at least 68 of the total the lines of accountability and increases the difficulties
249) for women. At the executive level, the Afghan presi- for voters who wish to reward or punish the incumbents.
dent is elected by absolute majority in a two-round system For example, the citizen asked to evaluate with one vote
similar to that employed in France. the performance of a three-party coalition government
may find it hard to express support for that government’s
fiscal policy (headed by a Conservative Party finance
Policy Implications minister) while rejecting its policies on education
(headed by a Christian Democratic Party education min-
The choice of election system can potentially impact the ister), as well as those on immigration (headed by a
quality and kind of policy pursued by an incumbent gov- Nationalist Party interior minister).
ernment. If elections are the essential ingredient in repre- If elections are central to the functioning of democratic
sentative democracy, then presumably there should be political systems, then another set of policy implications
some apparent connection between the will of the people can be found in the promotion of democratization through
as expressed through elections and the policies they elections and electoral reform. The foreign policies of
receive from the subsequently invested government. If many established democracies, as well as those of inter-
citizens are engaging in issue voting, as some research has governmental organizations and donor agencies, are inti-
consistently found, then it is important to gauge whether mately tied to this kind of promotion. The conduct of free
the governments they get are actually responsive to those and fair elections is frequently the litmus test for legitimacy
issues. If our fundamental expectations about democracy in the eyes of the democratic international community, and
require a close connection between elections and policy everything from diplomatic recognition to commercial rela-
outcomes, then the reality may sometimes disappoint tions can hinge on the successful holding of competitive
(Ginsberg & Stone, 1996). As Downs (1957) contended, elections. As such, governmental entities such as the United
political parties adopt policies in order to win elections Nations or the European Union will regularly send election
rather than win elections in order to adopt policies. The monitoring missions to observe voter registration and the
achievement of public policy goals may actually be casting of ballots to gauge openness, extent of fraud, and
instrumental to the more power-seeking ambitions of par- incidents of intimidation. Nongovernmental organizations,
ties and politicians. such as the Carter Center, have also played this monitoring
Political scientists therefore examine the ways in role in Africa, Latin America, and Asia. International finan-
which different electoral systems hold officials account- cial institutions, such as the World Bank, also incorporate
able for their fidelity to campaign promises once in elections into decisions about granting development assis-
office. Indeed, elections provide a kind of ex post tance funds to countries in need. This process of political
accountability for policy pledges. The more a particular conditionality is the stipulation of the conduct of democra-
model of election creates the perception among elected tic elections as a necessary occurrence prior to the alloca-
officials that those they claim to represent will oust them tion of foreign aid. Such conditionality has been part of the
for poor past performance, the stronger the democratic Structural Adjustment Programs implemented by the World
accountability linkage is said to be. Although several Bank and the International Monetary Fund in exchange for
scholars (e.g., Lijphart, 1999; Powell, 2000) demonstrate lower interest loans to developing countries. Critics of
that citizen satisfaction and the correspondence between these policies contend that tethering development assis-
median voters and the policy positions represented in a tance to political reforms is tantamount to threats that will
legislature are enhanced by consensual, proportional rep- lead to the rapid importation of electoral mechanisms that
resentation, there is also reason to find that majoritarian ultimately fail to take root.
systems provide the kind of clarity that voters need to
hold leaders accountable for policy choices. In a majori-
tarian, winner-take-all system, if a party campaigns on Future Directions
the basis of very clear policy pledges, wins the election,
and then proceeds to depart dramatically from its public As democracy expands (and sometimes contracts) across
promises (what political scientists call engaging in moral the globe, research on elections likewise adapts. Political
hazard), the voters should be able to easily identify this scientists continue to focus on formal rules and designs; on
lack of fidelity and then “throw the rascals out” at the individual-level attitudinal and behavioral responses to
next electoral opportunity. By contrast, in electoral sys- those formal mechanisms; and on the connections among
tems characterized by proportional representation, the elections, party systems, and policy outcomes. One of the
likelihood of multiparty coalition governments forming particular growth areas for future research in the area will
after protracted negotiations is great. In such cases, the be that addressing referenda and other forms of direct
166 • COMPARATIVE POLITICS

democracy. Referenda can take many different forms, with electoral systems have produced important findings about
some being ad hoc and others constituting routine and regu- the impact of alternative majoritarian and proportional sys-
lar procedures. They are advocated on the logic that circum- tems and the many hybrid models in between. Duverger’s
venting the normal representative institutions in favor of early assertions about the influence of ballot and district
direct votes by the entire electorate will encourage more cit- type on the size and character of political party systems
izens to become better informed and more involved in the have risen to the status of “law” in the discipline and
democratic decision-making process. Singling out a policy spawned subsequent and more sophisticated theorizing. A
choice for a decision by the people should, moreover, grant considerable body of evidence now exists to help explain
clarity to the direction desired by citizens; in contrast, the how the choice of electoral system can influence the qual-
normal process of bundling multiple policy choices within ity of a country’s democracy. Introducing or reforming
legislative bills makes it hard to achieve such clarity. electoral rules can alter citizen participation and satisfac-
Finally, it is presumed that decisions arrived at through ref- tion, can enhance or diminish the congruence between
erendum elections will enjoy much greater legitimacy than voter preferences and public policy outputs, and can have
those achieved through competition or cooperation of the profound consequences for system stability. Electoral
political elite. With greater legitimacy, we should expect, in engineering as such is one of the clearest issue areas in
turn, the greater likelihood that those policies are success- which political science research speaks directly to decision
fully implemented. Referenda, plebiscites, and citizens’ ini- makers. Indeed, given that for centuries revolutions have
tiatives take place in many countries: Switzerland uses the been fought and blood spilt for the right to live in a democ-
people’s initiative with relative frequency (more than 400 racy, it is imperative to understand how elections can sup-
national referenda since 1945); French and Dutch voters port or undermine transitions from authoritarian rule.
were asked in their respective 2005 national referenda Some of the most tenuous polities around the globe strug-
whether they supported a proposed European Union gle with legitimacy and leadership transitions, and design-
Constitution; and voters in East Timor chose to part from ing an appropriate electoral system shapes—if not
Indonesia in a 1999 referendum. Although the United States determines—those countries’ futures. Theories and data
does not hold national referenda, some states and many assembled for study of the United States, western Europe,
localities do hold frequent initiatives and ballot propositions and other cases of consolidated democracy offer much to
(for example, California’s 1978 Proposition 13 on property the electoral engineer, policymaker, and student observer;
taxes and its 2008 Proposition 8 on same-sex marriage). they cannot, however, be casually transported across the
Future research should build on the increasing rele- globe to nascent democracies without due consideration of
vance of referenda and initiatives to explore the impact of the opportunities and constraints defined by a country’s
question wording and ballot structure on voting outcomes. individual context.
Assumptions about referenda being decided by informed
and engaged citizens need to be tested thoroughly, across
time and across countries. If empirical support does not References and Further Readings
emerge to substantiate that there are informational gains
associated with referenda to a greater extent than in regu- Amorim Neto, O., & Cox, G. (1997). Electoral institutions,
lar elections, then some normative red flags need to be cleavage structures, and the number of parties. American
raised regarding the utility of this form of democratic deci- Journal of Political Science, 41(1), 149 174.
sion making. Fodder for further research comes with the Amy, D. (2000). Behind the ballot box: A citizen’s guide to elec
toral systems. Westport, CT: Praeger.
varying turnout requirements that countries impose for the
Anderson, C., & Guillory, C. (1997). Political institutions and
verdicts of direct democracy votes to be implemented: satisfaction with democracy. American Political Science
Why do different countries place the turnout threshold at Review, 91(1), 66 81.
different levels, and do high thresholds unfairly violate Bogdanor, V. (1984). What is proportional representation?
majorities that fail to reach them? Do voters behave differ- Oxford, UK: Martin Robertson.
ently when referendum elections are merely advisory Bogdanor, V., & Butler, D. (Eds.). (1983). Democracy and elec
rather than binding on the government? Most important, is tions. Cambridge, UK: Cambridge University Press.
there any empirical reason to expect that direct democracy Cox, G. (1997). Making votes count: Strategic coordination in
elections are supplanting conventional political elections the world’s electoral systems. Cambridge, UK: Cambridge
in any meaningful way? These trends and prospects clearly University Press.
deserve the attention of 21st-century political scientists. Crewe, I., & Denver, D. (1985). Electoral change in Western
democracies: Patterns and sources of electoral volatility.
New York: St. Martin’s Press.
Dahl, R. (1971). Polyarchy: Participation and opposition. New
Conclusion Haven, CT: Yale University Press.
Dalton, R., Flanagan, S., & Beck, P. (Eds.). (1984). Electoral
The rules governing and guiding voting are central to the change in advanced industrial democracies: Realignment or
study of contemporary politics. Decades of comparing dealignment? Princeton, NJ: Princeton University Press.
 Electoral Systems in Comparative Perspective • 167

Downs, A. (1957). An economic theory of democracy. New York: Powell, G. (2000). Elections as instruments of democracy:
Harper & Row. Majoritarian and proportional visions. New Haven, CT:
Duverger, M. (1954). Political parties: Their organization and Yale University Press.
activity in the modern state. London: Methuen. Przeworski, A., Alvarez, M., Cheibub, J., & Limongi, F. (2000).
Farrell, D. (1997). Comparing electoral systems. London: Democracy and development. Cambridge, UK: Cambridge
Prentice Hall/Harvester Wheatsheaf. University Press.
Farrell, D. (2001). Electoral systems: A comparative introduc Rae, D. (1967). The political consequences of electoral laws.
tion. New York: Palgrave. New Haven, CT: Yale University Press.
Ginsberg, B., & Stone, A. (Eds.). (1996). Do elections matter? Reeve, A., & Ware, A. (1992). Electoral systems: A comparative
Armonk, NY: M. E. Sharpe. and theoretical introduction. New York: Routledge.
Grofman, B., & Lijphart, A. (Eds.). (1986). Electoral laws and Riker, W. (1982). The two party system and Duverger’s law: An
their political consequences. New York: Agathon Press. essay on the history of political science. American Political
Huntington, S. (1993). The third wave: Democratization in the late Science Review, 76(4), 753 766.
twentieth century. Norman: University of Oklahoma Press. Rule, W., & Zimmerman, J. (1994). Electoral systems in com
Katz, R. (1980). A theory of parties and electoral systems. parative perspective: Their impact on women and minori
Baltimore: Johns Hopkins University Press. ties. Westport, CT: Greenwood Press.
Klingemann, H. D. (1999). Mapping political support in the Sartori, G. (1968). Political development and political engineer
1990s: A global analysis. In P. Norris (Ed.), Critical citizens: ing. In J. Montgomery & A. Hirschman (Eds.), Public policy
Global support for democratic government (pp. 31 56). (pp. 261 298). Cambridge, MA: Harvard University Press.
Oxford, UK: Oxford University Press. Schumpeter, J. (1942). Capitalism, socialism, and democracy.
Lijphart, A. (1994). Electoral systems and party systems: A study New York: Harper & Row.
of twenty seven democracies, 1945 1990. New York: Shugart, M., & Wattenberg, M. (Eds.). (2001). Mixed member
Oxford University Press. electoral systems: The best of both worlds? Oxford, UK:
Lijphart, A. (1999). Patterns of democracy: Government forms Oxford University Press.
and performance in thirty six countries. New Haven, CT: Taagepera, R., & Shugart, M. (1989). Seats and votes: The effects
Yale University Press. and determinants of electoral systems. New Haven, CT:
Lijphart, A., & Grofman, B. (Eds.). (1984). Choosing an elec Yale University Press.
toral system: Issues and alternatives. New York: Praeger. Wall, A., Ellis, A., Ayoub, A., Dundas, C. W., Rukambe, J., &
Norris, P. (2004). Electoral engineering: Voting rules and polit Staino, S. (2006). Electoral management design: The inter
ical behavior. Cambridge, UK: Cambridge University national IDEA handbook. Stockholm, Sweden: International
Press. Institute for Democracy and Electoral Assistance.