Companion piece: Michigan OLPR vs. MMP values quiz — a 10-question self-assessment that translates the tradeoffs reviewed below into a score and a concrete design recommendation.

---

Michigan is considering proportional representation (PR) for its state legislature. Two families of PR systems dominate the real conversation. Open-list proportional representation (OLPR) uses multi-member districts, a single party vote, and preference votes that let voters rank candidates within their chosen party. Mixed-member proportional (MMP) uses two tiers — a single-member district (SMD) tier plus a list tier — and two votes per ballot. Each has passionate advocates. Each has been tried in dozens of countries with genuinely varying results. Each solves some problems better than the other and introduces its own pathologies.

This piece reviews seven papers from the comparative electoral systems literature, chosen because they speak to the specific questions Michigan reformers will have to answer. Who gets elected under each system, and why? Who gets represented — majorities, concentrated minorities, dispersed minorities? What fails, and how badly? The goal is not to declare a winner but to clarify what is actually known, what is contested, and what Michigan would have to decide on its own.

The short version: the evidence favors neither system unambiguously. Both can work well. MMP’s reputation rests heavily on Germany and New Zealand, and its failures outside those cases are serious. OLPR is more robust to manipulation but places larger demands on voters and hands party gatekeeping power from party elites to the electorate — a feature or a bug depending on your theory of parties. The design parameters that most affect outcomes (district magnitude, threshold, number of preference votes, list flexibility) matter as much as the system-level choice. A well-designed OLPR or MMP will outperform a badly-designed version of the other.

---

I. How the two systems work

Under OLPR, voters cast one ballot. They pick a party and — in most versions — mark one or more preferred candidates within that party’s list. Seats are allocated across parties in proportion to party votes. Within each party, the candidates who accumulated the most preference votes win. Countries currently using OLPR include Brazil, Finland, Poland, Chile, Denmark (flexible-list variant), and Sweden (flexible-list variant).

Under MMP, voters cast two ballots. The first elects a single-member district representative by plurality. The second is a party-list vote that determines each party’s total seat share in the legislature. After the SMD winners are seated, list seats are filled from closed (usually) party-ranked lists to bring each party’s total up to its list-vote entitlement. Countries currently using MMP include Germany, New Zealand, Scotland (regional MSP tier), Wales, Bolivia, and Lesotho.

Both deliver roughly proportional outcomes. The differences are in the mechanics — who controls nominations, how voters express candidate preferences, what fails when the system is gamed, and how minorities and factions fare in each. These mechanics are what the literature actually studies, and what Michigan’s design choice hinges on.

---

II. Seven papers

1. Linhart, Raabe & Statsch (2019): Is MMP actually “the best of both worlds”?

Lijphart (1984) introduced the question of whether mixed electoral systems could deliver the advantages of both plurality and PR simultaneously. Since then, MMP has acquired a reputation as the strongest answer to that question — a system that produces proportional outcomes while concentrating the party system enough to support stable governments with clear voter-government links.

Linhart, Raabe, and Statsch stress-test this reputation empirically. They analyze all 187 elections ever held under MMP rules: in Albania, Bolivia, Germany, Lesotho, New Zealand, Venezuela, Scotland, Wales, and sixteen German Länder.

Their finding complicates the story substantially. Only 44 percent of non-Länder MMP elections land in the “best of both worlds” quadrant (above-average proportionality and above-average party-system concentration).¹ Three cases — Albania 2005, Scotland 2003, Venezuela 1998 — land in the “worst of both worlds” quadrant, with elections producing both high disproportionality (Gallagher LSI values around 15-30, comparable to British plurality) and fragmented party systems. Germany and New Zealand account for the bulk of the success cases.

The dominant failure mode outside Germany and New Zealand is strategic manipulation. Parties in Albania, Lesotho, and Venezuela gamed the compensation mechanism by creating decoy parties: the main party competed only in SMDs, and a dummy twin party competed only for list seats. Because the two formally distinct parties evaded MMP’s cross-tier proportionality correction, the SMD wins stacked without being offset by list seats for other parties. Bochsler (2012) documents this pattern in detail.²

Scotland and Wales fail in a different way. Their MMP systems use higher SMD-to-list ratios and smaller PR regions, producing above-average concentration but below-average proportionality. The design is too majoritarian-leaning to deliver the proportional half of the “both worlds” promise.

Linhart and colleagues conclude that “system designers should be careful with rather complicated MMP systems in non-stable democracies with rather volatile and unstable party systems” (p. 36). They further observe that “the principles of concentration and proportionality can also be balanced out with help of legal thresholds or moderate district magnitudes in PR systems (Carey and Hix 2011) which are much easier to understand for voters.” This is a thumb on the scale against MMP as the default choice, particularly for jurisdictions where simpler alternatives are available.

For Michigan: The state is an established democracy with stable institutions, so the decoy-list manipulation risk is lower than in Albania or Lesotho. But “stable” is not “immune.” Germany itself responded to surplus-seat problems with major reforms in 2013 (leveling seats) and 2023 (formal cap on Bundestag size). Any Michigan MMP implementation should include leveling seats from day one. The simpler alternative — moderate-magnitude OLPR with a 5 percent threshold — may deliver similar outcomes with materially less design complexity and less manipulation exposure.

2. Blumenau, Eggers, Hangartner & Hix (2017): What OLPR does to party brands

Which parties benefit when a PR system switches from closed lists to open lists? Blumenau, Eggers, Hangartner, and Hix run a 9,096-respondent survey experiment on British voters, using hypothetical European Parliament elections. The effect they identify generalizes well beyond that specific case.

When respondents were shown a closed-list ballot, UKIP (the unified Eurosceptic party) received 25 percent of the vote. When shown an open-list ballot with candidate endorsement information, UKIP dropped to 19 percent — a seven-point loss. The Conservatives gained six points, from 22 to 28 percent. Among respondents with strongly Eurosceptic views, the UKIP loss was 17 points.

The mechanism: voters with strong views on a salient cross-cutting issue abandoned the unified niche party in favor of heterodox candidates of mainstream parties. A Eurosceptic voter who would have backed UKIP under closed lists could instead back a Eurosceptic Conservative candidate under open lists, getting the policy signal she wanted without the cost of voting outside a governing coalition.

The generalizable claim: OLPR disadvantages parties whose brand rests on unity around a single salient issue, and advantages parties whose candidates are internally diverse on that issue. Niche parties lose. Big-tent parties gain — but at the cost of their internal coherence, because the mechanism only works when mainstream parties actually field heterodox candidates.

For Michigan: Two implications. First, Michigan’s existing small parties (Libertarian, Working Class, Greens, US Taxpayers) that depend on single-issue branding would likely shrink under OLPR, because their voters could back heterodox candidates in the major parties. Second — and more consequentially — Michigan’s two major parties themselves would likely fragment. Each is already internally divided: progressives vs. moderates vs. labor-aligned Democrats; MAGA populists vs. traditional Republicans vs. libertarian-leaning Republicans. OLPR gives voters tools to reward specific factions directly. Over time, this tends to crystallize factions into formal parties, as happened in Italy after 1993 and to a lesser extent in New Zealand after 1996.

The realistic Michigan prediction under OLPR with a 5 percent threshold and district magnitude 5-7 is five to six parties, but not the five to six one might naively expect. Expect progressive Democrats, moderate/establishment Democrats, possibly a labor Democratic party, traditional Republicans, MAGA Republicans, and possibly a libertarian-leaning party. Not the current Libertarians, Greens, Working Class Party, and US Taxpayers as separate winners — unless they absorb defecting major-party voters.

3. Bergman, Shugart & Watt (2013): Intraparty competition under OLPR

What happens inside parties under OLPR? Bergman, Shugart, and Watt analyze 2,033 party-district observations across nine countries using either OLPR or single-non-transferable-vote (SNTV) systems, deriving a logical model of expected intraparty vote distributions and validating it empirically.

Their core finding: OLPR combines individual-candidate incentives (preference votes) with collective-party incentives (vote pooling across the list) in a way that produces intense but functional intraparty competition. The critical difference from SNTV is that OLPR pools votes across candidates within a party. A vote for candidate X helps X personally and helps X’s list collectively. SNTV doesn’t pool, so a vote for X only helps X — which means popular candidates can cannibalize seats that more evenly-distributed votes would have won.

The empirical consequence: OLPR parties typically have 51 to 58 percent of their votes cast for winning candidates. The rest go to losing co-partisans. In SNTV systems, this rate is around 90 percent, because parties actively manage their nominations to prevent vote-splitting and equalize candidate vote shares.

This sounds like an OLPR failure, but it isn’t. It means OLPR embraces intraparty competition as a feature: parties nominate full slates, candidates compete for voter preference, and the intraparty wasted votes are the cost of giving voters real choice among co-partisans. Dominance is rare: only 3.9 percent of OLPR parties had a first candidate with more than two-thirds of the party’s votes.

Bergman and colleagues derive a formal model of expected vote distributions and conclude with an explicit design recommendation: district magnitudes of four to eight seats represent the “sweet spot” where intraparty competition is live but not chaotic. They note this range “would coincide roughly, but conveniently” with Carey and Hix’s (2011) sweet spot on the interparty dimension. Below four, the system loses most of its proportional benefits. Above eight, intraparty competition becomes fragmenting.

For Michigan: Aim for districts of five to seven seats. The 110-seat House divided into districts of six would produce roughly eighteen districts. This is the range where both the proportionality and intraparty-choice benefits of OLPR are realized without the fragmentation costs of larger magnitudes.

4. Buisseret & Prato (2022): When list flexibility backfires

Peter Buisseret and Carlo Prato develop a formal model that challenges a central assumption in the reform literature: that more list openness always improves voter-legislator accountability. Their finding complicates a simple “open is better” story.

They model a continuous “flexibility” parameter between fully closed and fully open lists. The model’s key result: more flexibility can worsen local representation above a critical threshold. The mechanism is what they call “pandering-via-obstruction.” Under high flexibility, a legislator who wants to build a personal brand may oppose policies that actually benefit her own voters, because opposition creates a visible signal of alignment with a political principle, while supporting bipartisan deal-making looks like co-option by party leadership.

They state this bluntly: “sometimes legislators obstruct policies that benefit their local voters” (p. 157).

The threshold at which flexibility starts backfiring depends on the strength of partisan attachment in the electorate. Where partisanship is strong (established party systems), flexibility helps voters discipline their representatives. Where partisanship is weak or clientelistic, flexibility becomes a tool for pandering. This reinforces Linhart and colleagues’ context-sensitivity point.

Buisseret and Prato also contradict a second piece of reform orthodoxy, from Carey and Shugart (1995). Carey and Shugart predicted that under open lists, district magnitude amplifies personal-vote incentives, so larger districts should produce more particularistic behavior. Buisseret and Prato’s model predicts the opposite: district magnitude reduces personal-vote incentives across all list types. Their simulations show the probability of dissent falling from ~0.50 to ~0.38 as magnitude rises from 5 to 40, across closed, flexible, and open lists alike.

For Michigan: This reinforces the design parameter of moderate district magnitude (five to seven seats — not larger). It also suggests a flexible-list variant (Swedish or Bavarian style) may dominate pure OLPR on local-representation grounds, because it gives voters some preference power without going all the way to the pandering-prone extreme. A Swedish-style Michigan OLPR — parties rank their candidates, voters can override the ranking with sufficient preference votes at an 8-10 percent threshold — is a genuinely plausible design.

5. Saalfeld (2005): Germany’s MMP as the archetype

Thomas Saalfeld’s chapter in the Gallagher and Mitchell Politics of Electoral Systems volume is the canonical description of how MMP actually works in its most successful application. Two features stand out for reform analysis.

First, the mandate divide problem — the fear that SMD-elected and list-elected legislators form two classes with different loyalties and different constituent bases — has been substantially mitigated in Germany by dual candidacies. About 50 percent of candidates run on both tiers in recent elections, roughly 60 percent for the major parties; 20 percent are SMD-only; 33 percent are list-only. This blurs the two-class distinction in practice. Stratmann and Baur (2002) found that SMD-elected MPs gravitate to committees serving geographic constituencies while list MPs serve wider party constituencies, but no significant difference in voting cohesion emerges between the two types. Parliamentary party discipline dominates tier of origin.

Second, ticket-splitting has grown substantially, from 6.4 percent of voters in 1957 to over 20 percent by 1998. Strategic ticket-splitting specifically — where voters split their first (SMD) and second (list) votes to signal a coalition preference — reached about 7 percent in 1994 and 1998. This is a genuine MMP-specific feature: voters can simultaneously support a local Social Democratic representative and cast a list vote for the Greens to signal a preferred SPD-Green coalition. Bawn (1999) summarized the logic: “The partisan composition of the Bundestag may be the same as it would under a simpler form of PR, but the electoral incentives of the individuals who occupy those seats are not.”

But Saalfeld also documents persistent problems. Surplus seats (Überhangmandate) repeatedly produced disproportional outcomes: 16 in 1994, 13 in 1998, 5 in 2002. In the 1994 election, Helmut Kohl’s CDU-led coalition had a majority of 10 with surplus seats — and just 2 without them. The Constitutional Court split 4-4 on their constitutionality in 1997, the procedure surviving on the tie. Germany ultimately addressed the problem with leveling seats starting in 2013; before that, surplus seats were a persistent legitimacy problem.

The most revealing fact in Saalfeld’s chapter concerns a failed reform attempt. An independent 1976 expert commission on constitutional reform recommended replacing closed lists with open lists. The major parties refused. Saalfeld’s explanation: “such a reform would limit the power of mid-level party elites at the regional and Land levels to control the composition of the Land lists.” Closed lists preserve party organizational power over who actually sits in the legislature. Whether this is desirable depends entirely on what one thinks parties should be.

For Michigan: MMP preserves party organizational power over candidate selection. If the view is that well-organized parties should have meaningful control over legislative personnel quality and demographic composition, MMP delivers that capacity. If the view is that voters should have that power directly, OLPR delivers it. This is not a neutral design choice. It is a choice about the legitimate role of political parties in a democracy, and it should be debated on those terms.

6. Buisseret, Folke, Prato & Rickne (2022): What parties actually do with list control

Assuming MMP gives parties list-ranking power, do they use it well? Buisseret, Folke, Prato, and Rickne address this with extraordinarily detailed Swedish municipal data: 217,734 candidate observations across 290 municipalities over seven election cycles, 1991-2014.

Their finding cuts against conventional principal-agent skepticism about parties. Parties top-load their lists with high-quality candidates, measured four ways: military leadership scores (psychologist-rated), cognitive test scores, earnings adjusted for education and sector, and educational attainment. The monotone pattern holds across all four valence measures. Better candidates appear higher on the list; worse candidates appear lower.

The counter-intuitive twist: advantaged parties — those most likely to win executive control — do this most strongly, not least. Conventional theory predicts that electorally constrained parties, facing the tightest voter discipline, should be the most disciplined about candidate quality. The Swedish data shows the opposite. Quality sorting rises with party electoral strength.

The mechanism Buisseret and colleagues identify is that high-quality candidates are complements to formal legislative control. When a party is likely to govern, the quality of its legislators matters more, because they actually shape policy. So parties with realistic expectations of controlling the executive invest most heavily in legislator quality. Parties that will spend the coming term in opposition invest less.

For Michigan: If Michigan Democrats and Michigan Republicans are reasonably organized and electorally consequential — which they are — MMP’s list mechanism will likely produce high-quality legislators at the top of party lists. This is a genuine strength of MMP, and it’s absent in OLPR, which hands the quality-sorting function to voters. Dal Bó, Finan, Folke, Persson, and Rickne (2017) find that Swedish voters do reward quality via preference votes in flexible-list elections, but the effect is weaker than the party-sorting effect. Voters typically have less information than party gatekeepers about candidate characteristics.

This cuts against the populist reading of OLPR as straightforwardly more democratic. Voters may genuinely produce lower-quality legislators on average than well-organized parties would — though they produce a more representative legislature across other dimensions.

7. Negri (2018): Why preference-vote count matters for minority representation

Margherita Negri’s formal model addresses a question most reform conversations miss entirely: how does OLPR affect minority representation specifically, holding everything else constant?

She models a society divided into a majority group and a minority group. Voters cast a party vote plus some number of preference votes for candidates within their chosen party’s list. The core parameter is π, the number of preference votes each voter can cast. Seats are assigned proportionally to parties; within parties, the candidates with the most preference votes win.

Her finding is sharp and consequential. If π = 1 (one preference vote per ballot), minorities are proportionally represented. As π grows, a “multiplier effect” kicks in: the majority can split its votes across parties and concentrate preference votes on majority candidates within both parties’ lists. With enough preference votes, the majority can shut minorities out entirely, even when the minority retains proportional vote share.

Her worked example: in a district with 1 minority voter and 4 majority voters, at π=1 the minority gets proportional representation (one seat). At π=3, the majority can cast 12 preference votes total (4 voters × 3 preferences) against the minority’s 3, concentrated across two parties — and the minority elects zero candidates despite proportional voting weight.

Countries using OLPR handle this parameter very differently. Sweden, Finland, Denmark, and Brazil use π=1. Greece varies from 1 to 5 depending on district. Belgium allows preferences for every candidate on the list. Negri’s theorem predicts — and cross-country empirics support — that minority representation declines as π rises.

For Michigan: This is not a nice-to-have design parameter. It is a first-order constraint on whether OLPR can deliver proportional minority representation at all.

The interaction with Michigan demographics is specific:

• Black voters (~14 percent statewide, concentrated in Detroit, Flint, Pontiac, Saginaw, Benton Harbor, Ypsilanti): Within a multi-member Detroit district, Black voters would be the majority and the relevant minorities would be Arab American, Latino, and others within Detroit. Negri’s theorem applies within-district: π=1 protects Arab American and Latino representation inside Detroit’s multi-member district.
• Arab American voters (~3 percent statewide, heavily concentrated in Dearborn and Hamtramck): In a Wayne County or Dearborn-centered district with magnitude 5-7, Arab American voters would be 15-25 percent — enough to elect 1-2 representatives with coordinated π=1 voting.
• Latino voters (~5 percent statewide, dispersed across Grand Rapids, Detroit, Pontiac, and agricultural regions of the western Lower Peninsula): Geographic dispersal hurts under any single-district scheme. OLPR cannot solve this without favorable district-drawing, or without a compensation mechanism that OLPR structurally lacks.
• Native American voters (~1 percent statewide, concentrated in the Upper Peninsula and northern Lower Peninsula): Sub-threshold almost anywhere in an 18-22 district system. Would require special provision (as in New Zealand’s reserved Māori seats) to achieve reliable representation.

The structural contrast with MMP is important. MMP’s list tier is genuinely better at representing dispersed minorities — Latino, Asian American, and any other group that doesn’t achieve concentrated majorities in specific districts — because statewide or regional list ranking can elevate dispersed-minority candidates regardless of where they live. Germany’s Turkish German community has been underrepresented relative to population share precisely because closed-list ranking is party-controlled and parties under-ranked Turkish candidates (Wüst 2014); but New Zealand’s Māori representation has been enhanced by the MMP list tier plus dedicated Māori electorates. The mechanism MMP gives you is powerful but depends on party behavior.

OLPR, by contrast, is better at concentrated minorities — Black voters in Detroit, Arab Americans in Dearborn, Latino voters where they concentrate — because preference-vote coordination within a multi-member district translates concentrated minority support directly into seats. It does this without requiring party gatekeepers to rank minority candidates favorably.

Neither system is unambiguously better on minority representation. They privilege different types of minorities, and the choice between them is partly a choice about which pattern of minority concentration Michigan wants to prioritize.

---

III. What the evidence supports, briefly

Six conclusions the literature supports with reasonable confidence:

1. MMP’s “best of both worlds” reputation rests heavily on Germany and New Zealand. Outside those cases, MMP performs like average PR at best and worst-of-both-worlds at worst, usually because of strategic manipulation or poor SMD-to-list ratio design (Linhart, Raabe & Statsch 2019).

2. OLPR with moderate district magnitude (4-8) and a 5 percent threshold approximates the Carey-Hix electoral sweet spot. Simpler to implement than MMP, more transparent to voters, fewer strategic vulnerabilities (Bergman, Shugart & Watt 2013; Carey & Hix 2011).

3. OLPR disadvantages single-issue niche parties and advantages factionalized mainstream parties. Michigan’s two major parties would likely fragment into 5-6 parties under OLPR; current small parties would likely not survive a 5 percent threshold (Blumenau et al. 2017).

4. List flexibility is not monotonically good. Above a threshold that depends on partisan attachment, more flexibility can worsen local representation via pandering-via-obstruction (Buisseret & Prato 2022).

5. Party organizational capacity to rank candidates by quality is real and has democratic value. Electorally advantaged parties top-load lists with higher-quality candidates; OLPR loses this mechanism entirely (Buisseret, Folke, Prato & Rickne 2022).

6. OLPR with multiple preference votes structurally disadvantages minority representation. Single preference vote (π=1) is a first-order design constraint, not a detail (Negri 2018).

Three conclusions the literature supports more weakly:

7. MMP’s list tier is better at representing dispersed minorities; OLPR with concentrated districts is better at representing concentrated minorities. Both patterns are observable in practice but depend on implementation choices.

8. Ticket-splitting under MMP enables genuine coalition signaling at the voter level, something OLPR structurally cannot do.

9. Strategic manipulation (decoy lists under MMP, intraparty vote-splitting under OLPR) are both real but decoy-list manipulation has been empirically more damaging, particularly in less-institutionalized party systems.

---

IV. Design parameters Michigan would have to decide

Assuming Michigan adopts OLPR:

• District magnitude. 5-7 seats (18-22 districts across the 110-seat House). Below 4, too majoritarian; above 8, too fragmenting.
• Preference vote count (π). One, to protect minority representation (Negri).
• Threshold. 5 percent statewide for party-level representation. Lower thresholds enable fragmentation; higher thresholds defeat the point of PR.
• List flexibility. Pure open list (voters fully determine ranking) or Swedish-style flexible list (party-ranked with voter override at ~8 percent preference-vote threshold). The flexible-list version preserves some party quality-sorting while maintaining voter discipline.
• Minority protection. Statutory requirement that MICRC consider minority community preservation in drawing multi-member districts. This is likely required to satisfy Voting Rights Act obligations under any plausible federal ruling, though no federal case law yet exists on preference voting in multi-member districts.

Assuming Michigan adopts MMP:

• SMD-to-list ratio. Roughly 60:50 (66 SMDs + 44 list) or 50:60 (55 + 55). Germany’s 50:50 is the archetype; Scotland’s 57:43 is the more majoritarian-leaning alternative.
• List regions. 4-6 regions (e.g., Upper Peninsula, Detroit Metro, West Michigan, Thumb/Central) rather than a single statewide list. Larger regions enable better proportionality; smaller regions preserve geographic identity.
• Leveling seats. Mandatory. Germany’s 2013 reform is the template. Without leveling, surplus seats repeatedly distort proportionality (Saalfeld 2005).
• Threshold. 5 percent statewide, with a bypass for parties winning a minimum number of SMDs (Germany allows 3).
• Dual candidacies. Permitted, to mitigate mandate-divide concerns.
• Diverse list ranking. Voluntary via party rules, statutory via a diversity requirement, or monitored via transparency requirement. This determines whether MMP’s list tier actually delivers the dispersed-minority-representation advantage it is theoretically capable of.

---

V. What the evidence cannot tell Michigan

Four questions on which empirical guidance is thin or absent:

1. How VRA Section 2 applies under multi-member preference voting. No federal case law exists on majority-minority protections in multi-member OLPR districts. Municipal STV cases in the United States have navigated this ad hoc. Michigan would likely need statutory state-level protections above the federal floor.

2. Whether Michigan’s minor parties would survive OLPR with a 5 percent threshold. No close comparable case exists. Canada’s (former) minor parties faced first-past-the-post, not proportional thresholds. The Libertarian, Working Class, Green, and US Taxpayers parties’ fate under OLPR is genuinely uncertain.

3. How Michigan partisan identity resembles German, Swedish, or Finnish partisan identity. The Buisseret & Prato threshold for when list flexibility starts backfiring depends on this comparison, and the comparison is not straightforwardly measurable.

4. What happens to the Michigan Legislative Black Caucus under either system. Both paths are mechanistically viable — Detroit-area multi-member districts under OLPR, preserved majority-minority SMDs under MMP — but the paths are different and produce different types of caucus. OLPR would enable more intraparty diversity within Black representation (progressive, moderate, church-based, youth-aligned caucus members elected simultaneously); MMP would preserve the current single-member-district accountability model.

---

VI. The companion quiz

The design tradeoffs reviewed above translate into ten values questions, which are what the companion quiz assesses. If the evidence above is the “what,” the quiz is the “what do you actually prioritize among these.” Two sophisticated readers can absorb the same evidence and reach opposite conclusions because they weight the tradeoffs differently. That is legitimate. The quiz is designed to make those weights explicit.

---

Works cited

Bergman, Matthew E., Matthew S. Shugart, and Kevin A. Watt. 2013. “Patterns of Intraparty Competition in Open-List & SNTV Systems.” Electoral Studies 32 (2): 321-333. https://doi.org/10.1016/j.electstud.2013.02.002

Blumenau, Jack, Andrew C. Eggers, Dominik Hangartner, and Simon Hix. 2017. “Open/Closed List and Party Choice: Experimental Evidence from the UK.” British Journal of Political Science 47 (4): 809-827. https://doi.org/10.1017/S0007123415000629

Bochsler, Daniel. 2012. “A Quasi-Proportional Electoral System ‘Only for Honest Men’? The Hidden Potential for Manipulating Mixed Compensatory Electoral Systems.” International Political Science Review 33 (4): 401-420. https://doi.org/10.1177/0192512111418445

Buisseret, Peter, Olle Folke, Carlo Prato, and Johanna Rickne. 2022. “Party Nomination Strategies in List Proportional Representation Systems.” American Journal of Political Science 66 (3): 714-729. https://doi.org/10.1111/ajps.12691

Buisseret, Peter, and Carlo Prato. 2022. “Competing Principals? Legislative Representation in List Proportional Representation Systems.” American Journal of Political Science 66 (1): 156-170. https://doi.org/10.1111/ajps.12644

Carey, John M., and Simon Hix. 2011. “The Electoral Sweet Spot: Low-Magnitude Proportional Electoral Systems.” American Journal of Political Science 55 (2): 383-397. https://doi.org/10.1111/j.1540-5907.2010.00495.x

Carey, John M., and Matthew Soberg Shugart. 1995. “Incentives to Cultivate a Personal Vote: A Rank Ordering of Electoral Formulas.” Electoral Studies 14 (4): 417-439.

Dal Bó, Ernesto, Frederico Finan, Olle Folke, Torsten Persson, and Johanna Rickne. 2017. “Who Becomes a Politician?” Quarterly Journal of Economics 132 (4): 1877-1914. https://doi.org/10.1093/qje/qjx016

Linhart, Eric, Johannes Raabe, and Patrick Statsch. 2019. “Mixed-Member Proportional Electoral Systems — The Best of Both Worlds?” Journal of Elections, Public Opinion and Parties 29 (1): 21-40. https://doi.org/10.1080/17457289.2018.1443464

Negri, Margherita. 2018. “Preferential Votes and Minority Representation in Open List Proportional Representation Systems.” Social Choice and Welfare 50 (2): 281-303. https://doi.org/10.1007/s00355-017-1084-2

Saalfeld, Thomas. 2005. “Germany: Stability and Strategy in a Mixed-Member Proportional System.” In The Politics of Electoral Systems, edited by Michael Gallagher and Paul Mitchell, 209-229. Oxford: Oxford University Press.

Shugart, Matthew Soberg, and Martin P. Wattenberg, eds. 2001. Mixed-Member Electoral Systems: The Best of Both Worlds? Oxford: Oxford University Press.

Stratmann, Thomas, and Martin Baur. 2002. “Plurality Rule, Proportional Representation, and the German Bundestag: How Incentives to Pork-Barrel Differ Across Electoral Systems.” American Journal of Political Science 46 (3): 506-514.