Reforming Big-City Elections

There are pressures to find an acceptable set of electoral system reforms for big cities. These include poor accountability due to lack of party competition,[1] demand for change manifest in the diffusion of forms of ranked-choice voting, and policy challenges that divide one or both major-party coalitions internally. An acceptable package would aspire to channel that factionalism (rather than let it run wild), facilitate the coordination needed to make difficult policy choices,[2] and not compromise representation along the way.[3] Earlier reform approaches have not been able to do these things simultaneously.[4]

We propose a new-to-the-United States approach centered on party-list proportional representation (PR), one form of which could preserve the single-seat district. The party list presumes that candidates are grouped into preelection slates and that a vote for some candidate benefits their slate as a whole. PR presumes multi-seat districts and allocates seats to parties in proportion to their vote shares. While many associate PR with multiparty politics, its main function is to permit parties (preelection slates) to aggregate their votes over larger areas than is possible under single-seat or otherwise small districts. Those groups could be minor parties or, as is more likely, identifiable factions of the majors. Together, these features could reconcile coordination with representation—two goals long thought to be at odds in American political science scholarship.[5]

We stress that our approach is not inherently liberal or conservative. Many issues facing cities today do not fit into these ideological boxes: housing, transit, policing, climate change, and so on.[6]

Our core points:

1. Proportional representation aggregates votes over large areas, alleviating the effect on a group of being “packed” or “cracked” under existing districts. Party-list variants are better at this because they do not invite voters to divide their support among competing slates.
2. It is common to think that PR would make minor parties proliferate. This misunderstands its mechanics and adoption circumstances, as well as the factional nature of a two-party system.
3. While party-list systems are widely thought to be at odds with factionalism, they might help bind the coalitions that ongoing reform activity is likely to create.
4. Several models are available to policymakers interested in the list approach.
5. Of particular interest is how one of these models (mixed-member PR, which includes single-seat districts) might interact with the traditional role of neighborhoods in city politics. This may have implications for policy in housing, development, and other areas.
6. New York City could adopt PR by expanding its 51-seat city council or by merging existing districts. The latter would permit retaining a 51-seat council. The former has other advantages.

Imagine the following scenario: a group of Democratic and Republican candidates gets together, possibly with others, with a plan to reduce crime, increase housing supply, and prepare for rising sea levels. If the plan is popular enough, the group can expect a citywide majority of votes at the next election. Yet many things could impede that, one of which is the current electoral system.

The first problem is geography. Either or both groups may be too residentially concentrated for the coalition to win a council majority. In other words, the group is “packed” from the perspective of the current electoral rules. Or it may be too dispersed to win as many districts as its numbers otherwise suggest (i.e., the group is “cracked”). Further, existing indicators of group strength might not reflect its potential, since voters and others may see election results as foregone conclusions. For example, they might perceive the dominant party’s primary to be the only game in town.[7]

The second problem is that many forces could split the coalition. Two potential fissures are: politicians from both left and right, and the potential winners and losers of a proposed public policy. No electoral system can prevent a split in coalition leadership.[8] However, some designs are more sensitive than others to small defections.[9]

One solution to the geography problem is a system of proportional representation (PR). The solution to the fissure problem is an electoral system that rewards coalition maintenance.

Here it is useful to introduce a distinction between list forms of PR and non-list approximations. The latter category includes systems often called “semi-proportional”: limited voting, cumulative voting, and the single transferable vote (STV, recently rebranded as proportional ranked choice).[10] Limited voting permits fewer ballot markings than there are seats in a district; winners are the candidates with the most votes. Cumulative voting permits multiple votes for a single candidate, and winners are again those with the most votes. STV asks voters to rank candidates, and each winner must achieve a quota (or be elected by process of elimination). The common features of these non-list systems’ rules is that they do not presume preelection grouping of candidates into slates. All list systems do. More technically, in a list system, the vote for a candidate (if this is an option) also is a vote for his/her party.[11] We return to some options in this paper’s second half. For now, here is an example of how a non-list rule might become problematic.

Imagine that our fictive city is divided into several three-seat districts. Within each district, roughly one-third of the votes are needed to win each seat. This design is similar to the STV system used in Portland, Oregon, in 2024. It also bears resemblance to one used in New York City between 1937 and 1947.

Now imagine that the bipartisan coalition has formed a party called the Better City Association (BCA). In a simple list system, a vote for one BCA candidate also is a vote for the party (the technical name for this property of list PR is “vote pooling”).[12] So if BCA candidates get two-thirds of votes collectively, their party gets two seats. Then the two highest-polling BCA candidates get those seats. STV produces this result only if a sufficient number of BCA voters rank BCA candidates above all others. With STV, there is no formal pooling mechanism. With other non-list systems (like limited and cumulative voting), there is no possibility of pooling. Everything depends on getting voters to individually mark multiple candidates from the same party. We return to the running BCA example later.

Two myths about PR are in common currency. One is that adopting it would unleash multiple parties, which many proponents want and their detractors fear. The second myth is that semi-proportional and proportional systems are functionally equivalent to PR. It is useful to consult the political science literature on how electoral systems shape the number of parties, how this number shapes the electoral system, and how these dynamics have been different in the United States. We further argue that the U.S. system of big-tent parties should be thought of as two collections of factions, not as two unitary actors. In turn, this helps explain both past and current attraction to non-list systems: people are allergic to the idea that someone else might get to decide who is on a party list. Put differently, electoral system reform in U.S. state and local government has tended to be about breaking out of a big-tent party, whereas electoral reform in other countries often has been about top-down management of already-existing multiparty competition.[13] If one appreciates these points, one is less concerned about party system fragmentation. One also begins to see the value of a party-list approach.

One reason that PR does not necessarily unleash multiparty politics where none had existed is that the details matter. More important than the use of PR is the overall size of an assembly. Then, with PR in place, the number of seats per district becomes important. These two factors together determine how many preelection coalitions can win seats (typically measured as parties).[14] Recent research has extended these conclusions to local elections in a sample of cities worldwide, including several from the United States. Notably, the only U.S. city in the data with anything resembling multiparty politics is Chicago. Here, a large assembly lets two factions of the Democratic Party sit in council.[15]

Second, the U.S. party system is essentially two fluid coalitions of factions ultimately organized to win the presidency and navigate the national separation of powers. By “faction,” we mean an ideologically kindred group of people (and interest groups) with identifiable leadership and duration beyond one election.[16] Here, the point of contrast might be a European social-democratic party that arose out of a disciplined labor movement. Such parties were common between the world wars and just after.[17] Neither major American party has approximated that condition.[18] As for U.S. third parties, those that have been successful enough to seat members of Congress also existed to navigate peculiar federal-local coalition structures. Examples include the People’s Party of the 1890s, which often allied with different parties for presidential and lower elections, and the American Labor Party of the 1930s–1940s, which allied with Republicans in New York State (and City) but with Democrats at higher levels.[19] Modern-day fusion parties like the Conservatives or Working Families perform similar roles.[20] American political institutions present strong incentives to stick with the major parties, either as a fusion party or a faction.[21]

Third, to say that the electoral system creates the number of parties is somewhat backward. Scholars have long observed that multiparty politics precedes PR.[22] More recent research therefore focuses on: (1) how multiparty politics leads to PR adoption; (2) why a country with multiple parties might not adopt PR; and (3) why PR might get adopted in the absence of multiparty politics.

Multiparty politics historically led to list PR adoption for two reasons. One was a need to prevent insurgent parties from numerically dominating parliaments, which both list and non-list rules can do.[23] The other was for legislative leadership to gain control of party nominations. It often is forgotten that the move to PR in most countries replaced runoffs,[24] not plurality, because runoffs permitted insurgent parties to play kingmaker in decisive-round elections.[25] List PR centralized control of nominations and therefore gave party leaders control of their own coalitions. As for countries with multiparty politics but no PR (like the United Kingdom), the reigning explanation is that levers of nomination control already existed.[26] In turn, this may enable dominant parties (which typically need to ratify the change) to adapt to demands that otherwise might lead to sustained pressures for electoral reform.[27]

The story is different in countries like the U.S., where specifically non-list systems found traction amid factional politics. Prominent examples are the 110-year use of cumulative voting in Illinois (1870–1980) and the adoption of STV by 22 cities, between 1915 and 1947. Most of these cases were conspicuous for a lack of multiparty politics (New York City, one of the 22, is the glaring exception. This is because it already had multiple parties on reform adoption, owing to the coalition that elected Mayor Fiorello La Guardia in 1932).[28] Rather, the point of these reforms was for party factions to compete as factions. Also, these semi-PR systems were imposed by referenda, often following elected charter commissions or, in Illinois’ case, an elected constitutional convention. One of us has termed this adoption mode a “realigning” one because it involves rearrangement of the existing majority coalition (e.g., defectors working with formerly excluded groups).[29] When factions understand themselves as such, they are naturally averse to the nomination control that lists imply.

Recent adoptions of ranked-choice voting (RCV)–based systems are consistent with the continued role of party factions in American reform episodes. Because this push has been national in aspiration, some have advised reformers to consider list-based options not traditionally associated with their focus on local government. In turn, this has produced cautionary advice to limit experimentation to the state and local levels and/or find reforms that work with the two-party system.[30] (These goals are not exclusive.) STV advocates also counter that members of Congress are habituated to weak party discipline and courting independent voters.[31] Yet in practice, actual adoptions are aimed at either rearranging or maintaining factional coalitions at the state and local levels. Prominent examples of maintenance are Alaska, where nonpartisan RCV is thought to perpetuate a bipartisan coalition of moderates,[32] and NYC, where its use within party primaries can be seen as a strategy for managing factional conflict.[33] More recently, Portland (Oregon) has adopted STV alongside a larger city council both to diversify that body and to make it address problems of housing, homelessness, crime, and so forth.[34] One hears little about third parties playing prominent roles in these local contexts.

We return to the interplay of factions and non-list rules below. The core points for now are: (1) list and non-list rules differ fundamentally by whether they presume preelection groups of candidates who appear on ballots as such; (2) this distinction affects the way that PR is adopted; and (3) the two-party system may need to be reconceptualized as two fluid coalitions of factions that exist alongside the federal structure of U.S. government. For these reasons, we are skeptical that PR would unleash multiparty politics much beyond what already exists in advance of adoption. Governability is a different story, and we address that below.

Our case for lists rests largely on the problems with non-list rules under factional conditions. This is because the usual approach to electoral reforms in U.S. cities has been to turn to nonpartisan elections in some form. Then, if PR is desired, the nonpartisan ballot limits reformers to one of the semi-proportional options.[35]

Historically, coalitions supporting PR-style reforms have been bipartisan. More precisely, they tended to comprise the nondominant faction of a city’s larger major party and the dominant faction of its main opposition. One of us has researched observable traces of this characterization. There are two key results. First, recall that one attractive property of PR is the multi-seat district, which can unpack or un-crack a voting block not otherwise likely to win seats.[36] Note that PR is not necessary for this outcome; it can result from any system based on districts of sufficient geographic scope. This sets us up to interpret the first result: reform charters with STV tended to win adoption in relatively competitive locales (where the dominant/nondominant ratio was roughly 2 to 1), whereas non-STV reform charters based on similar district structures tended to be adopted in much more lopsided jurisdictions. Second, based on available case history, an STV reform charter tended to fail at referendum when the dominant party avoided a split. Third parties played little role, perhaps to the chagrin of modern-day reformers seeking multiparty politics.[37]

The second relevant point from history concerns our fictive Better City Association. This example is modeled on actual nonpartisan slating groups (NPSGs) that emerged to contest U.S. STV elections. While NPSGs existed alongside many urban electoral reforms during the Progressive Era,[38] they played special roles in STV. One was balancing the interests of the factions that constituted them.

Another special role was engaging in a set of practices called “vote management.” These are discussed at length with respect to STV in several publications.[39] During the 1990s, interest in cumulative voting began to produce similar literature.[40] Broadly, vote management involves nominating a number of candidates who seem likely to win, evening out the distribution of votes among them, and then taking steps to keep votes within the party generally. Difficulties with these strategies was one factor in STV’s U.S. abandonment,[41] in the repeal of cumulative voting in Illinois,[42] and in the replacement of limited voting in Japan with a system based partly on lists.[43]

Can vote management work under factional conditions? Much of the literature on non-U.S. cases suggests that it works best where voters’ party attachments are stable, possibly because the parties contesting local elections are primarily organized to contest national ones,[44] and also in parties where internal policy disagreement does not spill over into decisions about which or how many people will get nominated.[45]

The questions for reform-minded persons are whether they can perform vote management effectively and, if so, for how long—especially if national issues split local coalitions. Political scientists have been investigating such tensions from the opposite perspective, suggesting that city politics under the current “urban-rural” alignment could produce policy disagreement within the congressional parties.[46] Other work on specifically local elections suggests that voters whose party attachments conflict with their policy views might be unpredictable.[47] Ultimately, local and higher-level party alignments could be at odds under a BCA-type reform package, and this could frustrate vote management at elite and voter levels alike.

An additional issue in non-list systems that invite multiple ballot markings (like STV, cumulative voting, and some forms of limited voting) concerns voter decision-making with respect to candidates other than the first choice. This is an active research area, partly because iconic STV cases (such as Australia and Ireland) only recently have begun producing unexpected results.[48] Voters might mark their first ballot for a favorite candidate, and then rely on other heuristics when deciding about others. These include inferred race,[49] gender,[50] incumbent status,[51] and others unrelated to political party. One recent study of ranked-choice ballots found appreciable evidence of pattern marking—bubbling in grid-style ballots to make zigzags and other doodles.[52]

The demands of vote management in a factional two-party system make a strong case for list-based reforms over non-list options now in vogue in many cities. The simple list system removes several needs: to limit nominations, urge voters to vote for or rank an entire slate, and overcome nonparty heuristics in the use of secondary ballot markings. It also is worth mentioning that multiple-vote rules invite politicians to craft electoral coalitions that straddle party lines.[53] This may sound attractive from a pre-reform perspective, but it makes little sense if organizations like BCA will emerge under such reforms (and perhaps propel them).

Rigorous descriptions of electoral systems typically begin with five or six components: assembly size,[54] district magnitude (number of seats per district), list type (e.g., candidate vs. party voting), number of votes, ballot structure (e.g., ranked or something else), number and operation of tiers (e.g., see MMP below), rules about seat allocation,[55] and rules about nominations.[56] Rather than address each at length, we focus on a few and then describe some model systems. It is critical to remember that assessing various trade-offs requires understanding the components that make up these models. Some of the most important such components are as follows:

Assembly size—The number of seats on a city council, known as assembly size, exerts the greatest influence on how many interests can win direct representation (Illustration 1). Take a city in which 10% of voters choose Bull Moose Party candidates. Assuming party-loyal voting, the council must have at least 10 seats for a Bull Moose candidate to win one. Assembly size is therefore at least as important as the choice of a model electoral system if one aims to take representation seriously. Recent research also shows that large U.S. cities’ councils are much smaller, on average, than those of similar-size cities in other countries.[57] This is largely a result of the Progressive Era’s nonpartisan reforms discussed above. Reductions in assembly size should be scrutinized closely, even if they are twinned with reforms otherwise seen as desirable.[58]

Open, closed, or flexible lists—The list is the fundamental unit of vote-earning and seat allocation. Once a list has won enough votes to secure some number of seats, the question is which candidates get them. Policymakers must choose: open, closed, or something in between? A purely open-list system lets voters set list order through their choice of candidates, and the top candidates get seats (Figure 1). A closed system offers voters choice among lists but not candidates. Here, some internal party procedure decides the order in which candidates win seats.“Flexible” lists represent a middle position, letting voters change a candidate’s list position only if votes for that candidate cross some threshold. Open- and closed-list ballots are presented below.

Closed lists shouldn’t be dismissed out of hand as nonstarters (Figure 2). The process of setting list order can involve a primary election, as was recently proposed in Wyoming.[59] And closed-list ballots often feature party leaders’ names or photos. This is consistent with presidential voting in most U.S. states, whereby voters have chosen party tickets since the Twelfth Amendment was ratified. Moreover, scholars who study women’s descriptive representation tend to recommend closed lists because such candidates can be distributed throughout a list to improve their prospects of election.[60] In short, closed lists may not be as unreasonable a reform choice as many have assumed.

District magnitude—Along with assembly size, the number of seats per district sets an upper limit on the number of organized interests that can win direct representation. It also affects an electoral system’s ability to provide seats-to-votes proportionality. As district magnitude rises, so does the system’s ability to produce fine-grained seat distributions. That can mean that minor parties become better represented, or it can lead to a closer fit between the major lists’ shares of votes and seats.

Very large magnitudes, however, probably imply that more candidate names will appear on each party’s list (if the list is open or flexible). This can make ballots unwieldy and cause voter confusion and therefore should be avoided.[61]

Allocation rule—List systems allocate seats to parties in either of two main ways. One is to calculate a quota by dividing total votes cast in a district by that district’s magnitude. Another is to distribute seats one by one, giving the seat at each round to the party with the highest average number of votes per seat that it already has. There are several variations on each process, and these are debated in practice for whether they might favor small, medium, or large parties.[62] Illustration 2 uses a common highest-average method over a range of district magnitudes.

Single-seat districts—Forms of PR are available that incorporate single-seat districts (SSDs), or in which the district magnitude is equal to one. The paradigmatic example of this is mixed-member proportional (MMP) representation (Figure 3). The most common version of MMP gives voters two votes: one within their SSD, and a second for a party (sometimes called “compensation seats”). SSDs usually comprise most of the seats in the legislature, but not all; 33%–50% of seats might be set aside for party-list allocation (usually with closed lists). Determining winners involves, first, using whatever rule applies to the SSD, which is almost always plurality. Second, the compensation seats are distributed to each party in proportion to their vote shares earned in the party-level vote, taking account of seats already won in the SSDs.[63] The system thus combines SSD seats with seats allocated from closed lists.

MMP also can employ a single vote, in which a voter’s district-tier selection also governs seat allocation in the compensation (list) tier. As before, after the SSD candidates have been elected, the compensation seats are allocated to each party according to each one’s share of the vote, taking account of seats already won in districts. Other models incorporating SSD are invented with some frequency.[64]

There are two main reasons that one might want to incorporate SSDs via MMP. One is that the model has been shown to balance district-level interests against those concerning polity-wide policy (e.g., citywide). It should be noted, however, that these findings emerge from cross-national research on parliamentary regimes, not local elections.[65] (MMP is used locally in South Africa.) A second reason is that SSDs continue to enjoy support among voting-rights litigators, civil rights activists, and others who study groups with lower-than-average propensity to vote.[66]

Finally, as a less important but still noteworthy consideration, MMP may make it easier to graft PR onto an existing system of SSDs. The assembly size can simply expand by the number of “compensation” seats allocated from party lists, elegantly allowing all incumbents to retain their districts. Or, if assembly size is to remain the same, an existing SSD can be combined into a smaller number of geographically larger ones.

One likely objection to PR is the potential to replace single-party domination with multiparty paralysis. With only small numbers of votes necessary for parties to win seats, the argument goes, there results a high possibility for a tenuous coalition government.[67] Related to this objection is the possibility that ideologically extreme parties will gain a foothold in city councils elected via PR.[68]

We believe that the risk of multiparty fragmentation is overstated. First, local PR systems likely would comprise several districts of modest magnitude. As shown above, using a magnitude of roughly five members per district would imply a sufficiently high threshold to constrain proliferation of parties in the long run (indicated by the Bull Moose Party in Illustration 2).[69] Second, as argued above, the federal system and attendant imperative to maintain big-tent parties means that radical third parties might be fleeting. Experience is consistent with this conjecture: one such party under STV in Cincinnati in the 1930s which folded into the local Republican Party, and then the American Labor and Communist Parties under STV in NYC in the 1940s, much of which had folded into the Democratic Party by the mid-1950s.[70] This historical context should be taken into account when evaluating the foothold argument.

The proliferation of small and potentially extreme parties also might be limited through ballot-access requirements. For example, rather than ask candidates to collect signatures for ballot access, as is the current practice, parties could instead be required to collect signatures to ensure their slate’s place on each district ballot. A higher signature requirement would increase the effort, resources, coordination, and support necessary to secure ballot access, thereby making it harder for poorly funded or uncoordinated groups to enter the council. Requiring parties to file as such also might cause voters to closely evaluate party brands, in contrast with a petition system based fundamentally on candidates. The party’s need to provide a list of candidates for its slate also requires the party to have multiple credible candidates willing to carry the party banner.

PR systems vary, and this variation might have different impacts locally from what has been observed in cross-national research. At first glance, the larger geographic scope of simple list-PR districts seems as though it might dilute the power of groups opposed to new housing development. Currently, in many American cities, hyper-local issues like rezonings of discrete parcels of land (or “spot rezonings”) usually involve relatively small numbers of highly motivated constituents within a district who share a strong interest in preserving some aspect of the status quo. In PR elections (which, again, presume multi-seat districts), candidates favoring denser rezoning can appeal to voters across a wider geographic area, who may outnumber the antidevelopment constituency.

Multi-seat districts therefore might ameliorate not-in-my-backyard pressure against housing construction. For example, in cities such as New York and Chicago, councilmembers have an informal veto over proposed rezonings in their districts. Because such land-use decisions require city council approval, the council generally votes down rezonings opposed by the member from the district in which the proposed change is taking place. This “member deference” or “aldermanic privilege” chills new housing construction and business development, as homeowner-voters know that they can hold their member accountable at the ballot box for approving denser housing uses.[71] With an odd number of members per district, this veto power might be weakened, for example, by requiring a majority of a district’s representatives to disapprove of a proposed rezoning.

Additionally, those who favor more housing supply do so for several potentially unrelated reasons: lowering housing prices, protecting property rights, unlocking economic opportunity and growth, advancing racial equity, correcting past discriminatory practices, reducing carbon usage, promoting public-housing ownership, and more.[72] Unlike the concentrated and localized interests shared by antidevelopment neighborhood groups, the motivations for more housing supply do not directly make supporters more likely to live near one another geographically.[73] This suggests the utility of an electoral system that allows these geographically disadvantaged voters to join behind a common slate across a wider area of the city than current districts allow.

Under list PR, voters and candidates might coalesce more easily under distinct party brands (“Common Sense,” Socialist, Conservative, Green, MAGA, or something like our BCA example) or around specific policy objectives with citywide appeal (housing supply, carbon-dioxide reduction, support for police officers, faster subway and bus service, and so on).[74] This might organize voters by interest and political philosophy, but at the potential expense of reduced focus on local-level issues. Representatives may, however, develop informal territories in which they divide up multi-seat districts to provide constituent services, such as in the neighborhoods where members live.[75]

Mixed-member PR most obviously aims to balance district-specific and citywide interests. Members elected from SSDs might have an incentive to focus on local constituent services and respond to acute pressures for district-specific matters like spot rezonings, street cleanliness, and public safety. Meanwhile, the closed-list compensation contest might induce a mix of neighborhood-specific and citywide appeals. The exact mix likely would depend on a party’s overall priorities and willingness to enforce discipline across the tiers.

Preserving SSDs through MMP therefore might facilitate the continuation of member deference and other anti-density measures. Compensation-seat (or “at-large”) delegates might find it difficult to support a spot rezoning if their district-level co-partisans decline it. It is hard to say exactly what would happen under MMP, except that member deference and similar practices may become less stable (or “institutionalized,” in the parlance of social scientists).

That list PR hasn’t been adopted in any major American city might signal to policymakers that theirs should not be the first. However, any city with partisan nominations and ballots already has two of its key features. Primaries could be retained or not, and the act of voting itself could be made as simple as whatever currently exists in some city (if not made simpler). From this perspective, list PR is less of an administrative and behavioral departure than going to nonpartisan elections under some semi-proportional system like STV or cumulative voting. That said, understanding potential trade-offs within the family of list systems, as well as their downsides and how to mitigate them, is critical to successful electoral system design and implementation.

In New York City, voters in 2019 adopted single-seat ranked-choice voting (also known as the Alternative Vote) for local primary and special elections.[76] After two city council election cycles under these reforms, NYC’s political system remains largely the same as before. One party controls nearly 90% of city council seats, despite commanding less than 80% of the vote in the city council elections of 2021 and 2023.[77] Further, the single-seat variety of RCV now used in New York does not seek to reflect the electorate’s political diversity in the resulting legislative body. In this context, polling conducted by the Manhattan Institute in April 2024 and January 2025 finds broad appetite for reforms meant to open up primaries.[78] Given what we have said above, it might be worthwhile to consider something else.

One obvious proposal would be to resurrect STV, i.e., “proportional ranked-choice voting,” as now used in Portland, Oregon, and as used in the Big Apple in 1937–47. Proponents might point to the fact that that, during this time, at least four parties were represented in the city council, which became the site of lively, and even entertaining, debates.[79]

Two big factors contributed to STV’s undoing: factional splits throughout the local party system and an increasing share of voters ranking candidates from both sides of it. This period demonstrates the importance of the need for greater party discipline than offered by the semi-proportional PR approximations now in vogue.

New York City Council elections could be conducted through list PR by amending the city charter. This might involve any of the options sketched above, each with its own pros and cons. Most designs would obviate primary elections altogether, unless a party decided to retain them for drawing up its list. Such a package could make space for more parties, although we have emphasized the importance of party factions in U.S. politics. Borough-specific interests seem as though they might generate such factions, and that would be consistent with the 1937–47 experience.

What might a reformed New York City Council look like under list PR or its MMP variant? For simplicity’s sake, assume that the council retains its 51-member assembly size. The current breakdown of city council seats by borough is as follows:

Current Allocation of Single-Seat Districts

• The Bronx: 8
• Brooklyn: 16
• Manhattan: 10
• Queens: 14
• Staten Island: 3

Existing districts could be combined to facilitate a smooth transition, as they already comply with federal, state, and local districting rules. More realistically, the process to amend the city charter would take years, so PR districts would likely not need to be drawn before the next redistricting cycle, after 2030. Under PR, districts might be distributed as follows:

Multi-Seat Districts Under List PR

• The Bronx: two four-member districts[80]
• Brooklyn: two five-member districts and one six-member district
• Manhattan: two five-member districts
• Queens: two five-member districts and one four-member district
• Staten Island: one three-member district (alternatively, a four-member district by combining the area of Bay Ridge, Brooklyn, with Staten Island, allowing Brooklyn three five-member districts)

To retain 51 seats under an MMP system, current districts could be combined—especially those whose incumbents are term-limited or otherwise not running for reelection. Alternatively, the city charter could be amended to expand the city council to add new seats that would correspond to the new compensation tier. An expanded city council and its resulting expenses would thus have to be weighed against the desirability of retaining council districts similar to today’s.

In response to the potential loss of control by the local Democratic Party in a place like NYC, it’s important to recall that in many major American cities, voters are habituated to vote for Democratic Party candidates (or, in nonpartisan races, for candidates who align with Democrats). It’s reasonable to presume that this dynamic will continue under list PR, at least in the short run.[81] In NYC, for example, the overwhelming registration advantage that Democrats now have likely would translate into city council majorities, barring some shock to the city’s party system that induces mass party-switching or widespread ticket-splitting. Our running BCA Party might be seen as the outgrowth of such a shock. Others might involve voters getting accustomed to using local parties to support specifically local policy bundles.

American cities need not accept electoral systems that perpetuate single-party domination and all that it implies. As big cities act on a range of pressures for electoral reform, list PR should be included in the menu of options available to their leaders. List-based options could give structure to local coalitions on issues that don’t track national politics, opening new opportunities for bipartisan, or maybe multiparty, local democracy.

There also may be reasons to expect different policy effects from different electoral systems. Housing policy, for example, appears to be more pro-supply under at-large elections than under systems in which single-seat districts predominate, but the costs do not fall evenly on all socioeconomic groups.[82] Though political compromises and policy effects cannot be predicted with any comfortable degree of confidence, one of the models (MMP) may provide ways of balancing district-level concerns with city-level needs.

Finally, we have not raised questions about mayor-council versus council-manager government or how to manage conflict between an elected mayor and the city council.[83] List PR might mitigate this risk by making mayoral and council electoral majorities more likely to overlap.[84] Staggered elections are another consideration and are probably to be avoided. These issues are all worth debating.

We are living through a rare period in which many Americans want electoral reform in cities. List PR is worth including in that conversation.

Jack Santucci is a professorial lecturer at the George Washington University and a senior lecturer at Western New England University. He is an expert on electoral systems, subnational politics, and voting behavior. Santucci is the author of More Parties or No Parties: The Politics of Electoral Reform in America, coauthor of “Toward a Different Kind of Party Government: Proportional Representation for Federal Elections,” and several journal articles on parties-and-elections issues. His public-facing writing has appeared in Democracy Journal, Philadelphia Inquirer, Washington Post, New York Daily News and other outlets. Santucci previously taught at Drexel University and Queens College–CUNY. He has also held research positions in private philanthropy and democracy promotion. Santucci holds a Ph.D. in government from Georgetown University, an M.A. in democracy studies from the same, and a B.A. in political science from McGill University.

John Ketcham is a legal policy fellow and the Director of Cities at the Manhattan Institute. His areas of expertise include housing, local elections, infrastructure, parental rights, and more, particularly in New York City and State. He holds a J.D. from Harvard Law School and a B.S. from Fordham University. While in law school, Ketcham spent a summer at the Permanent Observer Mission of the Holy See to the UN and was treasurer of the Harvard Federalist Society and managing editor of the Harvard Journal of Law & Public Policy. He has been published in the Wall Street Journal, Newsweek, The American Mind, New York Daily News, New York Post, and City Journal. Ketcham was awarded a New York City Leonine Forum Fellow in 2021–22 and the Napa Legal Institute’s Good Counselor Project Fellowship in 2023–24.

Please see Endnotes in PDF

Photo: FilippoBacci / E+ via Getty Images