Reconciling Approaches in the Study of Mixed-Member Electoral Systems
Abstract and Keywords
The diffusion of mixed-member electoral systems over the last three decades has prompted extensive scholarly attention to their consequences. One approach, labeled controlled comparison, emphasizes how the juxtaposition of two different election rules allows scholars to assess the implications of Duverger’s propositions. Another approach, labeled contamination effects, emphasizes how the simultaneous use of two different election rules creates interaction that potentially undermines the implications of Duverger’s propositions. The literature building upon these two approaches is inconclusive; some research finds evidence of contamination effects and other research does not. This chapter strives to reconcile the two approaches and proposes a research agenda that could uncover why scholarship on mixed-member systems has not reached consensus on their consequences.
1The incentives that electoral systems generate for voters and political actors are well established in the literature (e.g., Duverger 1954; Rae 1967; Riker 1982).2 Foundational research emphasizes how the institutional rules of constituency-based systems, most notably first past the post (FPTP), as well as majority runoff and proportional representation (PR) systems, affect the structure of the party system. Over time, the literature has expanded to address the connection between election rules and different outcomes, and to consider the effects of variation in election rules on those outcomes.3
One form of institutional variation—the mixed-member electoral system—was present at the time this early literature was published but was not subjected to extensive empirical scrutiny as a distinct electoral system. The first long-standing mixed-member electoral system,4 adopted in Germany after World War II, was at times categorized as a form of PR, or called semi-PR largely due to its proportional outcomes (e.g., Sartori 1994). Over the years, many countries and subnational units adopted mixed-member systems so that by the end of the 1990s, they were widespread (Ferrara, Herron, and Nishikawa 2005). While several societies that chose mixed-member systems during their extensive expansion in the 1990s subsequently abandoned them (e.g., Italy),5 they are used in prominent established democracies (e.g., Germany, Japan, New Zealand),6 in transitional societies and new democracies (e.g., Ukraine,7 South Korea, and Taiwan), and in authoritarian states (e.g., Russia).8 The diversity within mixed-member systems (p. 446) prompted seminal taxonomic assessments that established these electoral systems as a unique form (Massicotte and Blais 1999; Shugart and Wattenberg 2001).
The mixed-member system’s combination of different formulas to simultaneously select members of legislative bodies spawned two seemingly distinct approaches. One approach, labeled “controlled comparison,” focuses on how the use of different formulas in the same cultural, economic, and historical contexts and at the same time allows scholars to evaluate the effects of election rules. Another approach, labeled “contamination,” focuses on how the interaction between the component parts of mixed-member systems may alter the incentives and undermine controlled comparisons. Scholars have investigated the effects of mixed-member systems from both perspectives but have not reached consensus on which explanation best accounts for outcomes. In this chapter we argue that these two approaches are not mutually exclusive. Rather, they differ in degree, and the effects vary largely depending on whether or not electoral or legislative behavior is examined. We particularly argue that institutional arrangements of mixed-member systems influence political actors differently at the electoral and legislative levels.
To support this argument, we survey the published literature on mixed-member electoral systems and their effects, categorizing the findings along a spectrum in which the evidence suggests that contamination plays an important role, where the evidence does not support a role, and where the evidence is inconclusive. We further discuss approaches that could advance knowledge about mixed-member systems by evaluating the logic of controlled comparison and contamination and/or extending empirical assessments to better understand how the systems function.
What Is a Mixed-Member Electoral System?
The earliest comparative discussions of mixed-member electoral systems emphasize that they combine different electoral rules in an election for a single legislative body. Reynolds and Reilly (1997) identify mixed-member systems as those that combine PR list systems and constituency-based “winner takes all” systems. Massicotte and Blais (1999), Shugart and Wattenberg (2001), and Ferrara et al. (2005) widen the range of systems that are considered to be mixed. A minimalist definition encompasses any two-component system with PR list and nominal elections, such as FPTP, single transferable vote (STV), and single nontransferable vote (SNTV), overlaying one another in a geographic region.9
The challenge that all definitions of mixed-member systems confront is the wide variation in the institutional rules used to tally votes and allocate seats. Mixed-member systems may vary by the presence or absence of a formal linkage between the components, ballot structure, flexibility to contest in more than one component, and proportion of (p. 447) seats allocated to the different formulas. All of these features could affect incentives underlying contamination or affect the logic of controlled comparison.
Perhaps the most notable difference among mixed-member systems is the degree to which the two components are formally independent of one another for seat allocation. Mixed-member proportional (MMP) systems link the components in the seat allocation process, with the list component typically establishing the overall target distribution for seats. While the specific mechanisms to allot seats vary, parties typically receive all the seats that they win in the nominal component and garner additional seats in the list component. The target proportion of seats that the party receives overall is generally equivalent to the proportion “due” to that party as determined by the list component. Because MMP systems tend to generate proportional outcomes through the seat allocation mechanism up to the limit of district magnitude and thresholds, they have been characterized as a form of PR (Sartori 1994). Unlike a pure PR system, however, some seats are occupied by members of parliament who owe their success to local constituencies, and some are occupied by members of parliament who owe their success to the party’s performance and their list placement.
Mixed-member majoritarian (MMM) systems, by contrast, formally treat both components independently. Parties receive all constituency seats that they win and a proportion of the PR allocation based solely on the PR results. Results in MMM systems tend to be less proportional10 than those in MMP.
Ballot structure also differs among mixed-member systems. In some cases (e.g., Korea prior to 2000), voters’ choices on the nominal and list components are formally tied via a fused ballot, with the party affiliation of the constituency candidate connected to the party list vote. Ballots may also be printed on the same sheet (e.g., Albania) or on separate sheets (e.g., Ukraine), affecting the order in which voters may cast votes for the different races. The amount of information provided to voters may also vary, with some ballots providing the identities of party list candidates or candidate biographical information.
Another important institutional feature in mixed-member systems is the opportunity for candidates to contest races in both components simultaneously. Dual candidacy is permitted in many mixed-member systems like Germany, Japan, and New Zealand, although others such as South Korea, Taiwan, and Ukraine forbid it. Dual candidacy systems can generate dually nominated list winners who lose in the nominal component but win seats via the list component.11 From a strategic perspective, the possibility of candidates who lose in one component winning a seat in another could affect nomination and campaigning tactics.
The nominal and list components also differ across mixed-member systems in terms of the proportion of seats allocated to each component. In many cases, the balance is equal or nearly equal. But in some countries more seats are allocated by the nominal component (e.g., several East Asian countries12), and in others more seats are allocated by the list component (e.g., East Timor 2001 election, prereform Italy). The relative distribution of seats across formulas could change the power of incentives to influence behavior. (p. 448)
Mixed-member systems include a wide range of institutional features; no ideal form of mixed-member system exists. However, institutional variation among mixed-member systems is an important factor to explain differences in outcomes. The next section identifies these differences, discusses how scholars have treated mixed-member systems, and subsequently integrates the issue of institutional variation and theoretical expectations.
Controlled Comparison and Contamination in Mixed-Member Systems
Defining Controlled Comparison and Contamination
Although mixed-member systems were incorporated into the catalog of electoral systems, critical questions remained about how the incentives produced by mixed-member systems would influence the behavior of voters, candidates, parties, and elected legislators. Some scholars took advantage of the unique features of mixed-member systems to evaluate how institutional rules with competing incentives affect the behavior of actors. These scholars claimed that the independence of the nominal and list elections make “natural experiments” possible, since the two parts of the elections are conducted within the same country and any confounding factors, such as cultural, economic, or historical factors, are controlled. In effect, they treated the sections of mixed-member systems as functionally independent.
Figure 22.1 illustrates the basic argument of the “controlled comparison” approach. Under “controlled comparison,” Duvergerian incentives predominate, with the nominal component moderating the number of political parties in competition in constituencies, and the list component encouraging multiparty competition. The selection of legislators using two distinct types of election rules encourages representatives who are chosen in the nominal component to advocate for constituency interests and encourages representatives who are chosen in the list component to advocate for national-level (party) interests. This approach does not deny the possibility of interaction across the components, but suggests that the effects are inconsequential. Contextual features, such as the level of party system institutionalization and experience with democratic practices, influence how election rules affect outcomes more, yielding different results cross-nationally when similar rules are applied (Moser and Scheiner 2012). The studies that follow the controlled comparison approach imply that mixed-member electoral systems can produce the “best of both worlds” by generating local and national representation, encouraging party formation in the list component, and mitigating the size of the party system through the nominal component. However, this outcome is not guaranteed, (p. 449) as the question mark in the title of Shugart and Wattenberg’s (2001)The Best of Both Worlds? suggests.
Other scholars challenged the notion of independence, claiming that the simultaneous use of different institutional rules would generate interaction, or (p. 450) “contamination,” across the nominal and list components. The main consequence of contamination would be to alter the behaviors typically associated with “pure” versions of constituency-based or PR rules, illustrated in Figure 22.2. Instead of operating as functionally independent, the two components of the election system affect one another, generating a system with its own incentives. By the logic of contamination, voters, candidates, legislators, and party leaders take into account behavior in both components when making decisions about entry, voting, and advocacy. As a consequence, candidates may be “oversupplied” in the nominal component as parties that have decided to enter competition in the list component also nominate constituency candidates, even if they are not competitive. Pre-electoral coalitions may mitigate these effects, but the effective number of candidates should be higher than Duverger’s law suggests at the district level for pure FPTP systems. Depending on the costs of party formation, ballot access, and especially the accommodativeness of the system (e.g., MMP with large district magnitude in the list component), candidates may be encouraged to form parties for the list component to take advantage of the resources they may gain, magnifying multipartism.
The logic of contamination is based on political actors’ motivations, especially those associated with party leaders and candidates. Winning a specific seat is a primary goal for candidates; winning as many seats as possible is a primary goal for leaders. If party leaders believe that placing candidates in the nominal component is likely to increase the number of list votes the party receives, enhancing list seat acquisition, they should be motivated to place candidates in the nominal component even when those candidates have little chance of winning the particular seat. Party leaders also have multiple ways to entice candidates with resources available in government, the legislature, and the party (e.g., committee assignments, ranking in PR list, etc.), so that candidates can utilize resources to improve their electoral prospects and political careers (Pekkanen, Nyblade, and Krauss 2006).
In mixed-member electoral systems, candidates and leaders are expected to share the same incentives. Under any mixed-member systems—MMM or MMP—winning a constituency race guarantees a candidate a seat. Particularly for this reason, constituency seats may be perceived as more valuable prizes to acquire in the long term. For candidates with the goal of securing a nominal seat, contesting an election with the expectation that they can build a personal vote over more than one electoral cycle may mitigate concerns about short-term loss, especially if they are in a competitive party list position as well. If candidates adopt a longer time horizon, even candidates with little chance of winning may contest the nominal districts. The presence of these candidates may benefit the party as they may be especially motivated to campaign in their local districts, providing parties with electoral benefits.
A crucial mechanism for contamination is that the presence of such locally oriented candidates in the nominal component can make the party more visible in a given constituency. The visibility of the party is associated with contamination effects in two ways: first, supporters of locally oriented candidates are induced to cast list votes for the affiliated party, and second, party supporters cast straight tickets by voting for the (p. 451) candidates affiliated with the party regardless of competitiveness. Through either of the mechanisms (or both), the nominal component of the elections should witness some inflation in the number of candidates. In other words, party leaders’ motivations combined with ambitious candidates’ motivations promote contamination. We believe that this assumption is reasonable across mixed-member systems, although, as we explain later, the impact may be mitigated by a country’s constitutional structure, a party’s resources, and other factors.
Legislative behavior may also be contaminated like electoral behavior. Traditionally, the literature on electoral systems and parties has argued that candidates who are elected from pure FPTP systems are motivated to work for local interests, while candidates under PR work for broader interests (e.g., Katz 1980). Based on this intuition, a “stylized” version of controlled comparison in mixed-member systems would assume that nominal members are more likely than list members to defect from the party line, move from one party to another, and demand committee positions useful in providing parochial benefits.13 This is because nominal members’ principals (i.e., local constituents) may have conflicting interests with party leaders, who are interested in promoting their public policy platforms. In contrast, list members are more likely to work as loyal agents for party leaders because the power to rank candidates on the list gives leaders control.14
This “mandate divide” between nominal and list members may be ameliorated by the combined incentives.15 In other words, if contamination influences legislative behavior, nominal and list members’ theoretically expected behavioral differences should disappear. The specific hypotheses of contamination effects include expectations that nominal members may not be less loyal than list members in terms of roll-call voting or party switching, and that list members may act like nominal incumbents in committee assignments or constituency services. To put it differently, district incumbents act like list incumbents or list incumbents act like nominal incumbents in an effort to hedge bets. As discussed later, mixed-member systems’ incentive structures per se may not promote legislative contamination. Rather, nuanced but important institutional differences within mixed-member systems seem to alter the motivations of legislators and activate contamination in legislative behavior. We explain the incentives of candidates and parties and their relationship to contamination effects at the electoral and legislative levels in more detail later.
While the controlled comparison and contamination approaches have been characterized as in direct opposition to one another in much of the literature, we treat them as differences in degree in this chapter. Both approaches acknowledge the likelihood of interaction between the components. The controlled comparison approach indicates that the interaction effects are likely to be small and are often eclipsed by other factors that affect behavior. The contamination effects approach recognizes that there are conditions in which the effects may be small or overwhelmed by other factors, but contamination effects depend on the institutional arrangements of mixed-member electoral systems and the effects differ at the electoral and legislative levels. (p. 452)
Controlled Comparison and Contamination in the Literature
The late 1990s to early 2000s witnessed a substantial expansion in the discussion of mixed-member electoral systems16 due in large part to the diffusion of these rules worldwide and the increased interest in studying their effects. While this early literature is generally characterized as constituting two defined schools of thought, they share important features. In this section, we evaluate published research about mixed-member electoral systems that speaks directly to the issues raised by the controlled comparison and contamination effects approaches.
To systematically identify relevant literature, we relied on the Web of Science and Google Scholar, searching several combinations of keywords: “mixed systems17 and contamination,” “mixed systems and interaction,” and “mixed systems and controlled comparison.” By focusing on these terms alone, our search may have omitted research that indirectly speaks to the differences between these two approaches but does not formally link the theoretical approach to the controlled comparison/contamination dichotomy. In addition, some scholarship acknowledges the debate about contamination and controlled comparison but does not pursue an empirical strategy designed to adjudicate between these approaches (e.g., Thames and Edwards 2006) and is not included in our assessment. We also include research conducted by the primary proponents of the two arguments, potentially overstating the volume of evidence supporting particular approaches. Finally, we emphasized journal articles rather than books, although books from both perspectives have been published (e.g., Ferrara et al. 2005; Moser and Scheiner 2012). In sum, while we rely on our collected literature to assess trends in research on mixed-member systems, we must be circumspect in our conclusions as the sample may be biased.
Table 22.1 classifies the publications we gathered in three ways. Articles that find evidence supporting principles of contamination are in the column labeled “contamination is more important.” Articles that find no evidence supporting contamination, or evidence refuting it, are listed in the column labeled “contamination is less/not important.” Articles that produce some findings leaning toward both approaches are labeled as “mixed.” In addition to differentiating articles by their tendencies toward supporting the implications of contamination or controlled comparison, we separate research by the primary dependent variables, including the effective number of parties, list vote share, performance on nominal single-seat districts (SSDs), strategic/split-ticket voting, legislative behavior, and committee assignments. The table includes forty-six articles published between 2001 and 2016.
For most dependent variables, published research that we gathered in our search has not yielded a consensus on the controlled comparison/contamination debate. Twenty-four studies find evidence in support of the contamination approach. Even if we discount publications by the initial proponents of the argument, eighteen studies find evidence of contamination. Twelve studies find that contamination is not present or important to the outcomes, and nine find mixed evidence. This evidence is present in different national contexts, such as Italy, Russia, Japan, Korea, Spain, and Taiwan, as well (p. 453) as wider cross-national samples. Scholars have found the influence of contamination on the effective number of parties (Herron and Nishikawa 2001; Golosov 2003; Nishikawa and Herron 2004; Ferrara and Herron 2005; Ferrara 2006; Lago and Martinez 2007; Lago and Montero 2009; Rich 2015), list vote share (Ferrara 2004; Hainmueller and Kern 2008), SSD performance (Krauss, Nemoto, and Pekkanen 2012; Fortin-Rittberger (p. 454) and Eder, 2013; Shin 2014; Zittel and Gschwend 2008), split-ticket voting (Gschwend, Johnston, and Pattie 2003), and women’s representation (Fortin-Rittberger and Eder, 2013; Golosov 2014). Scholars have also found evidence of contamination in legislative behavior (Herron 2002b; Heitshusen, Young, and Wood, 2005; Bernauer and Munzert 2014; Hennl 2014; Ohmura 2014; Rich 2014; Stoffel 2014) and committee assignments (Pekkanen et al. 2006).
Table 22.1 Contamination Based on Different Dependent Variables
Main Dependent Variable
Contamination Is More Important
Contamination Is Less/Not Important
Effective number of electoral parties
List vote share
An almost equally diverse set of research finds limited or no evidence of contamination effects in electoral behavior. Using the effective number of parties as a dependent variable, scholars have found mixed (Bochsler, 2009) or no evidence that contamination provides leverage to explain variation in outcomes (Moser 2001; Moser and Scheiner 2004; Crisp, Potter, and Lee 2012). Similar findings are associated with other dependent variables like list vote share (Maeda 2008; Karp 2009) and split-ticket voting (Kostadinova 2006). Legislative dependent variables do not show solid support for the contamination argument, with mixed (Crisp 2007; Kerevel 2010; Rich 2012; Rich 2015) or limited effects in legislative behavior (Lundberg 2002; McLeay and Vowles 2007; Becher and Sieberer 2008; Sieberer 2010, 2015; Batto 2012; Olivella and Tavits 2014) and committee assignments (Stratmann and Baur 2002; Centellas 2015). These studies have not found contradictory results through replication, but rather have produced divergent interpretations by posing different questions, using different cases,18 or configuring models with different explanatory variables. Scholars’ varied approaches across the literature on mixed-member systems render a full adjudication infeasible.
Despite these challenges, our goal is to propose some potential mechanisms to develop scholarship that can better answer why we see such divergent results. One notable pattern on the table is the difference between studies emphasizing electoral dependent variables and those emphasizing legislative dependent variables. That is, contamination seems to find more robust support when it emphasizes electoral outcomes and less support when it emphasizes legislative outcomes. In our sample, five out of the nineteen studies in the electoral arena find no or unimportant effects of contamination. Contamination is more weakly associated with legislative outcomes; eight out of twenty-one studies in the legislative arena contend that contamination effects are unimportant. In the next section, we discuss how developing the research agenda on mixed-member systems may address these apparent discrepancies.
Advancing the Research Agenda on Mixed-Member Systems
The divergent outcomes may be produced by faulty analysis, biased data, or some other mistake in the conduct of research. While error is certainly possible, it seems unlikely that such a wide range of scholars analyzing data from different countries and time periods would consistently get the wrong answer. Another possibility that has been (p. 455) hinted at in the literature on mixed-member systems is that the story of contamination or controlled comparison may be more nuanced than current research suggests. That is, the phenomenon may exist at measurable levels only under certain conditions that are sometimes—but not always—present in mixed-member systems. Institutional variation among mixed-member systems, along with other factors that differentiate how the incentives function, may contribute to the disparate research outcomes. We explore ways that scholars can confront the lack of consensus in findings and advance research on mixed-member systems to answer how and under what circumstances contamination may matter.
Proximal and Distal Effects
The underlying logic of the contamination effects approach, noted earlier, is that voters, parties, and candidates pay close attention to both components when making strategic entry or voting decisions, disrupting the incentives that each component individually generates. But variation in the institutional features of mixed-member systems may influence these effects. If empirical results that both demonstrate contamination and challenge its presence accurately reflect the specific conditions scholars have identified (e.g., the particular array of institutional features and the social conditions of the societies under study), how can differences be reconciled?
One possibility is that the underlying theoretical accounts on both sides of the debate are underdeveloped. Later research from both approaches suggests that the argument should be more nuanced: institutional variation (Ferrara and Herron 2005) and social heterogeneity (Moser and Scheiner 2012) may be crucial intervening factors that condition the incentives and responses by political actors. In short, the mere presence of two simultaneous elections may not be enough to transform the incentives into contamination effects.
A potential avenue for further exploration comes from the controlled comparison literature. Scheiner (2008) (and earlier, Rae 1967) notes that electoral systems have proximal and distal effects. To influence outcomes, the proximal effects of electoral rules require few assumptions or limited causal chains, such as Duverger’s mechanical effects. Distal effects, by contrast, involve multiple causal chains that go beyond the electoral system’s mechanical effects.19 Synthesizing this theoretical observation with the study of mixed-member systems requires scholars to more clearly articulate how institutional incentives are connected across the two types of elections and the conditions under which we would expect political actors to respond to these incentives. That is, what are the features of mixed-member systems that could “transmit” contamination?
If we extend this idea of electoral proximity to the logic of contamination, electoral effects are relatively proximal, while legislative effects are relatively distal. The basic requirements of contamination in models of electoral behavior require minimal assumptions, such as the presence of nominal component candidates affecting party visibility. Models of legislative behavior require additional conditions to alter legislative (p. 456) behavior. However, this observation does not specifically define characteristics unique to mixed-member systems that could enhance or mitigate contamination.
One approach to advancing research on mixed-member systems is to more clearly identify how specific features may influence interaction at the electoral and/or legislative levels and to develop empirical strategies to test these expectations. Such specific features include dual candidacy, one ballot systems, and a linkage between the two components (i.e., the MMM/MMP difference). We pay the greatest attention to dual candidacy since we believe it is particularly vital for transmission of contamination effects, but we also briefly discuss the other features.
Dual candidacy provisions that permit a candidate to simultaneously contest in the nominal and list components could be a powerful conduit for the incentives promoting contamination at both electoral and legislative levels. During an election, candidates are directly faced with the challenge of managing a campaign. If they are only contesting in a constituency or on the party list, their efforts can be directed toward efforts to best mobilize the vote to ensure their victory. If the candidate is contesting a constituency race, mobilization and policy appeals would be primarily focused on the district. If a candidate is contesting in PR, the candidate may campaign more widely to represent the party’s national platform. That is, the principals of nominal candidates may be different from those of list candidates. Under such circumstances, two types of candidates under mixed-member systems are likely to possess distinctive incentives if they reach the legislature, making the mandate divide clear at the legislative level.
Once elected, a legislator must contemplate the possibility of running for re-election. The decision, and the rules under which the legislator will seek re-election, may be temporally distant. That is, if the next election is to be held four or five years in the future, many factors could affect how the legislator will contest that later election. Parties may permit legislators to promote local interests that challenge national party preferences during the term of office, but may also demand loyalty on certain key bills or when the election is approaching. In short, legislative behavior is more distal to electoral politics.
However, the potentially weaker distal effects should not automatically produce nonfindings in legislative behavior, since the way that dual candidacy affects decisions by parties and candidates in an election may make legislative behavior more directly connected to parties and legislators’ decisions (years in advance of the next elections). Japan’s MMM system used for the House of Representatives (HoR) provides an example. Candidates for the HoR are allowed to run on both the nominal and list components simultaneously. If they lose in the nominal election, they may gain a seat via the list component if they are ranked high enough. These dually nominated list winners are incentivized to serve local interests because their success as politicians depends more on their performance in the nominal component. Moreover, the MMM system used for Japan’s HoR allows parties to “clump” dually nominated candidates at the same rank on the list component. Since clumped candidates’ list rankings are determined by their vote margins in their nominal districts, (p. 457) this “best loser” provision is expected to intensify the effects of contamination.20 Under MMM, winning a nominal seat enhances a party’s seat acquisition, so a party leader also has the incentive to assign competitive candidates to local districts and have them work as nominal incumbents. Pekkanen et al. (2006), for instance, show that Japan’s dually nominated list winners are more likely to assume positions useful for helping them win back SSD seats, suggesting that their legislative behavior becomes indistinguishable from nominal incumbents’ behavior, leading to behavior consistent with the contamination logic at the legislative level.21
In contrast, the case of Japan’s House of Councillors (HoC) illustrates what happens to legislators’ characteristics if dual candidacy is prohibited. The Japanese HoC uses MMM, like the HoR, but candidates running for the HoC cannot be dually nominated. Candidates for the HoC’s list component, which elects around fifty members on a single nationwide district, tend to be those with close connections to national interest groups, such as trade unions, agricultural cooperatives, and doctors’ associations.22 In contrast to the counterparts in the HoR, members elected on the HoC’s list component are more likely to work as representatives for national policy interests instead of representatives for local constituents.
Barker and Levine’s (1999) study also suggests evidence for legislative contamination under dual candidacy in New Zealand. It is usually the case that list incumbents are assigned “shadow” electorates where the party failed to win a seat (Barker and Levine 1999).
The effects of dual candidacy at the electoral level are more evident because of its proximity. In Japan, for instance, dually nominated list winners are allocated to the districts where they lose and are expected to mobilize local supporters. The presence of these pseudo-nominal incumbents lowers the costs of looking for additional candidates. Therefore, districts containing such incumbents are marked by a substantially larger number of candidates.23 The list component can also exert contamination effects on the nominal component; a nominal candidate’s ranking on the list component affects his or her performance on the nominal component (Krauss et al. 2012).24
Mixed-Member Proportional and Mixed-Member Majoritarian Systems
We anticipate that MMP should promote contamination more than MMM at both the electoral and legislative levels. Because parties’ seat shares are ultimately determined by their list votes under MMP, leaders’ incentives to coordinate their candidates on the nominal component are expected to be weaker in MMP than MMM. Nominal component candidates themselves may also feel relatively free to remain uncoordinated in the districts under MMP, when compared to MMM. Since both party leaders and nominal component candidates are less motivated to coordinate in MMP than in MMM, the number of candidates is expected to be inflated more in MMP than MMM. Indeed, using fifty-three democracies between 1990 and 2001, Nishikawa and Herron (2004) empirically demonstrate that MMP produces a higher effective number of parties on average than MMM. (p. 458)
At the legislative level, nominal incumbents under MMP may act more like list incumbents than those under MMM. From a party leader’s standpoint, nominal incumbents do not have to maintain their seats, because parties’ seat shares are determined by list votes, not by the number of nominal seats they win, except for overhang seats. Thus, compared to their counterparts under MMM, party leaders’ incentives to let nominal incumbents focus on parochial interests are weaker under MMP.25 Similarly, nominal incumbents who were elected under MMP are also more motivated to work for broader interests, compared to those who are elected under MMM, since list elections influence their chance of re-election more than nominal elections. Without having local constituents as their principals, nominal and list candidates share the same incentives with party leaders under MMP, resulting in a blurred mandate divide between nominal and list incumbents.26
To be sure, the possibility of grabbing overhang seats may promote a clearer mandate divide. Party leaders would appreciate extra seats that they may receive beyond their party vote shares and would let nominal incumbents work for local interests. In addition, for nominal incumbents, winning nominal district seats necessarily allows them to maintain their incumbency.27 As a result, nominal incumbents act more like nominal incumbents to cultivate local votes. However, the number of overhang seats is relatively small and is likely to be only relevant to a few candidates from the major parties who have benefited from the mechanical effect of the nominal elections with a high disproportionate seat and vote ratio.
MMP combined with dual candidacy may promote contamination further. By analyzing incumbents’ utility calculations, Stoffel (2014) finds that most incumbents in Germany are motivated to serve broader party interests, since their likelihood of winning a seat is more likely to be determined by party vote shares than nominal vote shares (cf. Krauss et al. 2012). Stoffel suggests that list candidates with a safe slot on the party list and nominal candidates with secure nominal seats work for party interests, indicating the presence of legislative contamination.28
Proportion of the Seats Allocated to Nominal and List Components
The proportion of the seats allocated to the nominal and list components should interact with the institutional features of mixed-member systems in influencing contamination effects. Ferrara and Herron (2005) argue that when seat allocation is more balanced, candidates and parties are likely to face incentives to hedge bets and promote contamination. This argument can be extended to the analysis of contamination at both electoral and legislative levels.
At the electoral level, the effects of contamination appear to become substantively unimportant when one of the components dominates the elections. When more seats are allocated by list components, party leaders are still motivated to place candidates in the constituency elections, since it will mobilize the local votes in those districts. Candidates who are placed in nominal districts are also motivated to win nominal seats, since it will guarantee them those seats that they win. As a result, the (p. 459) effective number of candidates in the nominal districts may be inflated, suggesting the presence of contamination. However, the overall substantive importance of contamination effects may be minimal, given the small number of nominal districts. That is, the number of list incumbents actually changing their strategy to act like nominal incumbents would be small.
When nominal elections dominate with more seats, party leaders may be more motivated to coordinate by forming alliances with other parties and reducing the number of candidates, especially if single-member district systems are used in MMM systems.29 The number of seats that parties may be able to win by coordinating with other parties may yield more seats in the nominal component than placing candidates in nominal districts and gaining extra seats in the PR list component. Candidates are also pressured to exit. As a result, the number of candidates is reduced over time, minimizing the effects of contamination at the electoral level. The substantive importance of contamination should be at its peak when seats are distributed by the two components equally.
At the legislative level, the effects of contamination appear to vary depending on which component dominates. If more seats are allocated in the nominal component, contamination effects are expected to be weak. When the nominal component is dominant, nominal incumbents should serve local interests. An increased number of nominal seats is also associated with smaller-size districts that promote local interests further. In contrast, list candidates should serve broader interests to maximize the chance of winning future elections. However, list incumbents may also be motivated to serve local interests, since the chance of running for nominal seats in the next election may be high, because more seats are allocated by nominal elections. Under such circumstances, nominal and list candidates share similar interests, promoting legislative contamination. However, since the number of list candidates is small, the effects of contamination are substantively minimized. We will only observe a few list incumbents actually changing their behavior and acting like nominal incumbents. Overall, the incentives of “pure” nominal systems are promoted.
Similarly, if relatively more seats are allocated via a list component, nominal and list candidates would share similar interests with parties and are expected to serve broader interests. This similarity of incentives is also expected to blur the mandate divide between the two types of legislators, promoting legislative contamination. With fewer nominal seats contested, the nominal component becomes less important for party leaders in controlling electoral outcomes. Fewer seats also suggest that each district covers a larger region, which motivates nominal candidates to focus on broader interests. In turn, candidates are not strongly encouraged to serve local interests. The incentives of “pure” PR systems may be enhanced as the proportion of seats allocated by those mechanisms increases. However, since the number of nominal incumbents is small, few nominal incumbents actually change their strategy and act like list incumbents would. The substantive importance of contamination effects is minimized. Overall, the incentives of “pure” PR are promoted. (p. 460)
Nomination Patterns Unique to Mixed-Member Systems
Nomination patterns unique to mixed-member systems may promote contamination at the electoral and legislative levels. In some contexts, a party leader may not be able to effectively control list members through nominations or list ranking, resulting in a blurred mandate divide. The norm in South Korea, for instance, is that list incumbents do not receive list component nominations for two consecutive terms; to seek re-election, candidates must win nominations on the nominal component through primary elections. To win primaries, South Korea’s list members need to win support from local voters, creating incentives that may render list members indistinguishable from nominal members in terms of legislative behavior (Jun and Hix 2010) at the legislative level. This kind of arrangement does not occur outside of mixed-member systems.
This arrangement also increases the number of parties, promoting contamination at the electoral level. Since list incumbents, including minor party candidates, cultivate local votes, their incentives appear to inflate the number of parties when the next elections are held.
The previous section argued that variation in features of mixed-member systems should influence the relative power of the incentives, with some variants of mixed-member systems more likely to yield evidence of contamination. However, variation in electoral rules does not fully explain observations from Table 22.1. Some studies find contamination to be unimportant even in the electoral arena (e.g., see Moser 2001; Moser and Scheiner 2004; Crisp et al. 2012). Other studies also find less contamination in the legislative arena, even in mixed-member systems with dual candidacy, including MMM in Japan (e.g., Maeda 2008) and MMP in Germany (Becher and Sieberer 2008; Fortin-Rittberger and Eder 2013; Sieberer 2010, 2015) and New Zealand (McLeay and Vowles 2007), although the seat distributions between the nominal and list components are quite even.
Additional context may be needed to understand these outcomes. For instance, legislative contamination requires additional assumptions about multiple principals that incumbents face (i.e., local constituents and party leaders) (Carey 2009; Maltzman 1997). The hypothesis that SSD incumbents act like list incumbents implies that nominal incumbents need to, or are forced to, prioritize party leaders over local constituents. As Herron (2002b) shows in Ukraine, nominal incumbents might want to demonstrate their loyalty to the party leadership when they are dually nominated in an unsafe rank on the list. Still, voting against local constituents’ interests could backfire in later elections. Thus, for the hypothesis to hold, one would need to assume that nominal incumbents are relatively free from local constituents’ constraints. To continue the example, it could be the case that under Ukraine’s inchoate party system, the incumbency advantage was (p. 461) so weak that candidates would not need to cater to local demands. In other words, in addition to understanding the incentives introduced by institutional rules, the development of party organizations and level of party system institutionalization could influence the effects.
The previous discussion implies that institutional and contextual factors beyond the rules of mixed-member systems can dampen or accentuate contamination effects. The extant scholarship also suggests that factors such as constitutional structures, social homogeneity, candidate selection and campaign funding, and parties’ linkage strategies could play important roles (Crisp 2007; Moser and Scheiner 2012).30 In the remainder of this section, we describe how other institutional, social, and cultural features could explain variation in mixed-member systems’ outcomes. Specifically, we address presidentialism, party organizations and resources, and cultural factors.
Some scholarship notes that presidentialism should pressure candidates and voters to coordinate in the electoral arena,31 especially where the president is powerful and the presidential and legislative electoral cycles are concurrent (Hicken and Stoll 2011; Golder 2006).32 Since there is only a single office, parties and candidates are motivated to coordinate because of the all-or-nothing nature of the elections.33 As a result, small parties tend to align with electable presidential candidates, promoting two-party systems in legislative elections. Because of the alliances that parties form, the number of candidates in the nominal component of mixed-member systems may be suppressed. This feature dampens the parties’ incentives to nominate hopeless candidates for the legislative elections and suppress contamination at the electoral level.
To be sure, presidential elections may generate two blocks of parties instead of two parties, and the number of parties may not be reduced. However, party leaders may see more opportunities to coordinate their candidates in legislative elections when their parties recognize an alliance. For instance, in semipresidential Taiwan, small parties (the Taiwan Solidarity Union and the New Party) tried to coordinate their legislative candidates with one of the two major parties (the Democratic Progressive Party and the Kuomintang), even though nominating hopeless candidates could help shore up list votes (Nemoto and Tsai 2016).
In contrast, the incentive to coordinate candidates in legislative elections should be weaker under parliamentary systems without the pressure of having presidential elections. Because achieving coordination takes time and energy, party leaders are less likely to create party blocks as they do in presidential systems. Moreover, winning as many parliamentary seats as possible improves the odds of joining the cabinet and eventually winning the post of prime minister.34 Absent concurrent presidential elections that encourage coattail effects, parties also have stronger incentives to utilize the nominal component to influence list votes by populating constituencies with their candidates. (p. 462)
While mixed-member systems may produce incentives for political actors to create parties and for those parties to nominate candidates in district races, resource limitations may undermine contamination at the electoral level. Even if actors believe that, for example, nominating noncompetitive candidates in constituency races could improve their performance on the list component, they may lack the personnel, technical, or financial resources to find an appropriate candidate and gain ballot access. High-quality nominal candidates may produce more benefits to the party in the list component (Ferrara et al. 2005),35 and conversely problematic candidates could counter those effects. In addition, parties must have adequate resources to register and effectively support candidates. Some political actors, even when they expect contamination, may choose to nominate fewer constituency candidates (Herron 2002a) or not seek ballot access for a party because of resource limitations.
The way that parties distribute resources may also affect candidate behavior and influence contamination effects at the legislative level. Depending on how parties are organized and function, campaign funding can concentrate power in the hands of party leaders.36 If nominal candidates depend on a party for campaign funding and if campaigning is capital intensive, then the party becomes a more important principal controlling its members as agents. This condition implies that nominal candidates may act more like stylized list members. However, if nominal candidates depend more on party supporters for local canvassing than on funding from a party, then the mandate divide between nominal and list candidates could become sharper.
In addition to differences in resource management, parties may also adopt different strategies in approaching voters (Kitschelt 2000) that could affect contamination. With programmatic linkage strategies, parties access voters by developing ideologically cohesive sets of policy platforms. Under this model, candidates tend to become somewhat faceless conveying the party’s policy message to voters. Meanwhile, with clientelistic linkage strategies, parties exchange electoral support for particularistic policy favors, especially in the form of pork barrel benefits. With this approach, a party becomes a franchise of locally oriented members seeking parochial benefits for their own personal supporters.
The Japanese case suggests that party leaders might have perceived, at least in the first few elections under MMM after the electoral reform, that clientelistic linkage strategies were still important in winning votes and seats. This could be partially explained by path dependency (Krauss and Pekkanen 2010; Scheiner 2008): the electoral system previously used in Japan was the SNTV, which necessarily created intraparty competition and therefore the incentive to cultivate personal votes (Carey and Shugart 1995). Thus, even though the Japanese ruling party’s post allocation strategy described earlier (Pekkanen et al. 2006)—list incumbents assuming government and party positions useful in shoring up local votes—is compatible with the party incentive to maximize seats under MMM, the legacy from the past might have further blurred the mandate divide between nominal and list incumbents.37 (p. 463)
Social and Cultural Features
Moser and Scheiner (2012) point out several noninstitutional contextual factors that could affect mixed-member systems. Certain social characteristics, such as ethnicity, language, and religion, can serve as useful informational shortcuts for voters. Thus, in socially heterogeneous environments, ambitious politicians take advantage of such easy-to-understand heuristics to appeal to social minorities, resulting in party fragmentation (Amorim Neto and Cox 1997; Stoll 2013). In line with this argument, the presence of such useful heuristics under mixed-member systems might motivate party leaders to mobilize ethnic list votes by nominating as many nominal candidates as possible.38
Inherited traditions may also affect behaviors. For example, long experience with two-party systems and FPTP, as one finds in British colonial territories, could mitigate effects that we would expect from contamination (Crisp et al. 2012). Personalistic appeals, induced by decades of SNTV in Japan, may encourage list incumbents to work for local interests (Krauss and Pekkanen 2010; Nemoto, this volume; Crisp 2007).
A related factor that could affect behavior is the age of democracies. Party systems in fledgling democracies are often not institutionalized and may be accompanied by a large number of independents. For political actors to respond to incentives, they require consistent, systematic information to coordinate (e.g., withdraw candidates from nominal districts, engage in strategic voting to avoid wasting votes, or withdraw financial support from nonviable candidates). Adequate information may not be readily available among newly democratized countries. For example, Lesotho’s mixed-member system produced an inflated number of parties, but parties and voters in Lesotho did not fully understand the mechanical effects of the mixed-member system, especially in the first two elections, and voters did not have appropriate information to determine candidate viability. Although Rich, Banerjee, and Recker (2014) state that the results from Lesotho are consistent with the contamination hypothesis, the deficiency of pre-election polling data could have caused the number of parties to increase.
The expansion of mixed-member electoral systems at the end of the twentieth and beginning of the twenty-first centuries prompted extensive scholarly attention, particularly to the consequences of its diffusion. As we noted in this chapter, the scholarly debate about the effects of combining formulas in separate overlapping components for allocating seats into a single legislative election has not been conclusive. Some studies have found evidence of contamination effects in outcomes ranging from the number of political parties and women’s representation to legislative voting behavior and committee assignments. These effects have been documented in democratic and nondemocratic regime types. At the same time, the controlled comparison approach has yielded important findings about the factors that mitigate effects of electoral rules, and some (p. 464) scholars have generated empirical evidence that challenges the expectations of the contamination effects approach. While the traditional interpretation of the controlled comparison and contamination approaches has been to treat them as diametrically opposed to one another, we argue that they are better viewed as differences in degree. Both approaches define conditions in which the components of mixed-member systems may behave more or less independently, and neither approach suggests that contamination is always—or never—present.
Advancing our understanding of the complex combination of electoral and nonelectoral factors that induce different incentives in mixed-member systems will enhance our understanding of these rules, and institutional rules more generally. Building upon scholarship from the last two decades, we envision several ways to advance the research agenda on mixed-member systems: formalizing theoretical expectations, adopting different research designs, and extending data collection to societies where limited work has been conducted.
Most of the work on mixed-member systems has not mathematically formalized theoretical expectations, relying instead on less rigorous expository logic. Developing formal models and testing hypotheses derived from these models is an important step to advance the discussion about mixed-member systems. Bawn and Thies (2003) and Hennl (2014) provide a model.
Experimental approaches could also be meaningfully incorporated into the study of mixed-member systems.39 In the laboratory and potentially in the field,40 treatments that induce participants to engage with one component or another (or to look across components) could advance our understanding of the microfoundations under which contamination matters or does not matter. Moreover, conducting experimental work in varied cultural settings could take into account contextual features that may influence behavior.
At the same time, digging deeper into specific cases would advance the study of mixed-member systems. Qualitative work that engages directly with political actors to better understand their motivations for making decisions during campaigns and in the legislature could reveal if they take into account the factors that the contamination effects scholarship implies. To assess if these effects are artifacts of the analytical approach, it would be valuable to discover if political actors are aware of the effects and adopt strategies to take advantage of them.
Extending research on mixed-member systems to incorporate additional country- and region-based cases is also crucial to understanding how context could influence their effects. While scholars have investigated mixed-member systems in a wide range of countries, some world regions have been less intensively explored. Mixed-member systems adopted in Sub-Saharan Africa (e.g., Cameroon, Lesotho, Senegal) are notable for limited scholarly attention, and research on subnational use of mixed-member rules would also extend our knowledge about their effects. In addition, single-country studies dominate the area of legislative contamination effects. To empirically analyze some of the theoretical predictions that this article proposed, scholars should construct (p. 465) comprehensive datasets covering roll-call voting, party switching, and committee assignments in a wide range of mixed-member systems.
As we emphasized in this chapter, a key to understanding the consequences of mixed-member systems may come from better articulating theoretical expectations. We advocate more exploration of ideas such as the implications of distal and proximal effects, more attention to institutional variation in mixed-member systems (e.g., dual candidacy), and additional work on ancillary factors that could enhance or attenuate contamination effects. Scholarship that extends what we know about mixed-member systems will advance research not only on electoral systems but also on the interaction among institutional rules in general.
Amorim Neto, Octavio, and Gary W. Cox. “Electoral Institutions, Cleavage Structures, and the Number of Parties.” American Journal of Political Science 41, no. 1 (1997): 149–174.Find this resource:
Barker, Fiona, and Stephen Levine. “The Individual Parliamentary Member and Institutional Change: The Changing Role of the New Zealand Member of Parliament.” Journal of Legislative Studies 5, no. 3/4 (1999): 105–130.Find this resource:
Batto, Nathan F. “Differing Mandates and Party Loyalty in Mixed-member Systems: Taiwan as a Baseline Case.” Electoral Studies 31, no. 2 (2012): 384–392.Find this resource:
Bawn, Kathleen, and Michael Thies. “A Comparative Theory of Electoral Incentives: Representing the Unorganized Under PR, Plurality and Mixed-Member Electoral Systems.” Journal of Theoretical Politics 15, no. 1 (2003): 5–32.Find this resource:
Becher, Michael, and Ulrich Sieberer. “Discipline, Electoral Rules and Defection in the Bundestag, 1983–94.” German Politics 17, no. 3 (2008): 293–304.Find this resource:
Bernauer, Julian, and Simon Munzert. “Loyal to the Game? Strategic Policy Representation in Mixed Electoral Systems.” Representation 50, no. 1 (2014): 83–97.Find this resource:
Bochsler, Daniel. “Are Mixed Electoral Systems the Best Choice for Central and Eastern Europe or the Reason for Defective Party Systems?” Politics & Policy 37, no. 4 (2009): 735–767.Find this resource:
Carey, John M.Legislative Voting and Accountability. Cambridge: Cambridge University Press, 2009.Find this resource:
(p. 468) Carey, John M., and Matthew Soberg Shugart. “Incentives to Cultivate a Personal Vote: A Rank Ordering of Electoral Formulas.” Electoral Studies 14, no. 4 (1995): 417–439.Find this resource:
Centellas, Miguel. “Mixed-Member Election and Candidate Selection in Bolivia’s 1993 and 1997 Elections.” Latin Americanist 59, no. 1 (2015): 3–22.Find this resource:
Corral, Margarita, Francisco Sanchez, and Cristina Rivas Perez. “The Impact of Mixed-Member Districts on Legislators’ Behavior: The Case of Bolivia.” Latin American Politics and Society 58, no. 1 (2016): 29–48.Find this resource:
Crisp, Brian F. “Incentives in Mixed-Member Electoral Systems: General Election Laws, Candidate Selection Procedures, and Cameral Rules.” Comparative Political Studies 40, no. 12 (2007): 1460–1485.Find this resource:
Crisp, Brian F., Joshua D. Potter, and John J. W. Lee. “Entry and Coordination in Mixed-Member Systems: A Controlled Comparison Testing the Contamination Hypothesis.” Journal of Politics 74, no. 2 (2012): 571–583.Find this resource:
Desposato, Scott, ed. Ethics and Experiments: Problems and Solutions for Social Scientists and Policy Professionals. New York: Routledge, 2016.Find this resource:
Duverger, Maurice.Political Parties, Their Organization and Activity in the Modern State. London, New York: Methuen; Wiley, 1954.Find this resource:
Ferrara, Federico. “Electoral Coordination and the Strategic Desertion of Strong Parties in Compensatory Mixed Systems with Negative Vote Transfers.” Electoral Studies 23, no. 3 (2004): 391–413.Find this resource:
Ferrara, Federico. “Two in One: Party Competition in the Italian Single Ballot Mixed System.” Electoral Studies 25, no. 2 (2006): 329–350.Find this resource:
Ferrara, Federico, and Erik S. Herron. “Going It Alone? Strategic Entry under Mixed Electoral Rules.” American Journal of Political Science 49, no. 1 (2005): 16–31.Find this resource:
Ferrara, Federico, Erik S. Herron, and Misa Nishikawa. Mixed Electoral Systems: Contamination and Its Consequences. New York: Palgrave Macmillan, 2005.Find this resource:
Fortin-Rittberger, Jessica, and Christina Eder. “Towards a Gender-Equal Bundestag? The Impact of Electoral Rules on Women’s Representation.” West European Politics 36, no. 5 (2013): 969–985.Find this resource:
Golder, Matt. “Presidential Coattails and Legislative Fragmentation.” American Journal of Political Science 50 (2006): 34–48.Find this resource:
Golosov, Grigorii V. “Electoral Systems and Party Formation in Russia.” Comparative Political Studies 36, no. 8 (2003): 912–935.Find this resource:
Golosov, Grigorii V. “Interdependence Effects in Mixed-Superposition Electoral Systems: An Empirical Test on Women’s Participation in Sub-National Elections.” Journal of Elections, Public Opinion and Parties 24, no. 4 (2014): 434–454.Find this resource:
Gschwend, Thomas, Ron Johnston, and Charles Pattie. “Split-Ticket Patterns in Mixed-Member Proportional Election Systems: Estimates and Analyses of Their Spatial Variation at the German Federal Election, 1998.” British Journal of Political Science 33, no. 1 (2003): 109–127.Find this resource:
Hainmueller, Jens, and Holger Lutz Kern. “Incumbency as a Source of Spillover Effects in Mixed Electoral Systems: Evidence from a Regression-Discontinuity Design.” 27 (2008): 213–227.Find this resource:
Heitshusen, Valerie, Garry Young, and David M. Wood. “Electoral Context and MP Constituency Focus in Australia, Canada, Ireland, New Zealand, and the United Kingdom.” American Journal of Political Science 49, no. 1 (2005): 32–45.Find this resource:
(p. 469) Hennl, Annika. “Intra-Party Dynamics in Mixed-member Electoral Systems: How Strategies of Candidate Selection Impact Parliamentary Behaviour.” Journal of Theoretical Politics 26, no. 1 (2014): 93–116.Find this resource:
Herron, Erik S. “Mixed Electoral Rules and Party Strategies: Responses to Incentives by Ukraine’s Rukh and Russia’s Yabloko.” Party Politics 8 (November 2002a): 719–733.Find this resource:
Herron, Erik S. “Electoral Influences on Legislative Behavior in Mixed-Member Systems: Evidence from Ukraine’s Verkhovna Rada.” Legislative Studies Quarterly 27 (August 2002b): 361–382.Find this resource:
Herron, Erik S., and Misa Nishikawa. “Contamination Effects and the Number of Parties in Mixed-Superposition Electoral Systems.” Electoral Studies 20, no. 1 (2001): 63–86.Find this resource:
Hicken, Allen, and Heather Stoll. “Presidents and Parties: How Presidential Elections Shape Coordination in Legislative Elections.” Comparative Political Studies 44, no. 7 (2011): 854–883.Find this resource:
Jun, Hae-Won, and Simon Hix. “Electoral Systems, Political Career Paths and Legislative Behavior: Evidence from South Korea’s Mixed-Member System.” Japanese Journal of Political Science 11, no. 2 (2010): 153–171.Find this resource:
Karp, Jeffrey A. “Candidate Effects and Spill-Over in Mixed Systems: Evidence from New Zealand.” Electoral Studies 28 (2009): 41–50.Find this resource:
Katz, Richard S.A Theory of Parties and Electoral Systems. Baltimore: Johns Hopkins University Press, 1980.Find this resource:
Kerevel, Yann. “The Legislative Consequences of Mexico’s Mixed-Member Electoral System, 2000-2009.” Electoral Studies 29, no. 4 (2010): 691–703.Find this resource:
Kitschelt, Herbert. “Linkages between Citizens and Politicians in Democratic Polities.” Comparative Political Studies 33, no. 6/7 (2000): 845–879.Find this resource:
Kostadinova, Tatiana. “Party Strategies and Voter Behavior in the East European Mixed Election Systems.” Party Politics 12, no. 1 (2006): 121–143.Find this resource:
Krauss, Ellis S., Kuniaki Nemoto, and Robert Pekkanen. “Reverse Contamination: Burning and Building Bridges in Mixed-Member Systems.” Comparative Political Studies 45, no. 6 (2012): 747–773.Find this resource:
Krauss, Ellis S., and Robert Pekkanen. The Rise and Fall of Japan’s LDP. Ithaca, NY: Cornell University Press, 2010.Find this resource:
Lago, Ignacio, and Ferran Martinez. “The Importance of Electoral Rules: Comparing the Number of Parties in Spain’s Lower and Upper Houses.” Electoral Studies 26 (2007): 381–391.Find this resource:
Lago, Ignacio, and José Ramón Montero. “Coordination between Electoral Arenas in Multilevel Countries.” European Journal of Political Research 48, no. 2 (2009): 176–203.Find this resource:
Lancaster, Thomas D., and W. David Patterson. “Comparative Pork Barrel Politics: Perceptions from the West German Bundestag.” Comparative Political Studies 22, no. 4 (1990): 458–477.Find this resource:
Lundberg, Thomas. “Putting a Human Face on Proportional Representation: Early Experiences in Scotland and Wales.” Representation 38, no. 4 (2002): 271–283.Find this resource:
Lundberg, Thomas Carl. “Post-Communism and the Abandonment of Mixed-Member Electoral Systems.” Representation 45, no. 1 (2009): 15–27.Find this resource:
Maeda, Ko. “Re-Examining the Contamination Effect of Japan’s Mixed Electoral System Using the Treatment-Effects Model.” Electoral Studies 27, no. 4 (2008): 723–731.Find this resource:
Maltzman, Forrest.Competing Principals: Committees, Parties, and the Organization of Congress. Ann Arbor: University of Michigan Press, 1997.Find this resource:
Massicotte, Louis, and Andre Blais. “Mixed Electoral Systems: A Conceptual and Empirical Survey.” Electoral Studies 18 (1999): 344–366.Find this resource:
(p. 470) Mcleay, Elizabeth, and Jack Vowles. “Redefining Constituency Representation: The Roles of New Zealand MPs under MMP.” Regional and Federal Studies 17, no. 1 (2007): 71–95.Find this resource:
Moser, Robert G.Unexpected Outcomes: Electoral Systems, Political Parties, and Representation in Russia. Pittsburgh, PA: Pittsburgh University Press, 2001.Find this resource:
Moser, Robert G., and Ethan Scheiner. “Mixed Electoral Systems and Electoral System Effects: Controlled Comparison and Cross-National Analysis.” Electoral Studies 23 (2004): 575–599.Find this resource:
Moser, Robert G., and Ethan Scheiner. Electoral Systems and Political Context: How the Effects of Rules Vary across New and Established Democracies. Cambridge: Cambridge University Press, 2012.Find this resource:
Nemoto, Kuniaki, and Matthew S. Shugart. “Localism and Coordination under Three Different Electoral Systems: The National District of the Japanese House of Councillors.” Electoral Studies 32, no. 1 (2013): 1–12.Find this resource:
Nemoto, Kuniaki, and Chia-hung Tsai. “Post Allocation, List Nominations, and Pre-Electoral Coalitions under MMM.” In Mixed-Member Electoral Systems in Constitutional Context: Taiwan, Japan, and Beyond, edited by Nathan Batto, Gary Cox, Chi Huang, and Alex Tan, 165–193. Ann Arbor: University of Michigan Press, 2016.Find this resource:
Nishikawa, Misa, and Erik S. Herron. “Mixed Electoral Rules’ Impact on Party Systems.” Electoral Studies 23, no. 4 (2004): 753–768.Find this resource:
Ohmura, Tamaki. “When Your Name Is on the List, It Is Time to Party: The Candidacy Divide in a Mixed-Member Proportional System.” Representation 50, no. 1 (2014): 69–82.Find this resource:
Olivella, Santiago, and Margit Tavits. “Legislative Effects of Electoral Mandates.” British Journal of Political Science 44, no. 2 (2014): 301–321.Find this resource:
Pekkanen, Robert, Benjamin Nyblade, and Ellis S. Krauss. “Electoral Incentives in Mixed-Member Systems: Party, Posts, and Zombie Politicians in Japan.” American Political Science Review 100, no. 2 (2006): 183–193.Find this resource:
Rae, Douglas.The Political Consequences of Electoral Laws. New Haven, CT: Yale University Press, 1967.Find this resource:
Reynolds, Andrew, and B. Reilly. The International IDEA Handbook of Electoral System Design. Stockholm, Sweden: International Institute for Democracy and Electoral Assistance, 1997.Find this resource:
Rich, Timothy S. “The Effects of Election Reform on Legislator Perceptions: The Case of Taiwan.” Japanese Journal of Political Science 13, no. 3 (2012): 317–336.Find this resource:
Rich, Timothy S. “Party Voting Cohesion in Mixed Member Legislative Systems: Evidence from Korea and Taiwan.” Legislative Studies Quarterly 39, no. 1 (2014): 113–135.Find this resource:
Rich, Timothy S. “Duverger’s Law in Mixed Legislative Systems: The Impact of National Electoral Rules on District Competition.” European Journal of Political Research 54 (2015): 182–196.Find this resource:
Rich, Timothy S., Vasabjit Banerjee, and Sterling Recker. “Identifying the Institutional Effects of Mixed Systems in New Democracies: The Case of Lesotho.” Journal of Asian and African Studies 49, no. 6 (2014): 637–653.Find this resource:
Riker, William H. “The Two-Party System and Duverger’s Law: An Essay on the History of Political Science.” American Political Science Review 76 (1982): 753–766.Find this resource:
Sartori, Giovanni.Comparative Constitutional Engineering. London: Macmillan, 1994.Find this resource:
Scheiner, Ethan. “Does Electoral System Reform Work?: Electoral System Lessons from Reforms of the 1990s.” Annual Review of Political Science 11 (2008): 161–181.Find this resource:
Shin, Ki-young. “Women’s Sustainable Representation and the Spillover Effect of Electoral Gender Quotas in South Korea.” International Political Science Review 35, no. 1 (2014): 80–92.Find this resource:
(p. 471) Shugart, Matthew S., and Rein Taagepera. Votes from Seats. Cambridge: Cambridge University Press, 2017.Find this resource:
Shugart, Matthew S., and Martin P. Wattenberg, eds. Mixed-Member Electoral Systems: The Best of Both Worlds? Oxford: Oxford University Press, 2001.Find this resource:
Sieberer, Ulrich. “Behavioral Consequences of Mixed Electoral Systems: Deviating Voting Behavior of District and List MPs in the German Bundestag.” Electoral Studies 29, no. 3 (2010): 484–496.Find this resource:
Sieberer, Ulrich. “Using MP Statements to Explain Voting Behaviour in the German Bundestag: An Individual Level Test of the Competing Principals Theory.” Party Politics 21, no. 2 (2015): 284–294.Find this resource:
Stoffel, Michael F. “MP Behavior in Mixed-Member Electoral Systems.” Electoral Studies 35 (2014): 78–87.Find this resource:
Stoll, Heather.Changing Societies, Changing Party Systems. Cambridge: Cambridge University Press, 2013.Find this resource:
Stratmann, Thomas, and Martin Baur. “Plurality Rule, Proportional Representation, and the German Bundestag: How Incentives to Pork-Barrel Differ Across Electoral Systems.” American Journal of Political Science 46, no. 3 (2002): 506–514.Find this resource:
Thames, Frank G., and Martin Edwards. “Differentiating Mixed-Member Electoral Systems: Mixed-Member Majoritarian and Mixed-Member Proportional Systems and Government Expenditures.” Comparative Political Studies 39, no. 7 (2006): 905–927.Find this resource:
Zittel, Thomas, and Thomas Gschwend. “Individualised Constituency Campaigns in Mixed-Member Electoral Systems: Candidates in the 2005 German Elections.” West European Politics 31, no. 5 (2008): 978–1003. (p. 472) Find this resource:
(1.) Names are in alphabetical order to reflect equal authorship.
(9.) Scholars use different terms, such as “tiers” and “levels,” for “components.” In this chapter we use the term “components.”
(10.) To illustrate the differences between MMP and MMM, imagine a hypothetical legislature with two hundred seats, one hundred allocated to the PR list component, and one hundred allocated to constituency races via FPTP in the nominal component. Party A receives 20 percent of the list vote and forty seats in the nominal component. In an MMP system, party A’s total seat allocation should be 20 percent of the legislature, or forty seats. Since party A receives forty seats in the nominal constituency races, it would gain no additional list seats as MMP uses the list component to determine the overall allocation of seats in the legislature. In an MMM system, by contrast, party A would receive forty nominal constituency seats and twenty additional seats in the list component, for a total of sixty seats (30 percent of the legislature) because seats in the two components are allocated independently.
(11.) In the literature on Japan’s mixed electoral system, dually nominated list winners are called “zombies” (Pekkanen et al. 2006). From a normative perspective, this term implies that one tier carries more legitimacy than another. We use “dually nominated list winners” to preserve neutrality.
(12.) For example, in Japan, the number of nominal seats exceeds list seats by 295 to 180; in South Korea the difference is 253 to 47; and in Taiwan the difference is 79 to 34.
(13.) Typically, the literature uses members’ roll-call voting patterns, party switching records, and committee assignments to explore these theoretical predictions.
(14.) Assuming that the PR list is closed and that party rules privilege leaders in list construction.
(15.) The “mandate divide” describes the implication that behaviors vary based on the incentives of the two components and is consistent with a controlled comparison argument.
(16.) A Google N-gram assessment, focusing on the frequency of terms in books, shows an increase in the use of “mixed electoral system” and “mixed-member electoral system” beginning in the 1990s. However, the number of books the figure is based on is small. See http://bit.ly/2bgFXbY.
(17.) We use the term “mixed-member system” in the chapter, but we searched by “mixed system” to capture both labels.
(19.) According to Rae’s (1967) original formulation, distal and proximal effects are primarily based on time: the former require more time to take place than proximal effects. Scheiner (2008) generalizes the idea, suggesting that the distal effects are the more indirect effects of electoral systems.
(21.) Herron (2002b) uses the case of Ukraine to show that nominal incumbents’ party loyalty is mediated by their safety on the party list. That is, members placed in unsafe positions on the list tend to toe the party line to appeal their loyalty to the leadership.
(24.) One-ballot systems (fused ballot), used in countries such as Lesotho, should also promote contamination at the legislative and electoral levels. Rich, Banerjee, and Recker (2014) state that the introduction of the one-vote system in Lesotho encouraged parties to place noncompetitive nominal candidates in districts to win seats through party list competitions. Under such circumstances, as in dual candidacy, nominal candidates are motivated to serve broader party interests. In contrast, the introduction of two separate ballots (i.e., two votes) may be more likely to generate a clearer mandate divide between nominal and list members, ceteris paribus, although we cannot eliminate the possibility that two separate ballots actually promote contamination, since list candidates from minor parties may encourage split-ticket voting.
(25.) It is true that in New Zealand and elsewhere, nominal MPs provide constituency services to build up party supporters in local areas (Barker and Levine 1999). That said, however, our claim is that it should be relative in theory: such local orientations should be less pronounced under MMP than MMM. This is an empirical question that requires additional work, especially outside of New Zealand and Germany to evaluate how well the findings travel.
(27.) This assumes that their party satisfies other requirements, such as the minimum threshold. Under MMP in Germany, parties are required to receive 5 percent of the list votes or win three nominal seats.
(28.) Stoffel also suggests that those nominal candidates who are without such a secure seat are more motivated to work for local interests, potentially contributing to a clearer mandate divide. Although Stoffel’s (2014) focus is more on the effects of dual candidacy than on MMP, they seem to speak equally to the effects of MMP.
(29.) This observation especially applies to MMM systems.
(30.) As we note later, to explore these factors, it is imperative to cross-sectionally compare different mixed-member systems under different contexts and institutions.
(31.) At the same time, fragmentation is observed in many presidential systems. The evidence about the effects of presidentialism on coordination is not conclusive.
(33.) The presidential office is also the most important political prize in presidential systems. For example, US presidents are considered to be relatively weaker than the counterparts in Latin America and Africa, but they possess the power to appoint cabinet members and other positions, exercise veto power, negotiate treaties, and sign executive orders.
(34.) Governing parties usually divide cabinet positions in proportion to their legislative seat shares.
(38.) See the chapter by Robert G. Moser, Ethan Scheiner, and Heather Stoll in this volume for more on social heterogeneity.