Loading...
The URL can be used to link to this page
Your browser does not support the video tag.
Home
My WebLink
About
03-25-26 Study Commission Agenda and Packet Materials
A. Call to Order with Pledge of Allegiance and Moment of Silence - 4:00 PM, Commission Room, City Hall, 121 North Rouse B. Changes to the Agenda C. Public Comment on Anything within the Jurisdiction of the Study Commission THE STUDY COMMISSION OF BOZEMAN, MONTANA SC AGENDA Wednesday, March 25, 2026 How to Participate: If you are interested in commenting in writing on items on the agenda please send an email to govreview@bozeman.net prior to 12:00 p.m. on the day of the meeting. At the direction of the Study Commission, anonymous public comments are not distributed to the Study Commission. Public comments will also be accepted in-person and through video conference during the appropriate agenda items but you may only comment once per item. As always, the meeting will be recorded and streamed through the Meeting Videos and available in the City on cable channel 190. For more information please contact Ex Officio, Mike Maas, 406.582.2321, or visit bozemanstudy.com. This meeting will be held both in-person and also using an online video conferencing system. You can join this meeting: Via Video Conference: Click the Register link, enter the required information, and click submit. Click Join Now to enter the meeting. Via Phone: This is for listening only if you cannot watch the stream, channel 190, or attend in- person United States Toll +1 669 900 9128 Access code: 951 6442 0347 This is the time to comment on any matter falling within the scope of the Bozeman Study Commission. There will also be time in conjunction with each agenda item for public comment relating to that item but you may only speak once per topic. Please note, the Study Commission cannot take action on any item which does not appear on the agenda. All persons addressing the Study Commission shall speak in a civil and courteous manner and members of the audience shall be respectful of others. Please state your name, and state whether you are a resident of the city or a property owner within the city in an audible tone of voice for the record and limit your 1 D. Consent Agenda D.1 Approval of Study Commission Minutes(Heinen) E. Correspondence or Study Commission Update E.1 Final Survey Report for Community Engagement (Franks) E.2 New Calendar plan(Cestero/Franks) F. Unfinished Business F.1 Review and Discuss Suggested Edits for Article III Draft (Cestero/Franks) G. New Business G.1 Learning session for Wards/Districts(Clark ) H. Future Agenda Items H.1 Future Education Discussion and Needs for April 2 and April 15 Meetings (Heinen) I. Public Comment on Anything within the Jurisdiction of the Study Commission J. Announcements K. Adjournment comments to three minutes. Written comments can be located in the Public Comment Repository. Consider the Motion: I move to approve the Study Commission meeting minutes from March 5, 2026. May 7: Review of draft Charter language and submit for legal review and ask for opportunity of a Minority Report May 13: Final approval of draft Charter Language May 28: Approval of draft tentative report and release for public comment June 4: Review and discuss public and legal comments June 17: Public hearing on draft tentative report that includes the Charter language July 15: Study Commission input on draft Report and Charter language July 30: Approve Ballot Language to be submitted August 6: Final adoption of Study Commission Report August 10: Ballot language due to County election department August 19: Develop Ballot Election plan September 3 & 16 October 1 & 14 Consider the Motion: I move to approve the suggested edits for Article III of the Bozeman City Charter. Study Commission meetings are open to all members of the public. If you have a disability that requires assistance, please contact the City of Bozeman's ADA Coordinator, David Arnado, at 406.582.3232. 2 Study Commission meetings are televised live on cable channel 190 and streamed live on our Meeting Videos Page. 3 Memorandum REPORT TO:Study Commission FROM:Caeleb Heinen, Recording Secretary Mike Maas, Ex Officio SUBJECT:Approval of Study Commission Minutes MEETING DATE:March 25, 2026 AGENDA ITEM TYPE:Citizen Advisory Board/Commission RECOMMENDATION:Consider the Motion: I move to approve the Study Commission meeting minutes from March 5, 2026. STRATEGIC PLAN:1.1 Outreach: Continue to strengthen and innovate in how we deliver information to the community and our partners. BACKGROUND:Attached are the written minute summaries from the previous two meetings. Future meeting minutes ought to be approved at the next schedule Study Commission meeting. All past meeting recordings are available for review on the City's Meeting Videos page. UNRESOLVED ISSUES:None identified ALTERNATIVES:As per the Study Commission FISCAL EFFECTS:None Attachments: 03-05-26 Study Commission Meeting Minutes.docx Report compiled on: March 12, 2026 4 Bozeman City Commission Meeting Minutes, March 5, 2026 Page 1 of 6 THE CITY COMMMISSION MEETING OF BOZEMAN, MONTANA MINUTES March 5, 2026 A) 00:02:35 Call to Order with Pledge of Allegiance and Moment of Silence - 4:00 PM, Commission Room, City Hall, 121 North Rouse Present: Carson Taylor, Becky Franks, Barb Cestero, Deanna Campbell, Jan Strout, Mike Maas Absent: None Excused: None B) 00:03:50 Changes to the Agenda C) 00:03:58 Public Comment on Anything within the Jurisdiction of the Study Commission 00:05:13 Abby Jones gave public comment D) 00:08:51 Consent Agenda D.1 Study Commission Claims Review and Approval D.2 Approval of Study Commission Minutes 02-18-26 Study Commission Meeting Minutes.pdf 00:09:08 Motion to approve I move we approve the Consent Agenda for March 5 meeting. Barb Cestero: Motion Jan Strout: 2nd 00:09:22 Vote on the Motion to approve I move we approve the Consent Agenda for March 5 meeting. The Motion carried 5 - 0. Approve: Carson Taylor Becky Franks Barb Cestero 5 Bozeman City Commission Meeting Minutes, March 5, 2026 Page 2 of 6 Deanna Campbell Jan Strout Disapprove: None E) 00:09:32 Correspondence or Study Commission Update 00:09:40 Cmr. Cestero discussed Valley Unit Neighborhood Association (VUNA) presentation with Cmr. Taylor 00:12:03 Cmr. Taylor discussed Radio show with Dave Wooten 106.9 The Eagle with Cmr. Campbell, as well as their interview with Gallatin Valley Sentinel 00:12:39 Cmr. Franks discussed extra research by Bozeman Resident Hal Stanley, specifically from Bend Oregon 00:16:00 Cmr. Strout discussed a future program with League of Women Voters with Cmr. Franks (Monday March 23), Pecha Kucha (April 21 and 22), and Bozeman Professional Women for Voter Education. Cmr. Strout discussed a conversation with former Missoula Study Commissioner Sarah Bates 00:19:00 Cmr. Campbell reached out with Osher Lifelong Learning Institute (OLLI) for a follow up interview F) 00:19:29 Unfinished Business F.1 00:19:34 Review draft Charter language for sections 2.01, 2.03 and 2.06 DRAFT sec 2.01 2.03 and 2.06.pdf 00:19:43 Cmr. Taylor outlined the changes in Charter language 00:21:03 Cmr. Cestero gave an overview on the suggested edits for 2.01, 2.03 and 2.06 00:25:12 Cmr. Franks clarified the differences between Charter, Ballot and Recommendation language 00:26:13 Questions and discussion on suggested edits to Charter language 00:46:03 Discussion on the scope of recommendations for 2.01, 2.03 and 2.06 F.2 00:48:03 Discussion and Decisions Items Presiding Officer or Mayor.pdf Various Models of Executive Powers in Municipal Government.pdf Model City Charter 9th Edition.pdf 23- Ordinances 2152 - Amending the Salaries and Expenses of the City Commissioners and Mayor of the City of Bozeman.pdf BCSC - City Commission Notes.pdf BCSC - Board Survey Results.pdf RE_ Bozeman AMI.pdf 6 Bozeman City Commission Meeting Minutes, March 5, 2026 Page 3 of 6 Lansing Charter Summary Report Final.pdf Mayor and Commission 3 cities.pdf Recruitment Brochure - City Manager 2020.pdf 1150 City Manager 08 2022.doc Comparison Chart amended2.xlsx Clark on CM contracts.docx 00:48:18 Cmr. Franks introduced the item 00:50:47 Motion to approve I will move that we continue with the structure of Commissioner-City Manager within the context of our Charter form of Government. Carson Taylor: Motion Jan Strout: 2nd 00:51:15 Discussion on motion on the floor including Study Commissioner's position on the motion 00:58:14 Patti Steinmuller gave public comment 00:59:10 Vote on the Motion to approve I will move that we continue with the structure of Commissioner- City Manager within the context of our Charter form of Government. The Motion carried 5 - 0. Approve: Carson Taylor Becky Franks Barb Cestero Deanna Campbell Jan Strout Disapprove: None 00:49:13 Ex Officio Maas clarified language for a potential motion 00:59:34 Discussion on 3.01 of Bozeman Charter Bozeman Charter 01:03:08 City Attorney Greg Sullivan discussed his experience with previous City Managers and removals 01:08:04 Study Commissioner questions for City Attorney Greg Sullivan City Manager Contract (May 13 2025) 01:14:00 Study Commissioner's discussion on Article III and their viewpoints on it 7 Bozeman City Commission Meeting Minutes, March 5, 2026 Page 4 of 6 01:27:16 Discussion on reviewing City Manager annually in 3.01 01:31:13 Cmr. Franks introduced Recommendation opportunity on how staff, City Commission and City Manager communicate Resolution 5124 01:34:03 Discussion on City Manager, staff and City Commission communication 01:48:25 Public Comment related to all of F.2 01:49:09 Hillary Brown gave public comment 01:51:45 Jena Reno gave public comment 01:55:20 Summary of current edits to the City Charter 01:55:57 Recommendation on using established norms for City Manager as part of the current hiring process and discussion Resolution 5323 02:01:40 Discussion on the status of the mayor pertaining to workload and pay 02:27:28 Public Comment on status of the Mayor 02:27:46 Jena Reno gave public comment 02:29:15 Daniel Carty gave public comment 02:31:07 Hillary Brown gave public comment 02:32:42 Continued discussion on status of the Mayor, including job description 02:50:15 City Attorney Greg Sullivan clarified points about job descriptions written in the Charter 02:56:32 Motion to approve I move that we take the job description regarding Commissioners out of the Charter to combine it with the recommendation that the City Commission create a document for roles, responsibilities and expectations for the Mayor and the City Commissioners, and add after the and "it has been demonstrated to be a full time position and shall be a voting member" Becky Franks: Motion Jan Strout: 2nd 02:58:22 Discussion on the motion 8 Bozeman City Commission Meeting Minutes, March 5, 2026 Page 5 of 6 03:08:48 Cmr. Franks withdrew her original motion 03:08:51 Motion to approve We add that resolution about how they should make a document describing the roles, responsibilities, obligations, expectations for the Mayor and Commissioner Position have it reviewed every 5 years and then we just leave the full time thing out and we don't touch the pay. Becky Franks: Motion Deanna Campbell: 2nd 03:09:19 Discussion on the motion 03:14:27 Vote on the Motion to approve We add that resolution about how they should make a document describing the roles, responsibilities, obligations, expectations for the Mayor and Commissioner Position have it reviewed every 5 years and then we just leave the full time thing out and we don't touch the pay. The Motion carried 4 - 1. Approve: Becky Franks Barb Cestero Deanna Campbell Jan Strout Disapprove: Carson Taylor 03:14:37 Chairman Taylor passed the gavel to Vicechair Franks 03:14:43 Motion to approve I move that we amend 2.03 to read under 2.03 A.) Powers and Duties: the Mayor is a full time position and shall be a voting member. Carson Taylor: Motion Jan Strout: 2nd 03:15:15 Discussion on the motion 03:22:58 Vote on the Motion to approve I move that we amend 2.03 to read under 2.03 A.) Powers and Duties: the Mayor is a full time position and shall be a voting member. The Motion failed 2 - 3. Approve: Carson Taylor Jan Strout Disapprove: Becky Franks Barb Cestero Deanna Campbell 9 Bozeman City Commission Meeting Minutes, March 5, 2026 Page 6 of 6 03:23:18 Discussion on making a Recommendation on Mayor and Commission Salary 03:30:18 City Attorney Greg Sullivan suggested relating the conversation to compensation over salary 03:32:00 Discussion was tabled until April 15 MSU Local Government Center to find further research on Commissioner/Mayor Compensation G) 03:35:39 New Business G.1 Future Education Discussion and Needs for March 25, April 2 and April 15 Meetings 03:35:53 Educational needs for upcoming meetings discussion H) Future Agenda Items I) Public Comment on Anything within the Jurisdiction of the Study Commission J) Announcements K) 04:02:25 Adjournment 10 Memorandum REPORT TO:Study Commission FROM:Caeleb Heinen, Recording Secretary Mike Maas, Ex Officio Becky Franks, Study Commission Vice Chair Working Ventures SUBJECT:Final Survey Report for Community Engagement MEETING DATE: March 25, 2026 AGENDA ITEM TYPE: Citizen Advisory Board/Commission RECOMMENDATION: Final Survey Report for Community Engagement for distribution STRATEGIC PLAN: 1.1 Outreach: Continue to strengthen and innovate in how we deliver information to the community and our partners. BACKGROUND: Final Survey Report for Community Engagement for distribution UNRESOLVED ISSUES: None identified ALTERNATIVES: As per the Study Commission FISCAL EFFECTS: None identified Attachments: 2026.03.16 - BCSC - March '26 Communications Strategy Report.pdf Report compiled on: March 19, 2026 11 | PAID MEDIA ACTIVITY/PERFORMANCE Print and digital ads that offered insights on meeting dates and times; drove traffic to the Study Commission website. Social Media (Facebook & Instagram) SPRING 2026 BCSC COMMUICATIONS STRATEGY MARCH 2026 REPORT JANUARY 2026 FEBRUARY 2026 Cont. on next page. 12 | SPRING 2026 BCSC COMMUICATIONS STRATEGY MARCH 2026 REPORT MARCH 2026 SOCIAL MEDIA POSTS See the posts that have run.Interactive link below.Static link here. 13 | SPRING 2026 BCSC COMMUICATIONS STRATEGY MARCH 2026 REPORT PAID MEDIA ACTIVITY/PERFORMANCE Print and digital ads that offered insights on meeting dates and times; drove traffic to the Study Commission website. Newspaper (Bozeman Daily Chronicle) Cadence: Once‑per‑week print ad placements (tear sheet visuals available at the end of this report) Reach: Circulation is ~15,000 print; significantly higher digital reach (exact digital circulation varies by article) Newspaper (MSU Exponent) Cadence: Once per week through end of April 2026 (end of their publication cycle for the school year). Reach: Circulation is free access to MSU’s 54,000+ student body and 5,722 faculty and staff. Newspaper Ads Tear SheetsSee the ad in context. 14 EARNED MEDIA ACTIVITY/PERFORMANCE City of Bozeman Newsletter: Copy developed and deployed via city staff. Press Release: Developed and distributed via city staff to local and regional outlets. Study Commission to Hold February Meeting Focused on Executive Structure Third‑Party Media Coverage (February 1 - March 15, 2026): Gallatin Valley Sentinel interview with Carson Taylor and Deanna Campbell Gallatin Valley Sentinel: “The Quiet Power Grab You’ve Barely Heard About.” No major coverage found during this time period from: Bozeman Daily Chronicle Montana Free Press TV Station Websites RFP Notices Ballot Education RFP - Boeman Daily Chronicle | SPRING 2026 BCSC COMMUICATIONS STRATEGY MARCH 2026 REPORT 15 Memorandum REPORT TO:Study Commission FROM:Caeleb Heinen, Recording Secretary Mike Maas, Ex Officio Barb Cestero, Study Commissioner Becky Franks, Study Commissioner Vice Chair SUBJECT:New Calendar plan MEETING DATE: March 25, 2026 AGENDA ITEM TYPE: Citizen Advisory Board/Commission RECOMMENDATION: May 7: Review of draft Charter language and submit for legal review and ask for opportunity of a Minority Report May 13: Final approval of draft Charter Language May 28: Approval of draft tentative report and release for public comment June 4: Review and discuss public and legal comments June 17: Public hearing on draft tentative report that includes the Charter language July 15: Study Commission input on draft Report and Charter language July 30: Approve Ballot Language to be submitted August 6: Final adoption of Study Commission Report August 10: Ballot language due to County election department August 19: Develop Ballot Election plan September 3 & 16 October 1 & 14 STRATEGIC PLAN: 1.1 Outreach: Continue to strengthen and innovate in how we deliver information to the community and our partners. BACKGROUND: Study Commissioners Cestero and Franks have provided a new Calendar outline, to be review by the Study Commission body. UNRESOLVED ISSUES: None identified ALTERNATIVES: As per the Study Commission FISCAL EFFECTS: none Report compiled on: March 12, 2026 16 Memorandum REPORT TO:Study Commission FROM:Caeleb Heinen, Recording Secretary Mike Maas, Ex Officio Barb Cestero, Study Commissioner Becky Franks, Study Commissioner Vice Chair SUBJECT:Review and Discuss Suggested Edits for Article III Draft MEETING DATE: March 25, 2026 AGENDA ITEM TYPE: Citizen Advisory Board/Commission RECOMMENDATION: Consider the Motion: I move to approve the suggested edits for Article III of the Bozeman City Charter. STRATEGIC PLAN: 1.1 Outreach: Continue to strengthen and innovate in how we deliver information to the community and our partners. BACKGROUND: Further discussion and approval of the suggested edits for Article III made by Study Commissioners Cestero and Franks. The language of the edits is attached. UNRESOLVED ISSUES: None identified ALTERNATIVES: As per the Commission FISCAL EFFECTS: None identified Attachments: Ariticle III DRAFT Suggested Edits.docx.pdf Report compiled on: March 12, 2026 17 Article III Edits: Suggested Edits following direction of the majority vote for the Bozeman Study Commission ARTICLE III CITY MANAGER Section 3.01. Appointment; Qualifications; Compensation. The city commission, by a majority vote of its total membership, shall appoint a city manager for an indefinite term and fix the manager's compensation. The city manager shall be appointed solely on the basis of education and experience in the accepted competencies and practices of local government management. The manager need not be a resident of the city or state at the time of appointment but may reside outside the city while in office only with the approval of the commission. The City Manager shall be provided with an annual comprehensive performance review. Recommendation: In the process of hiring a City Manager, the Commission must discuss with the candidates the Resolutions related to the norms of how the Commissioners communicate with City employees. 18 Memorandum REPORT TO:Study Commission FROM:Caeleb Heinen, Recording Secretary Mike Maas, Ex Officio Dan Clark, MSU Local Government Center SUBJECT:Learning session for Wards/Districts MEETING DATE: March 25, 2026 AGENDA ITEM TYPE: Citizen Advisory Board/Commission RECOMMENDATION: Participate in a Learning Session with Montana State University Local Government Center's Dan Clark on the specificities of Wards and Districts. STRATEGIC PLAN: 1.1 Outreach: Continue to strengthen and innovate in how we deliver information to the community and our partners. BACKGROUND: Mandatory Reading: At-Large VS Wards Wards Documents for Bozeman 2026 Supplemental Information: District Based Elections and Class Based Representation Evidence from the California Voting Rights Act Free Cities Center Analysis City Representation MIT At Large Elections and Minority Representation in Local Government Primer Districts vs. At-Large Elections Digital Trade-Offs between at-large and single member districts UofHouston Wards at-large in Canadian Cities Wards vs At Large on Diversity At-Large vs Ward - Implications for Public Infrastructure Hybrid City Council Election Systems UNRESOLVED ISSUES: None Identified ALTERNATIVES: As per the Study Commission FISCAL EFFECTS: None Identified Attachments: 19 At-large VS Wards 3.12.26.pdf Wards Documents for Bozeman 2026.pdf 1. district-based-elections-and-class-based-representation- evidence-from-the-california-voting-rights-act (1).pdf 2. pacificresearch.org-Free Cities Center Analysis City representation Single-member districts versus at large (1) (1).pdf 3. MIT At Large Elections and Minority Representation in Local Government.pdf 4. Primer_DistrictvsAt-LargeElections_Digital-1.pdf 5. Trade-Offs between at-large and single member districts- UofHouston.pdf 6. Wards at-large in Canadian Cities Candian Journal of Poli Sci.pdf 7. Wards vs At large on Diversity.pdf 8. At-Large versus Ward -Implications for Public Infrastructure.pdf 9. Hybrid City Council Election Systems.pdf Report compiled on: March 19, 2026 20 Prepared by the MSU Local Government Center. For Educational use only. For interpretation of the law, please seek competent legal counsel. Comparison of Electoral Systems in Medium-Sized Cities Prepared by the MSU Extension Local Government Center Introduction The City of Bozeman has a long history of conducting its elections on an at-large basis. When the study commission first proposed a city charter in 1976, it recommended electing city commissioners nominated by wards but elected at-large; however, that charter was ultimately rejected by the voters. In the Final Report prepared by the City’s first local government review, the study commission documented findings from a community survey indicating concerns about responsiveness in local government. As the report noted, citizens expressed “some dissatisfaction with the apparent unwillingness of local government officials to listen to them. The commission feels that the election of some of the commissioners by districts may cause those, so elected, to be more responsive to residents of their [ward]. A recognized danger in this method of electing commissioners is that they may become overly responsive to the citizens of their area. This may result in less concern for issues that affect the whole city and may cause a lack of harmony among commissioners. Election of all commissioners by the total electorate of the city will reduce the likelihood that these situations will develop.” Fifty years ago, the study commission had identified many of the same advantages and disadvantages associated with ward-based and at-large elections that continue to shape the discussion today. Below is a review of the advantages and disadvantages of elections conducted by wards, at-large, and hybrid systems (a combination of ward-based and at-large elections). These conclusions are drawn from peer-reviewed academic literature and applied research guides, including publications from the American Journal of Political Science, Political Science Research and Methods, and the Canadian Journal of Political Science, as well as practitioner-focused resources such as the Democracy Reform Primer Series produced by the University of Chicago’s Center for Effective Government at the Harris School of Public Policy, and analyses from the Pacific Research Institute. Artificial intelligence tools were used to synthesize and organize findings across these sources. ELECTIONS BY WARD Below is a list of advantages and disadvantages of elections by ward. This is a familiar practice in Montana and is the statutory default option for those municipalities in Montana that have selected the commission-executive form of government. Advantages 21 Prepared by the MSU Local Government Center. For Educational use only. For interpretation of the law, please seek competent legal counsel. • Increases descriptive representation for racial and ethnic minorities o District elections increase representation for racial/ethnic minorities when groups are moderately sized and geographically concentrated. [Source 1 & 4] o Conversion from at-large to district elections produces dramatic gains in minority officeholding when minority population share is sufficiently large. [Source 3] • Enhances representation of lower-income neighborhoods o District systems increase the proportion of elected officials from lower-income and high-renter neighborhoods. • Stronger neighborhood advocacy o Ward councilors focus on geographically defined neighborhood interests and are more responsive to localized concerns. [Source 6] o Ward-based systems create a “representational focus” on smaller areas, improving neighborhood-level responsiveness. • Lower campaign costs and higher local engagement o Candidates campaign within a smaller geographic unit, which reduces cost barriers and encourages more candidates to run. [Source 4] Disadvantages • Risk of parochialism (“ward politics”) o Councilors prioritize neighborhood interests over citywide needs, which can fragment policy decision-making. • Encourages logrolling and pork-barrel spending o Ward systems incentivize vote trading to direct geographically concentrated spending to each member’s area. [Source 8] o This can raise total municipal spending on localized capital projects. • May disadvantage women candidates o Empirical evidence shows women are less likely to be elected in single-member districts compared to at-large elections. [Source 4] • Ineffective when minorities are not geographically concentrated o District representation benefits minorities only in cities where groups are sufficiently large and segregated. Otherwise, districting provides little or no improvement. [Source 4] Options for Elections by Wards If Bozeman were to retain five elected commissioners, one option would be to continue electing the mayor at-large while electing the remaining four commissioners from wards 22 Prepared by the MSU Local Government Center. For Educational use only. For interpretation of the law, please seek competent legal counsel. drawn to be as compact and equal in population and geographic area as practicable. Under this approach, each ward would be represented by a commissioner serving approximately 14,500 residents. By comparison, districts for the Montana House of Representatives generally include between 10,000 and 11,000 residents. To further reduce the number of residents represented by each commissioner, the study commission could propose adding two additional commission seats. Expanding the commission in this manner would allow for six wards, with each commissioner representing just under 9,700 residents and the mayor elected at-large. Possible charter language: Following each federal decennial census, the city commission shall divide the city into [the number of commissioner wards corresponding to the number of commissioners elected by ward], ensuring that the wards are as compact as practicable and equal in population and geographic area. Apportionment may occur at any time for the purpose of equalizing population and area among commissioner wards. However, a commissioner ward may not be changed in a way that affects the term of office of any city commissioner who has been elected. Additionally, changes to the boundaries of any commissioner ward may not be made between the date that is six months prior to a city commissioner primary election and the date of the general election. AT-LARGE ELECTIONS At-large elections are the statutory default option for selecting city commissioners in the Commission-Manager form of government. Advantages • Encourages citywide perspective and policy coordination o At-large councilors see their constituency as the entire city, which promotes citywide rather than neighborhood-based decision-making. [Source 6] • Reduces neighborhood parochialism o At-large elections were originally adopted to reduce corruption and the hyper-localism associated with ward systems. [Source 6] • Promotes gender diversity o Research finds at-large systems tend to elect more women compared to district systems. [Source 5] • Less susceptible to gerrymandering o Because the entire city votes, manipulation of district boundaries is not possible. [Source 4] 23 Prepared by the MSU Local Government Center. For Educational use only. For interpretation of the law, please seek competent legal counsel. Disadvantages • Marginalizes racial and ethnic minorities o At-large systems historically diluted minority voting power and still suppress minority representation in many contexts. [Source 4] o Conversion from at-large to district elections consistently increases minority representation when minority groups are sizable. [Source 3] • Higher campaign costs o Campaigning citywide is significantly more expensive, limiting candidate diversity. [Source 2] • Overrepresentation of business or elite interests o At-large representatives tend to align more with citywide economic interests (business elites) rather than neighborhood-level concerns. [Source 6] • Lower responsiveness to neighborhoods o Councilors elected citywide pay less attention to geographically specific issues. [Source 6] HYBRID (MIXED WARD + AT-LARGE) SYSTEMS Research on hybrid systems is more limited, but emerging studies offer clear patterns. Advantages • Balances neighborhood representation with citywide perspective o Hybrid councils that are majority district-based retain the representational benefits of ward systems but incorporate at-large seats that promote citywide coordination. • Improves gender diversity o Including at-large seats in an otherwise ward-based system can increase the election of women compared to pure ward systems. • Reduces extreme drawbacks of both systems o Hybrids can prevent excessive parochialism while still protecting minority and neighborhood voices. [Source 9] Disadvantages • Minority representation depends on the proportion of ward seats o Hybrids with only a small number of district seats may still dilute minority representation compared to pure districts. [Source 9] 24 Prepared by the MSU Local Government Center. For Educational use only. For interpretation of the law, please seek competent legal counsel. • More complex system design o Hybrid systems require balancing ward boundaries, at-large ballot design, and potential structural inequities. • Can perpetuate inequities if at-large seats dominate o Research warns that if at-large seats form a majority, hybrid systems replicate the underrepresentation issues of pure at-large models. [Source 4] Hybrid Options for Bozeman One option for structuring the City Council is to retain a five-member council, with one member serving as mayor. Under this approach, the mayor would continue to be elected at-large by voters citywide, while the remaining four council members would be associated with geographic wards. Those four wards would be drawn to be as compact and equal in population and geographic area as practicable. Within this structure, a key design choice is how the four ward associated council members would be elected. Under one variation, each of the four council members would be required to both reside in and be elected by voters within their respective wards. This approach would emphasize direct neighborhood representation and accountability between council members and the residents of their ward. Alternatively, the four council members could be required to reside in a ward but continue to be elected at-large by the entire electorate. This option would preserve a citywide voting base while ensuring geographic diversity among council members, blending elements of ward representation with at-large elections. Gallatin County uses this same method when electing the County Commissioners. Another option would be to expand the council to seven members, which would reduce the number of residents represented by each councilor and increase overall geographic representation. Under this model, one member would serve as mayor and be elected at-large, along with one additional council member elected at-large. The remaining five council members would reside in and be elected from five wards drawn to be as compact and equal in population and geographic area as practicable. This structure blends ward-based representation with at-large positions, balancing neighborhood responsiveness with a citywide perspective. It would also spread representation more evenly across the city while retaining at-large seats to address issues affecting the community as a whole. Summary Table 25 Prepared by the MSU Local Government Center. For Educational use only. For interpretation of the law, please seek competent legal counsel. Model Advantages Disadvantages Ward Strong minority & low income representation; localized responsiveness; lower campaign costs Parochialism; logrolling; may reduce gender representation; ineffective where minorities not concentrated At-Large Unified citywide policy view; more women elected; less gerrymandering Dilutes minority representation; higher costs; elite dominance; weak neighborhood responsiveness Hybrid Balanced citywide + neighborhood representation; better gender diversity; moderates extremes Minority gains depend on ward ratio; can replicate at large inequities; structurally complex Sources: 1. Cushing-Daniels M, Jones D and Shannon B (2026) District-based elections and class- based representation: evidence from the California Voting Rights Act. Political Science Research and Methods, 1–17. https://doi.org/10.1017/psrm.2025.10070 2. Warnken, M. (2025, February 13). City representation: Single-member districts versus at-large (Free Cities Center Analysis). Pacific Research Institute. https://www.pacificresearch.org/free-cities-center-analysis-city-representation-single- member-districts-versus-at-large/ 3. Abott, Carolyn and Magazinnik, Asya. 2020. "At-Large Elections and Minority Representation in Local Government." American Journal of Political Science, 64 (3). 4. Trounstine, J. (2024). District vs. at-large elections: A practical research guide (Democracy Reform Primer Series). University of Chicago Center for Effective Government. https://effectivegov.uchicago.edu 5. Hofer, S., Huang, C., & Murray, R. (2018, October). The trade-offs between at-large and single-member districts (Hobby School of Public Affairs White Paper Series No. 14). University of Houston, Hobby School of Public Affairs. 6. Koop, R., & Kraemer, J. (2016). Wards, at-large systems and the focus of representation in Canadian cities. Canadian Journal of Political Science / Revue canadienne de science politique, 49(3), 433–448. https://doi.org/10.1017/S0008423916000512 [Wards at-l...f Poli Sci | PDF] 7. Trounstine, J., & Valdini, M. E. (2008). The context matters: The effects of single-member versus at-large districts on city council diversity. American Journal of Political Science, 52(3), 554–569. 26 Prepared by the MSU Local Government Center. For Educational use only. For interpretation of the law, please seek competent legal counsel. 8. Dalenberg, D. R., & Duffy-Deno, K. T. (1991). At-large versus ward elections: Implications for public infrastructure. Public Choice, 70(3), 335–342. https://www.jstor.org/stable/30025474 [At-Large v...astructure | PDF] 9. Jasso, F., & Krebs, T. (n.d.). Hybrid city council election systems (Informational brief). Pat Brown Institute for Public Affairs, California State University, Los Angeles; University of New Mexico. [Hybrid Cit...on Systems | PDF] 27 LOCAL GOVERNMENT CENTER | QUASI JUDICIAL HANDOUT | 01LOCAL GOVERNMENT CENTER | WARDS vs. DISTRICTS | 01 DIFFERENCE BETWEEN “WARDS” AND “DISTRICTS” IN MONTANA In the Montana Code Annotated (MCA), the terms “ward” and “district” are not interchangeable, they apply to different types of local governments and serve different legal purposes. Created only in municipalities (cities and towns) Used for electing city/town council members and defining representation within the municipality Boundaries changed only by ordinance Must have at least 150 electors to form a new ward Number of wards depends on the class of the city (1st, 2nd, 3rd class or town) WARDS Used only in municipalities (cities and towns) Under MCA 7 5 4401, cities and towns must divide themselves into wards for election and other municipal purposes. These wards must be roughly equal in population, and statute prescribes how many wards different classes of cities must have. If the City of Bozeman were operating under general powers and elected its council members by wards, MCA 7 2 4401(2)(a) would control, requiring the city to establish no fewer than four wards and no more than ten wards. However, because Bozeman operates under a voter approved charter, the Charter itself may determine the number of wards or a hybrid of wards and at large and is not required to conform to this statutory expectation and range. Characteristics of Wards The purpose of Wards ensure representation within a city/town by dividing the municipality into smaller election units. Used primarily in counties (and sometimes in multi-jurisdictional special entities) DISTRICTS The MCA uses districts in two major ways: As county election districts Counties use commissioner districts or other “districts” for electing county officials. These are not called “wards” as that term is strictly used when referencing a municipal service or voting area. Butte Silver Bow Consolidated City/County and Anaconda Deer Lodge Consolidated City/County use commissioner districts As “special districts” for delivering services Special districts (park districts, fire districts, rural improvement districts, etc.) are legally defined in MCA 7-11-1002. A “special district” means a unit of local government that is authorized by law to perform a single function or a limited number of functions. These may be countywide or multi jurisdictional. 1 2 28 LOCAL GOVERNMENT CENTER | QUASI JUDICIAL HANDOUT | 02 Prepared by the MSU Local Government Center. For Educational use only. For interpretation of the law, please seek competent legal counsel. Characteristics of Districts Created by counties (or jointly by cities and counties) Used for: Electing county commissioners (in districted counties) Organizing service delivery (fire, road, RID, SID, mosquito, cemetery districts, etc.) Special districts can cross city boundaries and are defined in statute depending on the district type Purpose Districts are designed to: Structure county representation (e.g., commissioner districts) Provide specific public services in a defined geographic area Relationship Between Wards and Districts in Election Law MCA 13-3-104 clarifies that wards (municipal) and districts (county) are separate units, but precincts may be drawn in ways that intersect both, so long as voters can be clearly identified for each unit. This reinforces that wards = municipal units, and districts = county or service units. LOCAL GOVERNMENT CENTER | WARDS vs. DISTRICTS | 02 29 Political Science Research and Methods (2026), 1–17 doi:10.1017/psrm.2025.10070 ORIGINAL ARTICLE District-based elections and class-based representation: evidence from the California Voting Rights Act Michaela Cushing-Daniels1 , Daniel Jones1 and Brooke Shannon2 1School of Public and International Affairs, University of Pittsburgh, Pittsburgh, PA, USA and 2Department of Political Science, University of Memphis, Memphis, TN, USA Corresponding author:Daniel Jones; Email:dbj10@pitt.edu (Received 14 March 2024; revised 13 March 2025; accepted 22 April 2025) Abstract District-based elections are a central feature of local governance throughout the United States. Prior work hasexploredwhetherdistrict-basedelectionsimpactracial/ethnicdescriptiverepresentationinlocaloffice; much less is known about the impacts of local district-based elections on other dimensions of representa- tion. We consider another such dimension: socioeconomic class. To explore how district-based elections shape the composition of locally elected officials on class dimensions, we focus on city councils and study thedramaticshifttowardsdistrict-basedelectionsinCaliforniainthe2010s.Weconstructastatewidemap- ping of newly drawn council districts; we also draw on rich and partially hand-collected data on council candidates and members. We find that district-based elections increase the share of candidates and coun- cil members from lower-income and higher renter share neighborhoods, and lead to fewer members with business backgrounds. Keywords:American politics; class and ethnicity; race; regional; representation and electoral systems; urban and local politics Across the United States and for the past 40 years, district-based elections have served as a key tool for increasing the presence of otherwise underrepresented racial and ethnic groups in local elected office.A1982amendmenttotheVotingRightsAct(VRA)andasubsequentSupremeCourtdecision (ThornburgvGingles,1986)precipitatedadramaticshifttodistrict-basedlocalelectionsintheSouth; thatinturnledtoanequallydramaticincreaseinBlackrepresentationinlocallegislatures(Davidson, 1994). More recently, the California Voting Rights Act (CVRA)—signed into law in 2002—explicitly stated that at-large elections must not be used “in a manner that impairs the ability of a protected class to elect candidates of its choice,” which has led to a dramatic shift towards local district-based elections in that state; nearly one-third of California cities shifted to district-based elections since its passage. District-based elections are a central feature of local governance in the United States, and they are documented to have a clear impact on racial/ethnic representation in local legislatures. The shift to districts at the local level is largely from multi-member at-large elections, meaning a slate of coun- cil members is elected from one election with the entire city as the electorate (Bullock and Gaddie, 1993). District reforms, like those facilitated by the CVRA, divide the city up into geographically defined areas of similar population size, and the constituents of each district elect a single represen- tative. These elections are majoritarian, winner-take-all, and institute single-member representation © The Author(s), 2026. Published by Cambridge University Press on behalf of EPS Academic Ltd. This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.https://doi.org/10.1017/psrm.2025.10070Downloaded from https://www.cambridge.org/core. Montana State University Library, on 11 Mar 2026 at 17:50:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.30 2 Cushing-Daniels et al. in the city council (Gerber et al.,1998). Districts disperse electoral power throughout the city and ensure representatives are elected from different neighborhoods. Conversely, electoral power in at- large systems favors the median voter, and in these systems it is difficult for minority candidates to win (Gerber et al.,1998; Trounstine,2010). Extant literature on district reforms has focused on race and ethnic minority representation, pri- marily because these communities were central in adopting district policies in the first place. We reviewthisliteraturebelow,inparttounderscoreitscontinuingimportanceindistricting(Hutchings andValentino,2004).Weexpandthequestionofwhetherdistrictsincreaserepresentationforminori- ties to include class; wealthy, white-collar communities are consistently over-represented in terms of who gets elected, at every level of government (Carnes,2013;2020). Therefore, the central question of our paper is,Does the use of at-large vs. district-based elections affect the descriptive representation of less-affluent residents and neighborhoods on municipal councils? District reforms to ensure representation for racial and ethnic minorities may also likely impact otherdimensions,duetothepresenceofcorrelationsbetweenraceandclass.Inadditiontoprioritiz- ing representation of a racial or ethnic minority group, do district reforms disrupt the representation for wealthier individuals and neighborhoods, traditionally robust in at-large city council systems (Schaffner et al.,2020)? To address these questions, we construct a statewide map of all newly drawn city council districts in California in the 2010s—which accounts for council districts in well over 100 cities. We link our map to Census demographics and the residential locations of council candidates. We also draw on a dataset of council candidates and winners where we have, in some cases, via hand-collection of data, attached a variety of characteristics of these individuals, again for candidates and members across the state. With those data in hand, we use a difference-in-difference approach to estimate the causal impacts of district-based elections on the characteristics of council candidates and members, as well as the neighborhoods they are coming from. We operationalize “neighborhoods” as Census block groups, mapped to city council districts in California, for all cities in the state with district elections. Our results show clear impacts of district-based elections on class-based representation. Specifically, we find that council candidates and members are significantly more likely to reside in lower-income and higher renter-share neighborhoods and are less likely to self-report a “business”- relatedoccupation.Whilewealsoreplicatepriorfindingsonracial/ethnic(inthiscase,largely,Latino) representation, our class-based results are true even for (and perhaps especially for) white candi- dates and members. We include measures for percentages of renters and Supplemental Nutrition Assistance Program (SNAP) recipients, and average per capita income levels in the neighborhood. These measures reflect descriptive representation for the neighborhoods in which candidates and members reside. We view our results on socioeconomic representation on council as contributing to critical dis- cussionsonthelackofrepresentationalongthesedimensions—andwhatthatmightimplyforpolicy outcomesandinequality.Einsteinetal.(2022),forinstance,documentdramaticunderrepresentation ofrentersinelectedofficeandpointtothatasapotentialdriveroflongstandinghomeowner-favoring housing and zoning policy. City council members are also systematically wealthier than the popula- tions they govern (Schaffner et al.,2020); while not explicitly in the context of local government, CarnesandLupu(2023)reviewthevarietyofimpactsthatoverrepresentationofthewealthyinoffice may have on political and policy outcomes. Our results point to one institution that has an impact on these fronts: district-based elections. 1.Drawing theoretical predictions from prior work 1.1.Districts, homeownership, and socioeconomic status OverrepresentationofthewealthyisdocumentedineverylevelofAmericangovernment;only10%of citycouncilmembersarefromworking-classoccupationalbackgrounds,higherthanstatelegislators,https://doi.org/10.1017/psrm.2025.10070Downloaded from https://www.cambridge.org/core. Montana State University Library, on 11 Mar 2026 at 17:50:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.31 Political Science Research and Methods 3 members of Congress, governors, Supreme Court justices, or US presidents (Carnes,2020; Kirkland, 2021). Aside from occupation, a traditional indicator of wealth or class is homeowner status. Often extolled as a virtue in politics, homeownership is associated with familial economic prosperity and long-term investment in community (Basolo,2007). Homeownership is associated with external interest group activity in local politics; in cities with high rates of homeownership, there is less interest group activity in local politics (Anzia,2022). This indicates more citizen participation and more responsiveness by local leaders, as homeowners par- ticipatemoreinlocalpoliticsbyvotingandattendingcouncilmeetings(Einsteinetal.,2022).Recent home buyers begin to participate more after purchasing a home, and are activated most by local zon- ing policies (Yoder,2020; Hall and Yoder,2022). At the local level, land use is a dominant area of policymaking; zoning plans prioritize the interests of homeowners over others to protect most peo- ple’s most valuable assets (Trounstine,2018; Einstein et al.,2022). In both political participation and policy responsiveness to their concerns, homeowners are overrepresented at the local level. There is little variation in elected officials’ homeowner or renter status; most are owners (Carnes, 2013; Einstein et al.,2022). City council members do not rent in large numbers in any system, but Einstein et. al find that renters are no less likely to win than candidates who own their homes. However, they do find a gap earlier in the process, at the candidate recruitment stage (Einstein etal.,2022).Thisfindingiscommensuratewithotherscholarshipthatfindsinstitutionalorstructural forces, such as socioeconomic status, are major factors in the decision to run for office at all (Welch andBledsoe,1988;Spiceretal.,2017;Phillips,2021).Childhoodpovertycanhaveindelibledampen- ing effects on people’s confidence in their own qualifications for office when considering candidacy (Levine and Abromowitz,2023). The lasting insecurity left by poverty diminishes the likelihood that many people who experienced poverty as children will run. InasurveyofcandidatesforlocalofficeintheChicagoarea,Oliveretal.(2012)findalackofcon- gruence with the general population. Most candidates are long-term residents of their communities who own their homes, are over 50, typically white men with college degrees (98). The variety is in occupation; most candidates are motivated by a civic duty to public service with a responsibility to thecommunity.Thesepositionsareinthepublicsector,management,andlawindustriesthatcangive would-be candidates experience working with the public and like homeownership, can demonstrate a similar long-term investment in the community (99). Institutional reform that alters the parameters of an election may impact the decision to run for office at the recruitment stage, particularly for candidates with lower-income backgrounds. Due to economic variance in different districts, those who run for office as civic or economic elites are more likely to be diverse than in at-large systems. In this way, districts force diffusion of the centers of local political power. In a model of candidate recruitment, Jacob (1962) suggests that workers from “brokerage”occupationsfromanyclassoughttobeengaged.Brokerageoccupationsincludelawyers, restaurateursandhoteliers,insuranceandcarsalespeople,andlocalunionofficials(709).Workersin these industries have transferable skills for politics: “the ability to convince, the art of inspiring trust and confidence” (710). However, the majority of recruited candidates are white collar professionals. Blue-collar workers are harder to recruit for office because they have a harder time fundraising for campaigns that are not considered “safe bets” to win by political elites (Carnes,2020). In the absence of party influence in nonpartisan systems, other organizations can emerge in can- didate recruitment. For instance, labor unions have a vested interest in developing their members for political office. Where unions are stronger, more workers run and win (Sojourner,2013). District systems can diversify who is recruiting as well. Nationally, the vast majority of city council mem- bers come from a white-collar background (Carnes,2013;2020), but districts have been found to impact council representatives’ income and education level more than nonpartisan elections, local population, median education levels, and racial demographics (Welch and Bledsoe,1988). Theseareallimportantfactors,becausepolicyrepresentinglower-incomevoters’interestsmaybe affected by who is in office. For example, land use policies may be impacted by district reforms if thehttps://doi.org/10.1017/psrm.2025.10070Downloaded from https://www.cambridge.org/core. Montana State University Library, on 11 Mar 2026 at 17:50:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.32 4 Cushing-Daniels et al. council itself becomes more economically diverse. Wealthy, white neighborhoods are disproportion- ately represented in at-large systems (Lineberry,1977; Heilig and Mundt,(1984)). Fundamental to arguments for district reform is that they facilitate a more diverse city council, as a result of smaller costs of running a smaller-scale campaign in a single district rather than an at-large election with a city-wide electorate. Extant literature has found higher campaign expenditures in at-large races, including within hybrid cities featuring both districts and at-large members (Heilig and Mundt, (1984); Hogan and Simpson,2001; Malinowski,2013). Others challenge these conclusions, however, comparing cities to find the size of city and prestige of serving in a larger city’s local government may have more influence over campaign costs. For instance, Adams (2018) found that campaigns for office in small cities bring down absolute and per-voter costs, since holding public office in smaller cities is viewed as less prestigious, which makes candidate background more important. Candidate background, like occupation and socioeconomic status, may matter most in nonpartisan systems as well, in lieu of a party identification heuristic for voters (Krebs,2001). Given the uniformity of non- partisan elections in California,we hypothesize that district reforms allow for more economically diverse city government.1 1.2.District-based elections and racial/ethnic representation Extant literature on descriptive representation primarily considers dimensions of race and ethnicity, andforgoodreason,giventhecentralityofraceandethnicityinpoliciesadoptingdistrict-basedelec- tions, and the construction of districts themselves. Early in the 20th century, wards or districts as the vehicleforracialandethnicminorities’representationwasamongthemainargumentsbyProgressive Era reformers supporting at-large systems (Banfield and Wilson,1963; Trounstine,2008). Later, fol- lowingthepassageoftheVRAin1965(andespeciallyfollowinganamendmenttotheVRAin1982), many local jurisdictions in the South were court-ordered to adopt districts to address the barriers Black voters faced in electing their preferred candidates; others did so preemptively in anticipation of legal action (Davidson,1994). District-based elections increased descriptive representation for African Americans, a consistent finding owing to relative consistency in geographic concentration of Black communities, an ongo- ing legacy of segregation (Karnig,1976; Robinson and Dye,1978; Engstrom and McDonald,1981; Engstrom,1987; Welch,1990; Trounstine,2008). Districts make descriptive representation for these communitiesmorelikely,sincecity-wideturnoutandanappealtoawhitemajorityarenotnecessary to elect a council member (Trounstine and Valdini,2008; Hajnal,2010). Black communities winning city council seats through districts was particularly strong in the South, where they were prohibited from voting en masse until the VRA’s passage in 1965 (Davidson,1994; Shah et al.,2013). Latinosarethelargestminorityinthecountry,andthelargestracialorethnicgroupinCalifornia, where they comprise about 40% of the total population (Lopez,2014; Funk and Lopez,2022). Population growth and the CVRA have inspired recent studies on descriptive representation explor- ing the fortunes of Latino communities in local and state institutions. Collingwood and Long (2021) found CVRA-era district reforms in California improved minority representation in high-density Latino cities by more than 20%. Around the country, Latino communities have benefited from dis- trictswhendenselyconcentratedwithindistrictboundaries,aswellasfromlargercouncilswithmore members, and as Latino candidates may drive mobilization (Taebel,1978; Bullock and MacManus, 1990; Jones-Correa,1998; Sass,2000; Barreto,2007). Studies examining California school board elections find similar results. Abott and Magazinnik (2020) use the CVRA to find an increase in one additional Latino school board representative for every three elected seats following district reforms in areas with high Latino populations and a high degree of residential segregation. Leal et al. (2004) find at-large elections detrimental for Latino 1In the online appendix, we further overview relevant work on intersections of class and race.https://doi.org/10.1017/psrm.2025.10070Downloaded from https://www.cambridge.org/core. Montana State University Library, on 11 Mar 2026 at 17:50:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.33 Political Science Research and Methods 5 representation on California school boards, especially in areas where they constitute a minority. In both city councils and school boards, district reforms have resulted in increased representation for Latino populations, largely with high numbers and in concentrated communities, both consistent with traditional findings. 2.Institutional details 2.1.California city councils All city council elections in California are nonpartisan. Part of the Progressive Era reform move- ment that resulted in at-large city councils being the norm, nonpartisan elections were intended to support a motivated and informed public outside of the divisiveness in party politics (Macmanus, 1978;Schaffneretal.,2001).Nonpartisanelectionsaffectvoters’considerationsbyremovingapromi- nent heuristic in vote choice: party identification (Kirkland and Coppock,2018). Effects on minority representation in nonpartisan elections are inconclusive; some find depressed turnout overall and strengthened incumbent advantage, which can both reinforce the status quo (Schaffner et al.,2001). Others find that emphasis on candidates’ professional and political experience may be used to com- pensate for a lack of party heuristic, but does not translate to an automatic party advantage for candidates(Schaffneretal.,2007;KirklandandCoppock,2018).Wedonotconsiderthepartisanship ofcandidatesortheelectorate,becausethereisnochangeorvariationinthenonpartisanelectiontype in the state. However, nonpartisan elections contribute to the political environment in which district reforms take place. Californiacouncilmembersservefour-yearterms,electedviastaggeredelectionseverytwoyears.2 Most councils have five members; with the exception of the largest cities, deviations from the five- member structure still result in relatively small groups of seven- or nine-members. Weidentify—forallcitiesinCalifornia—whetherandwhentheyswitchedfromat-largetodistrict- based elections. We find that by 2011, only 40 cities (roughly 8%) of cities were using district-based elections.Asoutlinedabove,thischangeddramaticallyduringthe2010s:inourdataonallCalifornia cities, by 2020, roughly 150 cities (31%) use district-based elections. All but five of the cities that shifted to district-based elections after 2011 did so in the period from 2016 to 2020. We document which cities were always at-large during our sample period, always districted during our sample period, or switched from at-large to district-based in Figure 1. A primary takeaway is that the switching cities are relatively spread throughout the state. 3.Map of cities shifting to district-based elections 3.1.California Voting Rights Act To study the impacts of district-based elections on multiple dimensions of council composition, we leverage shifts to district-based elections stemming from the CVRA of 2002. The Act specified that “Anat-largemethodofelectionmaynotbeimposedorappliedinamannerthatimpairstheabilityof a protected class to elect candidates of its choice or its ability to influence the outcome of an election, as a result of the dilution or the abridgment of the rights of voters who are members of a protected class,” with protected class defined as “voters who are members of a race, color or language minority group.” As noted, at the time (and into the early 2010s), with the exception of the largest cities, most cities in California employed at-large elections. We direct readers to Abott and Magazinnik (2020) and Hankinson and Magazinnik (2023) for detailed discussions of the history of the CVRA and why the largest share of shifts to district-based elections occurred in the 2010s, a decade after the passage of the Act; in short, a major piece of the explanation for the gap in timing is that organizing groups 2For example, a five-seat council may fill seats 1 and 2 in yeart, seats 3, 4, and 5 in year t+2, and then hold an election for seats 1 and 2 again in yeart+4.https://doi.org/10.1017/psrm.2025.10070Downloaded from https://www.cambridge.org/core. Montana State University Library, on 11 Mar 2026 at 17:50:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.34 6 Cushing-Daniels et al. Figure 1.Cities by at-large vs. district-based elections during the sample period (2008–2020). Note: “Switched” refers to cities that switched from at-large to district-based city council elections between 2008 and 2020. whoinitiatedpotentiallegalactionundertheActfaced“internalcapacityconstraintsandcompeting priorities” (Hankinson and Magazinnik,2023). 4.Data We construct a unique database combining information from the California Elections Data Archive (CEDA), the US Census, the American Community Survey (ACS), the National Association of Latino Elected and Appointed Officials (NALEO), the California Voter Registration Database, and hand-collectedinformationoncandidatecharacteristics.Usingthisdataset,weanalyzeblock-group- level demographic data and individual-level candidate data to estimate the impact of switching to district-based elections on descriptive representation in 118 California cities. We outline the data construction in greater detail in the remainder of this section. 4.1.District maps and characteristics Most critical to our analysis is the mapping of Census block groups to city council districts for all cities in California with district-based council elections. Such a mapping allows us to compile and aggregate to district-level a variety of block-group-level demographic, socioeconomic, and partisan characteristics of residents. To do so, we begin from voter file data from 2020, which reports voters’https://doi.org/10.1017/psrm.2025.10070Downloaded from https://www.cambridge.org/core. Montana State University Library, on 11 Mar 2026 at 17:50:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.35 Political Science Research and Methods 7 Table 1.Average district characteristics, split by racial composition Maj. White (CVAP) Not Maj. White Maj. White Total N 208.00 (38.95%) 326.00 (61.05%) 534.00 (100.00%)Pop. (1000s) 19.12 (13.39) 16.11 (10.56) 17.28 (11.82)Share AAPI (CVAP) 0.15 (0.15) 0.08 (0.06) 0.11 (0.11)Share Black (CVAP) 0.08 (0.08) 0.03 (0.03) 0.05 (0.06)Share White (CVAP) 0.35 (0.10) 0.69 (0.11) 0.56 (0.20)Share Hisp. (CVAP) 0.40 (0.17) 0.18 (0.09) 0.27 (0.16) Share Other Race/Eth. (CVAP) 0.01 (0.01) 0.01 (0.01) 0.01 (0.01) Share Renter Occ. Housing 0.41 (0.15) 0.33 (0.14) 0.36 (0.15) Share Inc>2x Pov. 0.59 (0.17) 0.72 (0.13) 0.67 (0.16) addresses—which in turn are geocoded to Census block groups—and also the city council district they are assigned to as of 2020. This then immediately facilitates a mapping of block groups to dis- tricts.Someshare(29%) ofblockgroupsaresplitacrossdistricts. However,amongstthesesplitblock groups,mostareprimarilycontainedwithinasingledistrict.Wedropanyblock-group–districtpair- ings where less than 50% of the block-group population is contained within the district in question. This generates a unique block-group–district mapping. Amongst these, for each block group in the resulting dataset, the average share of the population contained within the district that we assign it to is 96%.3 We then merge a variety of block-group variables into this dataset and aggregate them to the district-level to facilitate our analysis of the composition of districts. We draw especially on block- group-leveldemographicandsocioeconomiccharacteristics,whicharedrawnfromthe2010Census and the five-year compilation of the 2008–2012 ACSs. Specifically, we draw in block-group-level counts of population and households by income level, renter/owner status, and receipt of social assistance. We draw on the Census’s special tabulation of citizen voting age population (CVAP) by race/ethnicityasasourceforblock-grouprace/ethniccomposition.Thatalsodrawsondatacollected from the 2008 to 2012 ACSs. Aggregating all of these characteristics to the city council district-level provides a snapshot of dis- trict composition in the 2010s, when a vast majority of cities that switched to district-based elections didso.Thesecharacteristicsatthedistrict-levelaresummarizedin Table1.Giventhattheaimsofthe CVRAandshiftstodistrict-basedelectionswerefocusedonracialandethnicrepresentation,wesplit summarystatisticsbydistrictracecomposition(majoritywhitevs.notmajoritywhite)inColumns1 and2,withfullsampleaveragesinColumn3.Doingsopreviewsthecorrelationsbetweenracial/eth- niccompositionofdistrictsandothercharacteristics(income,partisancomposition,etc.),whichmay in turn imply that these shifts have broader impacts on representation. 4.2.Candidate characteristics Our main results center around how district-based elections change who runs for council and who is elected. This subsection describes how we construct our dataset on candidate characteristics. The coredatasetusedistheCEDA,whichreportstheuniverseofcandidatesforlocalofficesinCalifornia from 1995 to 2020. We use data from city council elections from 2008 to 2020. We removed cities that held district-based elections throughout the entire period, as they did not experience a system change. Our sample therefore consists of candidates in cities that switch to district-based elections between2008and2020,aswellas(forthesakeofacleancontrolgroupinthedifference-in-difference analysis) cities that are at-large the entire time. There are 13,688 candidates in our resulting subset of the CEDA data. 3Additional notes on this process can be found in the supplemental appendix.https://doi.org/10.1017/psrm.2025.10070Downloaded from https://www.cambridge.org/core. Montana State University Library, on 11 Mar 2026 at 17:50:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.36 8 Cushing-Daniels et al. CEDA data list candidates’ names, office contested, voteshare, and victory or loss. The data also note the ballot designation that appeared on the ballot; this is often the candidate’s occupation. We take advantage of the ballot designation to provide at least one measure of candidates’ professional background. The largest share of candidates indicates that their occupation is “businessperson” or somethingrelated,soinouranalysis,wecomparepeoplewithbusinessorfinance-relatedoccupations to all others. Specifically, we tag a candidate as having a business or finance-related occupation if any of the following strings appear in their ballot designation:business, owner, entrep, executive, investor, capitalist, banking, finance, accounting, accountant. This categorization is similar in spirit to the work of Carnes (2013), categorizing members of legislatures as being in “profit-oriented professions.” Then, because candidates who have won in the past and are incumbents often simply list “incum- bent” or “city councilmember” as their ballot designation, we tag individual candidates as having a business/finance-related occupation if they ever list any of the above strings across the multiple years that we see them in the data. Most typically, this will be drawn from their first appearance as a candidate. With this approach, roughly 28% of candidates are tagged as having a business/finance- related occupation. Appendix Table A3 lists all categories that we coded, some examples of ballot designations from each category, and the relative frequency of each category.4 CEDA data report only the information that appears on the ballot and the results of the election; it therefore does not report candidates’ demographic characteristics. For candidate race and gender, wedrawondatafromBeachandJones(2017)andBeachetal.(2024).Theonlineappendixdescribes their data construction process in more detail; we extend their data to include additional years using the same process. We gather additional information about candidates by matching candidate observations in the CEDA data to the 2016 California Voter Registration Files, matching by name and city. These data provide candidates’ home address. The home address is then geocoded to Census block groups and matched to Census data to provide some information on characteristics of the neighborhood that a candidate lives in (e.g., income, share of renters, etc.).5 4.3.Date of adoption of district-based elections Finally, we also use the California Election Data Archive to identify the first election year that a city used district-based elections. In elections where cities use district-based elections, the district that a candidate is running within is listed. In cities with at-large elections, this variable is simply blank. As such, we identify the first election that a city used district-based elections by identifying the first year in the data for a city where the district variable is not blank. Appendix Table A2 provides city-level summary statistics for cities that never use district-based elections during our sample period (which form our control group,n=322), cities that switch to district-based elections during our period (which form our treatment group,n=116), and cities that hold district-based elections during our entire sample period (which we ultimately omit from our analysis,n=32).6 A key takeaway is that cities that form our treatment group are quite similar to cities that form our control group on most dimensions other than population. The cities that adopt district-based elections are larger. The fact that those cities are larger is not surprising; as Hankinson and Magazinnik (2023) note, organizers actively targeted their efforts towards higher population cities. The cities that already used district-based elections prior to 2008 are different on a num- ber of dimensions than the other two categories of cities; they are much larger (even dropping Los AngelesandSanDiego)andhavehigherracial/ethnicdiversity.AppendixFiguresA10,A11,andA12 4We also adopt an alternate means of coding candidates’ occupations, described in more detail in the online appendix. 5Toassesswhetherthereisanysystematicrelationshipbetweencandidatesmatchedvs.notmatched,theAppendixprovides two figures. See further discussion in the online appendix. 6The table omits the two largest population cities in the sample, Los Angeles and San Diego, which would otherwise fall in the always-districted group.https://doi.org/10.1017/psrm.2025.10070Downloaded from https://www.cambridge.org/core. Montana State University Library, on 11 Mar 2026 at 17:50:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.37 Political Science Research and Methods 9 graphically depict averages of key outcome variables (measured at city-by-year level) across years and for each of the three categories of cities noted above (switching cities, never districted, always districted). 5.Empirical approach: difference-in-differences We employ a difference-in-difference analysis of how district-based elections impact who runs and whoisultimatelyelectedtothecouncil.Thesespecificationswilladdresshowandwhetherthepoten- tial or “expected” shifts in representation captured in the previous section in fact translate into shifts in who is elected to the council. Our main specifications are estimated at the city-by-election year level, where we take as outcomes the share of candidates or share of winning candidates with some characteristic (e.g., percent white) in that city-by-election year.7 We estimate: (Winning) Candidate Char. Sharect = 𝛽 ×Post-Districtingct + 𝛿t + 𝛾c + 𝜖ct (1) Post-Districting isabinaryvariableequaltooneincities(c)andyears(t)thatareusingdistrict-based elections and zero for cities that have not yet switched to district-based elections or never do dur- ing our sample period. We exclude cities that were district-based for the entire sample period,8 but include all other California cities. Thus, cities that switch to district-based between 2008 and 2020 serve as the “treatment” group and cities that use at-large elections the entire time serve as the “con- trol” group. We include city and year fixed effects to difference out time-invariant city characteristics and across-city year effects.𝛽identifies the causal impact of switching to district-based elections. A critical assumption inherent in difference-in-difference analysis is that treated and control units would follow parallel paths in the absence of treatment. To provide some evidence on this, we also conduct event studies to ensure that there are no differences in trends in the years leading up to treatment (Figure 4). 6.Results 6.1.Main difference-in-difference results Figure2presentsresultsfromthedifference-in-differenceanalysisoncandidatecharacteristicsaggre- gated to the city-by-year level. (Our next figure will provide evidence on the types of neighborhoods candidates and winners come from.) Each point (with confidence intervals) in the figure is from a separate regression, matching Eqn. 1 above, but taking on a variety of outcome variables. The fig- ure plots both the shares of all candidates (red dots) and the shares of winning candidates (blue dots) in a given city-year who are of a particular characteristic, listed along the left-hand side of the figure. Even though race and ethnicity is not the central focus of this paper, we describe results on that front first, both to highlight that we observe similar results to prior literature and to orient readers to our figure. Consider the row labeled “Race/Eth: White.” The pair of dots there reports the impacts of district-based elections on the share of candidates (red dot) and winners (blue dot) who are, to use Census-styleracial-ethniccategories,non-Hispanicwhite.Wefindthatthereisasignificantdecrease both in the share of candidates and the share of winners who are white. As in other work on the CVRA, district-based elections generate increases in Hispanic representation; in our paper, we are able to document that a large part of why this happens (as depicted in the figure) is an increase in 7In calculating shares, the denominator is the set of candidates or winners in that city-year that we have matched for the outcome in question. For instance, if we identify the race/ethnicity of two of the three total candidates, the denominator in calculating the race/ethnic share of candidates is two. 8A key reason for doing so is that recent work documents bias introduced by using already-treated units as control units in a difference-in-difference analysis (Goodman-Bacon,2021).https://doi.org/10.1017/psrm.2025.10070Downloaded from https://www.cambridge.org/core. Montana State University Library, on 11 Mar 2026 at 17:50:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.38 10 Cushing-Daniels et al. Figure 2.Difference-in-difference analysis: how individual characteristics of candidates and winners change under district- based elections. Note: Each point in the figure is from a distinct regression. The regression plots the “Post-Districting” coefficient from the difference-in- difference specification. Also displayed are 90% and 95% confidence intervals, with standard errors clustered at the city level. Hispaniccandidateentry,astheimpactofdistrictsontheshareofcandidatesandwinnersisrelatively similar. Next, we turn to evidence in the figure on other dimensions of representation: especially class. The bottom pair of points in the figure reports the impacts of district-based elections on the profes- sional background of candidates and winners. While we cannot observe candidates’ individual-level incomes or socioeconomic status, we do observe their occupation as self-reported on the bal- lot. We find that the share of both winners and candidates who are in a business/finance-related occupation decreases, providing evidence that district-based elections shift councils away from overrepresentation of individuals from white-collar occupations.9 Wefurtherexplorehowdistrict-basedelectionsimpactclassandsocioeconomicstatus-basedrep- resentation in our analysis on the types of neighborhoods that candidates come from, which we turn tonext.Figure3 presentsresultsonthatsetofoutcomes.Specifically,wetakeasoutcomesthecharac- teristicsoftheCensusblockgroupthatcandidatesandwinnerscomefrom.Again,tostartwithresults that link us to much of the existing literature and orient readers to the figure, the first two pairs of estimates report results on the racial/ethnic composition of Census block-group candidates live in. The results reveal that the average winner comes from a neighborhood that is roughly 4 percentage points less (non-Hisp.) white and 4 percentage points more Hispanic. Results are similar, but more muted on neighborhood characteristics for all candidates. The estimation sample standard deviation of the winning candidates’ neighborhood Hispanic population share is 0.24, so the magnitude of our effect is roughly one-sixth of a standard deviation. The remainder of the figure contains results that we view as our primary contribution. The pair of resultsintherowlabel“%Renter”capturestherentershareofcandidates’andwinners’Censusblock 9Appendix Figures A1 and A2 provide more evidence on occupational background.https://doi.org/10.1017/psrm.2025.10070Downloaded from https://www.cambridge.org/core. Montana State University Library, on 11 Mar 2026 at 17:50:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.39 Political Science Research and Methods 11 Figure 3.Difference-in-difference analysis: how neighborhood characteristics of candidates and winners change under district-based elections. Note: Each point in the figure is from a distinct regression. The regression plots the “Post-Districting” coefficient from the difference-in- difference specification. Also displayed are 90% and 95% confidence intervals, with standard errors clustered at the city level. groups. While we cannot observe whether candidates/winners are themselves renters or homeown- ers, we do find that district-based elections are associated with increases in candidates/winners from higherrenter-shareneighborhoods;withtheaveragecandidateandwinnercomingfromaneighbor- hoodthatis3percentagepointshigherinrentershareunderdistrict-basedvs.at-largeelections.The estimation sample standard deviation of winning candidates’ neighborhood renter share is 0.17, so the magnitude of our effect is roughly 0.18 of a standard deviation of the outcome. Thenexttwopairsofresultsreportimpactsonthetypesofneighborhoodscandidatesandwinners are coming from with respect to income composition. The first pair of points takes as an outcome the average (logged) per capita income of candidates/winners neighborhoods. The second pair of points takes as an outcome the share of block-group residents who are SNAP recipients; because SNAP eligibility is income-based, we draw on this outcome largely as another measure of income distribution of neighborhoods, with this measure capturing the lower end of the distribution. The high-level result on both outcomes is that we observe a shift towards candidates and win- ners from lower-income neighborhoods. For instance, the average winning candidate comes from a block group with income that is roughly 6% lower. But here, we observe more sizable differences betweentheestimatesforcandidatesandwinnersthanmanyofourotheroutcomes—withmorepro- nounced impacts for winners. This indicates that while district-based elections have some impact on the entry of candidates from lower-income neighborhoods, a large part of the increased presence of these candidates on the council is from the increased likelihood of winning conditional on running. This is consistent with our predictions, which highlighted that district-based elections may decrease the cost of participating in a contest and therefore increase the likelihood of winning for individuals from a broader range of the income and wealth distribution. 6.2.Robustness of main results As noted in the description of our methods, we conduct event study analyses to assess the parallel trends assumption that is critical to a difference-in-difference analysis. Event studies from some ofhttps://doi.org/10.1017/psrm.2025.10070Downloaded from https://www.cambridge.org/core. Montana State University Library, on 11 Mar 2026 at 17:50:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.40 12 Cushing-Daniels et al. Figure 4.Event study: district-based elections and share of Hispanic winning candidates. Note: Each panel in the figure is from a distinct event study regression. Estimates are depicted by the solid line. Also reported are 90% (dark shaded area) and 95% (light shaded area) confidence intervals, with standard errors clustered at the city level. ourkeyoutcomevariablesaredepictedin Figure4.Thefiguresplotdifferencesbetweentreatmentand controlgroupsatdifferentpointsintimerelativetotheyearsimmediatelybeforetheimplementation of district-based elections. If there were confounding time-varying factors that relate both to our outcome but also to the adoption of district-based elections, this may materialize as trends in our outcomes in the years immediately prior to adoption. Instead, in all figures, there is no difference between treatment and control in the years prior to treatment—indicating parallel trends in the pre- periods. Additional robustness tests are presented and discussed in the online appendix. 6.3.Drivers of main results Next, we provide some additional evidence aimed at better understanding the mechanisms driving our main results. Before doing so, it is worth taking stock of our current results and the mechanisms that may be driving them. We have found that district-based elections are causally associated with an increase in shares of: Latino candidates and winners, candidates and winners from higher renter- shareneighborhoods,andcandidatesandwinnersfromlowerincomeneighborhoods.District-based elections, likewise, are causally associated with decreases in shares of: white candidates and winners andcandidatesandwinnersfromabusiness/finance-relatedoccupation.Ofcourse,therace/ethnicity resultssummarizedherearelargelyinlinewithexistingliterature;thechangesinclass-basedcompo- sition of candidates and winners are our primary focus. The fact that we observe changes in both the composition of candidates and winners already provides some evidence on the mechanism driving https://doi.org/10.1017/psrm.2025.10070Downloaded from https://www.cambridge.org/core. Montana State University Library, on 11 Mar 2026 at 17:50:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.41 Political Science Research and Methods 13 our results: results are at least partially driven by candidate entry and not strictly a change in candi- datesuccessconditionalonentering.Wedonote,however,thatinmostofourmainresults,theeffect of district-based elections is at least marginally larger when focusing on the composition of winners relative to the composition of candidates. If candidate entry was the sole driver of changes in council composition, the change in the composition of winners would roughly equal the change in the com- position of candidates. Our finding on this front parallels Abott and Magazinnik (2020), who find that district-based elections increased both the Latino share of school board candidates and school board members in California following the local adoption of district-based school board elections. Next, in the context of race/ethnic representation, many have argued that we should expect the strongest impacts of district-based elections on the composition of the elected body in relatively racially/ethnically segregated localities; Abott and Magazinnik (2020) and Trounstine and Valdini (2008) provide empirical evidence in support of this claim. We ask whether the same is true of the class-based representation that we document: are changes in class-based representation on council contingent on the degree of economic segregation in cities? Addressingthatquestioninoursettingcanhelpfurtherclarifypotentialmechanismsatplay.One mechanism that might drive our results is that district-based elections reduce the cost of campaign- ing or impact organizing efforts. If those campaign-focused channels were the primary mechanisms driving our results, results should not differ across more and less economically segregated cities. On the other hand, under the assumption that members of an underrepresented group are more likely to vote for own-group members, absent changes in campaigning or organizing, district-based elec- tions should only be expected to change the composition of the elected body in a city where the group members in question are spatially concentrated (Trounstine and Valdini,2008). Of course, the most relevant mechanisms impacting class-based representation may differ from those impacting racial/ethnic representation. Totesttheroleofeconomicsegregationinoursetting,wefirstconstructameasureofhowresiden- tially segregated higher-income individuals are. We construct a block-group measure of the share of residentswithinablockgroupwhoseincomesexceedtwo-timesthepovertyline;weusethatshareas our measure of higher-income individuals. The remainder in each block group is the share of lower- income individuals. We then use these data to construct a high-income/low-income dissimilarity index at the city level, an index of segregation frequently used to measure local racial/ethnic segre- gation. The index runs from 0 to 1, with zero indicating the maximum degree of integration and 1 indicating the maximum degree of segregation of higher and lower income individuals. We split our sampleofcitiesatthemedianvaluetodefinemoreandlessincome-segregatedcities.Wethenre-run our main specification with distinct “treatment” variables for our more and less income-segregated cities. Results are presented graphically in Figure 5. That top panel reports difference-in-difference coef- ficients for more income segregated cities; the bottom panel reports coefficients for less income segregated cities. As before, each row or pair of coefficients corresponds to a particular outcome. The directional pattern of our results is similar across the top and bottom panels. In fact, for any out- come in the figure, we are not able to reject the hypothesis that the effect of district-based elections is the same in more relative to less income segregated cities. These results do not occur just in newly districted cities with high levels of segregation, which points to the reforms as a driving force. The smaller scale of economy within districts raises the potential for a more diverse slate of candidates for seats in underrepresented neighborhoods, including non-business occupations. Next, in part because some of the factors assessed in our main analyses are correlated with one another, we test how district-based elections impact the composition of the council with regards to intersections between race/ethnic groupings (in particular, non-Hispanic white and Hispanic) and the other characteristics assessed. To do so, we run the same difference-in-difference specification usedelsewhereinthissection,buttakeasoutcomesintersectionalgroups—e.g.,shareofwinningcan- didates who reside in a Higher Income Neighborhood. Results are depicted in Figure 6 and uncover https://doi.org/10.1017/psrm.2025.10070Downloaded from https://www.cambridge.org/core. Montana State University Library, on 11 Mar 2026 at 17:50:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.42 14 Cushing-Daniels et al.Post X High BG Inc. Seg.Post X Low BG Inc. Seg.Race/Eth.: White Race/Eth.: Hisp. Bus./Fin. Occ. N'hood: % Renter N'hood: ln(PC Inc.) Race/Eth.: White Race/Eth.: Hisp. Bus./Fin. Occ. N'hood: % Renter N'hood: ln(PC Inc.) -.2 -.1 0 .1 .2 DiD Estimates - Impact of Districts All Cands. Winners Figure 5.Difference-in-difference analysis: distinct effects of district-based elections in more and less income segregated cities. Note: Each point in the figure is from a distinct regression. The regression plots the “Post-Districting” coefficient from the difference-in- differencespecification.Alsodisplayedare90%and95%confidenceintervals,withstandarderrorsclusteredatthecitylevel.“High(Low) BGInc.Seg.”referstocitiesthatareabove(below)medianinthedissimilarityindex,whichcapturesresidentialsegregationofhigherand lower income individuals. heterogeneity in the impacts of district-based elections that were concealed by the prior analyses. For example, the “Business Occ.” row reports the effect of treatment on the share of candidates who hold Business Occupations and are Hispanic or White (respectively). In this exercise, “High Income Neighborhood”and“HigherRenterNeighborhood”aredefinedasthoseblockgroupsthatfallabove the sample median in those measures. Most notably, we find that two of our main outcomes documenting shifts in class-based represen- tation (share of candidates from “business” occupations and incomes of neighborhoods candidates come from)seem tobe drivenbywhitecandidates in particular.Wefind thatdistrict-based elections impact racial/ethnic representation and separately impact class-based representation, a departure from the expectation that district-based elections may impact the race or ethnicity of candidates and winners, which would impact other intersecting characteristics. 7.Conclusion This paper has assessed the impacts of shifting from at-large to district-based elections for all 118 city councils in California that did so between 2008 and 2020. In doing so, it has been our aim to shed light on the broad impacts of district-based elections in local elections, both in who runs and wins elections, and how districts are drawn. While a sizable body of work has studied the impacts of district-based elections on descriptive racial and ethnic representation (Davidson,1994; Abott and Magazinnik,2020), there is less known in the modern context and on a wide scale about the two main focuses of this paper: the characteristics of candidates and elected councilmembers, and the far-reaching impacts of how the districts themselves are drawn. We aim to address these questions, focusing in particular on class-based representation. The rapid shift from at-large to district-based https://doi.org/10.1017/psrm.2025.10070Downloaded from https://www.cambridge.org/core. Montana State University Library, on 11 Mar 2026 at 17:50:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.43 Political Science Research and Methods 15 Figure 6.Difference-in-difference analysis: winning candidate characteristics interacted with race/eth. Note: Each point in the figure is from a distinct regression. The regression plots the “Post-Districting” coefficient from the difference- in-difference specification. Also displayed are 90% and 95% confidence intervals, with standard errors clustered at the city level. All regressions are at the city-by-year level. elections among city councils in California in the 2010s and our statewide district mapping efforts facilitate our contributions on these fronts. To briefly summarize our findings: we find increases in the representation of individuals from lower-income neighborhoods and neighborhoods with a higher share of renters. We also find a decrease in winners with “business”-related occupations. These findings are important in light of the dramatic underrepresentation of working-class individuals and renters in local government (Schaffner et al.,2020; Einstein et al.,2022) and also in light of some recent papers that document a shiftinhousingpolicyunderdistrict-basedelections(Mast,2024;HankinsonandMagazinnik,2023). Beyond the shift in incentives faced by councilmembers elected from districts, our results highlight that changes in housing policy (and likely other policy) may also be a consequence of who serves on the council under district-based elections, in line with evidence that working-class representation in Congress impacts substantive representation (Carnes,2012). Our results also contribute to enduring questions: which groups get represented, and by whom? The myriad implications of these two essential roles in democratic representation have animated decades of research. District reforms in city councils have the potential to increase substantive rep- resentation for minorities long excluded, namely, people of color and lower classes. While districts are able to remedy underrepresentation for these groups, they are most beneficial for geographi- cally concentrated populations, a direct result of racial and economic segregation. As cities continue to reform electoral institutions to reflect minority communities, we expect district representation to becomemoreubiquitous,andmunicipalpolicytomorecloselyreflectthepreferencesoftheseminor- ity groups. Future research is needed to address more substantive benefits of district reform in the policyareasmostimportanttocitycouncils,suchasdevelopment,parks,policeandfire,andutilities. Supplementary material.The supplementary material for this article can be found at https://doi.org/10.1017/psrm.2025. 10070. To obtain replication material for this article, please visit https://doi.org/10.7910/DVN/NXTJMD.https://doi.org/10.1017/psrm.2025.10070Downloaded from https://www.cambridge.org/core. Montana State University Library, on 11 Mar 2026 at 17:50:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.44 16 Cushing-Daniels et al. References Abott C and Magazinnik A (2020) At-Large Elections and Minority Representation in Local Government.American Journal of Political Science 64(3), (July) 717–733. AdamsB (2018)Campaigninginlilliput:Money’sinfluenceinsmallandmid-sizedcityelections.CaliforniaJournalofPolitics and Policy 10(2), 1–17. Anzia SF (2022)Local interests: Politics, policy, and Interest Groups in US City governments.University of Chicago Press. Banfield EC and Wilson JQ (1963)City Politics.Vantage Books, Random House. Barreto MA (2007) Si Se Puede! Latino Candidates and the Mobilization of Latino Voters.The American Political Science Review 101(3), 425–441. Basolo V (2007) Explaining the support for homeownership policy in us cities: A political economy perspective.Housing Studies 22(1), 99–119. BeachBandJonesDB (2017)Gridlock:Ethnicdiversityingovernmentandtheprovisionofpublicgoods.AmericanEconomic Journal: Economic Policy 9(1), 112–136. BeachB,JonesDB,TwinamTandWalshR(2024)Racialandethnicrepresentationinlocalgovernment.AmericanEconomic Journal: Economic Policy 16(2), 1–36 Bullock CS and Gaddie RK (1993) Changing from Multimember to Single-Member Districts: Partisan, Racial, and Gender Consequences.State and Local Government Review 25(3), 155–163. Bullock CS and MacManus SA (1990) Structural features of municipalities and the incidence of Hispanic councilmembers. Social Science Quarterly 71(4), 665–681. CarnesN(2012)Doesthenumericalunderrepresentationoftheworkingclassincongressmatter?.LegislativeStudiesquarterly 37(1), 5–34. Carnes N (2013)White-Collar government: the Hidden Role of Class in Economic Policy making. Chicago Studies in American politics. Chicago: The University of Chicago Press. Carnes N (2020)The Cash ceiling: why Only The Rich run for office–and What we can do About it.Princeton University Press. Carnes N and Lupu N(2023) The economic backgrounds of politicians.Annual Review of Political Science 26(1), 253–270. Collingwood L and Long S (2021) Can states promote minority representation? assessing the effects of the california voting rights act.Urban Affairs Review 57(3), 731–762. Davidson C(1994)Quiet Revolution in the South: The Impact of the Voting Rights Act, 1965-1990.Princeton University Press. Einstein KL, Ornstein JT and Palmer M (2022) Who Represents the Renters?Housing Policy Debate 33(6), 1554–1568. https://doi.org/10.1080/10511482.2022.2109710.https://www.tandfonline.com/doi/full/10.1080/10511482. EngstromRL (1987)DistrictmagnitudesandtheelectionofwomentotheIrishDáil.ElectoralStudies 6(2),(August)123–132. (accessed 24 April 2023).https://doi.org/10.1016/0261-3794(87)90019-9.https://linkinghub.elsevier.com/retrieve/pii/ 0261379487900199. Engstrom RL and McDonald MD (1981) The election of blacks to city councils: Clarifying the impact of electoral arrange- ments on the seats/population relationship.American Political Science Review 75(2), 344–354. Funk,C.M.H.Lopez(2022).Abriefstatisticalportraitofu.s.hispanics.Technicalreport,PewResearchCenter,Washington, D.C.https://www.pewresearch.org/science/2022/06/14/a-brief-statistical-portrait-of-u-s-hispanics/. Gerber ER, Morton RB and Rietz TA (1998) Minority representation in multimember districts.American Political Science Review 92(1), 127–144. Goodman-Bacon A (2021) Difference-in-differences with variation in treatment timing.Journal of Econometrics 225(2), 254–277. Hajnal Z (2010)America’s Uneven democracy: race, turnout, and Representation in City politics.Cambridge, New York: Cambridge University Press. OCLC: ocn318672982. HallABandYoderJ(2022)Doeshomeownershipinfluencepoliticalbehavior?evidencefromadministrativedata.TheJournal of Politics 84(1), 351–366. HankinsonMandMagazinnikA(2023)Thesupply-equitytrade-off:Theeffectofspatialrepresentationonthelocalhousing supply.The Journal of Politics 85(3), 1033–1047. HeiligPandRJMundt (1984)Yourvoiceatcityhall:thepolitics,procedures,andpoliciesofdistrictrepresentation.SUNYSeries in Urban Public policy.State University of New York Press, 3–8. Hogan S and Simpson D (2001) Campaign contributions and mayoral/aldermanic relationships: Building on krebs and pelissero.Urban Affairs Review 37(1), 85–95. Hutchings VL and Valentino NA (2004) The centrality of race in american politics.Annu. Rev. Polit. Sci.7(1), 383–408. Jacob H(1962) Initial recruitment of elected officials in the us–a model.The Journal of Politics24(4), 703–716. Jones-Correa M (1998)Between two nations: the Political Predicament of Latinos in New York City.Ithaca: Cornell University Press. KarnigAK (1976)BlackRepresentationonCityCouncils:TheImpactofDistrictElectionsandSocioeconomicFactors.Urban Affairs Quartely 12(2), (December) 223–242. Kirkland PA and Coppock A(2018) Candidate choice without party labels: New insights from conjoint survey experiments. Political Behavior 40, 571–591.https://doi.org/10.1017/psrm.2025.10070Downloaded from https://www.cambridge.org/core. Montana State University Library, on 11 Mar 2026 at 17:50:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.45 Political Science Research and Methods 17 Kirkland PA (2021) Business owners and executives as politicians: the effect on public policy.The Journal of Politics 83(4), 1652–1668. Krebs TB (2001) Political experience and fundraising in city council elections.Social Science quarterly 82(3), 536–551. Leal DL, Martinez-Ebers V and Meier KJ (2004) The Politics of Latino Education: The Biases of At-Large Elections.The Journal of Politics66(4), (November) 1224–1244. Levine P and Abromowitz D (2023) Challenges reported by candidates for local office.State and Local Government Review 55(2), 139–152. Lineberry RL (1977) Equality and urban policy: The distribution of municipal public services, 36–38; 53–57. Lopez, M. H.(2014). In 2014, latinos will surpass whites as largest racial/ethnic group in california. Technical report, Pew Research Center, Washington, D.C.https://www.pewresearch.org/short-reads/2014/01/24/in-2014-latinos-will-surpass- whites-as-largest-racialethnic-group-in-california/. Macmanus SA (1978) City Council Election Procedures and Minority Representation: Are They Related?.Social Science Quarterly 59(1), (June) 153–161. Malinowski J(2013) Master’s thesis, University of Washington, Master’s thesis. Campaign spending in city council elections: A comparison of at-large and district contests. Mast E (2024) Warding off development: Local control, housing supply, and nimbys.Review of Economics and Statistics 106(3), 671–680. Oliver JE, Ha SE and Callen Z(2012)Local Elections and the Politics of small-Scale democracy.Princeton University Press. Phillips CD (2021)Nowhere to Run: Race, Gender, and Immigration in American Elections.Oxford University Press. Robinson TP and Dye TR (1978) Reformism and Black Representation on City Councils.Social Science Quarterly 59(1), (June) 133–141. Sass TR (2000) The Determinants of Hispanic Representation in Municipal Government.Southern Economic Journal 66(3), (January) 609. SchaffnerBF,RhodesJHandRajaRJL (2020)Hometowninequality:Race,class,andRepresentationinAmericanLocalpolitics. Cambridge University Press. Schaffner BF, Streb M and Wright G(2001) Tearns without uniforms: the nonpartisan ballot in state and local elections. Political Research quarterly 54(1), 7–30. Schaffner BF, Streb MJ and Wright GC (2007) A new look at the republican advantage in nonpartisan elections.Political Research Quarterly 60(2), 240–249. Shah PR, Marschall MJ and Ruhil AVS (2013) Are We There Yet? The Voting Rights Act and Black Representation on City Councils, 1981–2006.The Journal of Politics 75(4), (October) 993–1008. Sojourner AJ (2013) Do unions promote members’ electoral office holding? evidence from correlates of state legislatures’ occupational shares.ILR Review 66(2), 467–486. Spicer Z, McGregor M and Alcantara C(2017) Political opportunity structures and the representation of women and visible minorities in municipal elections.Electoral Studies 48, 10–18. Taebel D (1978) Minority Representation on City Councils: The Impact of Structure on Blacks and Hispanics.Social Science Quarterly 59(1), (June) 142–152. Trounstine J (2008)Political Monopolies in American cities: the Rise and Fall of Bosses and reformers.Chicago: University of Chicago Press. OCLC: ocn196310795. Trounstine J (2010) Representation and Accountability in Cities.Annual Review of Political Science 13(1), (May) 407–423. Trounstine J (2018)Segregation by Design: Local Politics and Inequality in American Cities.New York, NY: Cambridge University Press. Trounstine J and Valdini ME (2008) The Context Matters: The Effects of Single-Member versus At-Large Districts on City Council Diversity.American Journal of Political Science 52(3), (July) 554–569. Welch S and Bledsoe T (1988)Urban Reform and its consequences: a Study in representation.Chicago: University of Chicago Press. Welch S (1990) The impact of at-large elections on the representation of blacks and hispanics.The Journal of Politics 52(4), 1050–1076. Yoder J (2020) Does property ownership lead to participation in local politics? evidence from property records and meeting minutes.American Political Science Review 114(4), 1213–1229. Cite this article:Cushing-Daniels M, Jones D and Shannon B (2026) District-based elections and class-based representa- tion: evidence from the California Voting Rights Act.Political Science Research and Methods, 1–17.https://doi.org/10.1017/ psrm.2025.10070 https://doi.org/10.1017/psrm.2025.10070Downloaded from https://www.cambridge.org/core. Montana State University Library, on 11 Mar 2026 at 17:50:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.46 February 13, 2025 Free Cities Center Analysis: City representation – Single- member districts versus at large pacificresearch.org/free-cities-center-analysis-city-representation-single-member-districts-versus-at-large At-Large Elections: Advantages and Disadvantages In at-large elections, all elected representatives represent the entire polity, be it a city or county. Proponents argue that this method promotes a unified vision among members since they all serve the same district. Additionally, at-large elections eliminate the issue of gerrymandering within the city, as there are no individual districts to manipulate. It also allows every citizen to vote for every member of the city council. One can argue that no one is left out. 1/547 Those reasons are compelling for many people, but the most significant drawback of at-large elections is the very real marginalization of minorities. For example, if a city is 60% white and Democrat, it is common for all city council members to be white and Democrats (or vice versa). The majority easily forms a majority while they hold the majority. The cost of running is higher at large, which also reinforces the majority and further marginalizes other voters. While occasionally someone outside this paradigm may be elected, they are immediately in the minority if they vote differently from the rest. Though just being there has an effect and that is substantive and moves the ball so to speak, they are easily voted down. At-large districts often result in every member of a city council being from one party, and one race and other factors leading to a lack of diverse perspectives and often an oppression of opposing views. Single-Member Districts: Benefits and Concerns In single-member district elections, the city is divided into districts with relatively equal population sizes as required by law, and each member runs from their respective district. One advantage is that the cost of running for office is significantly lower, as candidates only need to campaign within their specific district and not the entire community. Research shows that voter participation tends to be higher in district elections, possibly due to increased engagement and a sense of direct representation. Candidates do not have the burden of running in the entire district. Thus, they can spend more time with their smaller group of voters. Critics of single-member districts express concern about each representative focusing primarily on their own district’s interests, a phenomenon known as parochialism. This can lead to conflicts between districts and a lack of a citywide perspective. However, proponents argue that this localized focus ensures that each area’s unique needs and concerns are addressed. One exemplary problem could be that if one rep had a post office put in their district, everyone else would want one too. This can be true, but it also represents a form of fairness. Everyone in one way or another contributes taxes. If one district is given something, it is only natural that other districts seek something to even things out. It might lead to overall general consumption that people might not want, but it also prevents one part of the city from getting all the spoils so to speak. People do keep score. There is a point for each person when you can no longer ignore the disparities and that is human nature. Key Differences: Access, Cost and Accountability The most critical differences between at-large and district elections lie in the access afforded to minorities, the cost of elections and the accountability of members to their constituents. Single-member districts provide better opportunities for minority representation, as 2/548 historically underrepresented communities can have a stronger voice in electing their own representatives. The lower cost of running in district elections allows for a more diverse pool of candidates. At-large races are generally left to those with significant financial resources. Finally, district representatives are more directly accountable to their constituents, as they are elected by and responsible for a specific area. When something happens in a district, it is that representative whose duty it falls on to fix the matter. At-large elections allow the representatives to escape responsibility. History and Practice in California In California, every county appears to have district elections, while cities have a mixed history. Research shows that all cities initially had district elections and strong mayors with veto power. Over time, many cities transitioned to at-large elections. The author believes this is likely due to the involvement and greater influence of government officials and employees, but this trend is now reversing due to the passage of the California Voting Rights Act (CVRA). Santa Barbara Case Study: From At-Large to District Elections The author’s experience in Santa Barbara, Calif., provides a compelling example of the transition from at-large to district elections. When elections were held at large, the cost of running increased significantly, with each winning candidates eventually raising around $100,000 to have a chance at winning. Campaigns focused on prime voters – those who voted four of the last four elections were the doors that were generally knocked on. Most focus went particularly to an affluent area coined as the “golden triangle” area, leading to a lack of diversity on the City Council. The more focus over time simply reinforced its prominence to the detriment to other areas. One strange period had all the members of the City Council living on the mesa within a few blocks of each other. At-large elections simply have a proclivity to electing people who were often living in the same area, were from the same political party and had the same racial background. The Santa Barbara Tea Fire incident highlighted the problems with at-large elections at the end of 2008. During the fire, City Council members went on television and individually addressed issues based upon the fire in general terms, leaving citizens unsure of whom to contact with specific concerns. This lack of direct representation and accountability was clearly a significant drawback of the at-large system. A lawsuit based on the CVRA forced Santa Barbara to adopt district elections. The first election under the new system saw a significant drop in campaign costs and increased voter participation. Two majority-minority districts for Latinos were created, and the winner of one 3/549 of these districts ensured that more resources were allocated to previously neglected areas. Interestingly, when district elections were brought back, the city police chose to no longer participate in the elections with support or endorsements, likely to avoid finding themselves with an opponent in office. Again, it appears that government employees may favor at-large elections as it appears to increase their influence on the electoral process. In one word, it gives those in government more power and in turn, many minorities are not given choices. Legal Perspective: The California Voting Rights Act The CVRA has been a driving force behind the shift from at-large to district elections in California. Enacted in 2002, the CVRA prohibits the use of at-large elections in local government if they impair the ability of a protected class to elect candidates of its choice or influence the outcome of an election (California Elections Code § 14027). In Sanchez v. City of Modesto (2006), the California Court of Appeal held that the CVRA applies to both charter cities and general law cities. This decision has led to numerous lawsuits and settlements, resulting in many cities transitioning from at-large to district elections to avoid costly litigation. The CVRA has withstood constitutional challenges. In Jauregui v. City of Palmdale (2014), the Court of Appeal rejected arguments that the CVRA violated the Equal Protection Clause of the U.S. Constitution or the California Constitution. The Importance of Minority Representation and Access In district elections, citizens with specific problems can address their concerns directly to their representative, increasing the likelihood of their issues being addressed. In at-large elections, issues affecting a small number of people are less likely to receive attention, especially as the city’s population grows. The most reasonable and proper solution is to move to single-member districts and increase the number of representatives on the council. In a republican form of government, minority rights must be vigilantly protected. The American system of government is styled as a democratic system, meaning that each person should have their part in government and not be structurally excluded from the process. Access to representative bodies should be a fundamental right, as the purpose of representative government is to provide access to those being represented. Districts Elections are the Better Model Single-member district elections offer a more equitable and accessible form of representation, particularly for minorities and underrepresented groups. The CVRA has been instrumental in promoting this transition in California, with courts upholding its 4/550 constitutionality and applicability to all cities. While at-large elections may have some advantages, the potential for minority marginalization and lack of direct accountability make single-member districts a more democratic and representative system. As cities and counties continue to grapple with issues of representation and governance, the move towards single-member districts is a step in the right direction for ensuring fair and inclusive political participation. Michael Warnken works on court cases relating to representation and other similar issues. He is an advisor for the nonprofit Citizens Rising. 5/551 MIT Open Access Articles At#Large Elections and Minority Representation in Local Government The MIT Faculty has made this article openly available. Please share how this access benefits you. Your story matters. Citation: Abott, Carolyn and Magazinnik, Asya. 2020. "At#Large Elections and Minority Representation in Local Government." American Journal of Political Science, 64 (3). As Published: http://dx.doi.org/10.1111/ajps.12512 Publisher: Wiley Persistent URL: https://hdl.handle.net/1721.1/140916 Version: Author's final manuscript: final author's manuscript post peer review, without publisher's formatting or copy editing Terms of use: Creative Commons Attribution-Noncommercial-Share Alike 52 Author Manuscript This article is protected by copyright. All rights reserved. At-Large Elections and Minority Representation in Local Government Carolyn Abott † Asya Magazinnik ‡ December 19, 2019 Keywords: at-large elections, local government, education politics, minority representation, voting rights This work would not have been possible without a grant from the Spencer Foundation that allowed us to collect this data for the Education Governance and Accountability Project (EGAP) at The Ohio State University. Special thanks go to Vlad Kogan and St´ephane Lavertu at EGAP, and to Nolan McCarty, Kosuke Imai, and members of the Imai Research Group at Princeton University for their invaluable feedback on this project. Additional thanks to participants of the 2017 CSAP American Politics Conference at Yale University. We are also grateful to Robert Rubin for many informative conversations. †Department of Government and Politics, St. John’s University; abottc@stjohns.edu. ‡Department of Political Science, Massachusetts Institute of Technology; asyam@mit.edu. 1 This is the author manuscript accepted for publication and has undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to dierences between this version and the Version of Record. Please cite this article as doi: 10.1111/ajps.12512 53 Author Manuscript This article is protected by copyright. All rights reserved. Abstract Despite a long history of legal challenges alleging that elections conducted at-large suppress minority representation, this remains the dominant electoral system in local governments throughout the United States. Moreover, a large empirical literature remains divided over the present-day impact of at-large elections on the political success of underrepresented groups. We reconcile the competing findings in this literature by providing contingent, causal estimates of the eect of conversion from at-large to ward elections on minority oceholding, using a novel identification strategy aorded by the California Voting Rights Act of 2001. We find a dramatic positive eect of conversion in districts where Latinos constitute a suciently large share of the voting population, and in large and residentially segregated districts. When these conditions are not satisfied, we consistently see null estimated eects. Replication materials: The data and materials required to verify the computational re- producibility of the results, procedures and analyses in this article are available on the American Journal of Political Science Dataverse within the Harvard Dataverse Network, at: https://doi.org/10.7910/DVN/6OEPWE. Word Count:9,583 54 Author Manuscript This article is protected by copyright. All rights reserved. 1 Introduction Free and fair elections are the very essence of modern democracy. The institutions that structure how this exercise of popular will is translated into political representation are no less important. In the United States, electoral rules governing who can vote and how votes are aggregated to legislative seats have been used to exclude women and ethnic minorities from full political participation. By the same token, the past half century has seen a number of attempts at reforming the rules of the game to correct these historical inequities. One prominent example has been the push to eliminate at-large voting in local elections across the country. In at-large elections, voters across an entire constituency have the oppor- tunity to select candidates for every available seat in its legislative body. This is in contrast to the classic majoritarian scheme that divides the constituency into wards, each having its own seat in the legislature. 1 In at-large systems, its opponents claim, the minority vote is diluted by majority interests voting as a bloc. If minorities are highly concentrated in particular regions, as they have been historically, then switching to ward representation can at least guarantee them seats wherever they constitute a local majority, thereby increasing their voice in the political process. If these claims are accurate, then the elimination of at-large voting would have profound, far-reaching consequences for political representation across the United States. At-large sys- tems are still the prevailing institution in American local elections: as of 2012, approximately 64 percent of U.S. cities relied exclusively on at-large voting for their city council elections, with another 21 percent employing some combination of at-large and ward systems (Clark and Krebs 2012). 2 Governing bodies elected at-large—city councils, school boards, and mu- 1This representational scheme is known by many dierent names, including “district,” “by-trustee,” and “single-member.” For the sake of clarity we will use the term “ward” throughout, and reserve “district” to refer to the entire political unit—in our empirical case, the school district—whether it has at-large elections or is further subdivided into wards. 2These figures are estimates that come from surveys conducted by the International 1 55 Author Manuscript This article is protected by copyright. All rights reserved. nicipal boards—make decisions about how education is funded, where roads are built, and how water and sanitation services are delivered; they determine housing, economic devel- opment, transportation, and urban planning policies that shape their constituents’ daily lives. Recognizing the importance of local electoral institutions, coalitions of minority groups and civil rights advocates have mobilized to take legal action against at-large voting districts, resulting in several historical waves of conversion to wards (Browning, Marshall and Tabb 1986). Most recently, the California Voting Rights Act (CVRA) of 2001 lowered the legal standard of victory for plaintis suing at-large districts for suppressing minority represen- tation. The result was a flurry of successful litigation across the state to break up at-large school board and city council districts into wards, amassing millions of dollars in legal fees and settlements, and more than tripling the incidence of ward-based systems in California over the fifteen years following the CVRA’s passage (California Common Cause 2017). But these institutional reform eorts have also generated significant controversy—even among supporters of their general goals. Ward elections, critics argue, breed parochialism, fragmen- tation, and less responsiveness to the interests of the larger constituency; moreover, reforms have not been nearly as eective as expected at increasing the number of Latino representa- tives elected to oce. 3 Meanwhile, costly lawsuits deplete already overstretched municipal budgets, hurting the very constituents that the law was meant to empower. At least one of the focal points of this debate—the eectiveness of conversion to ward representation in propelling minority candidates to oce—is an empirical question that a careful study of electoral institutions and outcomes ought to resolve. Unfortunately, a vast City/County Management Association. Though imperfect, and subject to survey measure- ment error, these estimates are the best measure we have of the frequency of local at-large elections. 3See, for instance, https://www.latimes.com/politics/la-pol-ca-voting-rights-minorities- california-20170409-story.html (last accessed 09/30/19). 2 56 Author Manuscript This article is protected by copyright. All rights reserved. research agenda attempting to measure the relative merits of the two systems for minority oceholding has been unable to produce conclusive results. While there is substantial evi- dence that ward elections for local oce result in greater representation for Black and Latino voters (Molina Jr. and Meier 2016; Marschall, Ruhil and Shah 2010; Trounstine and Valdini 2008; Meier, Juenke, Wrinkle and Polinard 2005; Leal, Martinez-Ebers and Meier 2004; Poli- nard 1994; Davidson and Grofman 1994; Moncrief and Thompson 1992; Stewart, England and Meier 1989; Karnig and Welch 1982; Davidson and Korbel 1981; Engstrom and McDon- ald 1981; Robinson and England 1981), a number of additional studies have found no eect (Fraga 2009; Welch 1990; Bullock and MacManus 1987; MacManus 1978; Cole 1974), with still others that posit a negative association between ward elections and minority electoral success (Meier and Rutherford 2014; Welch and Karnig 1978). Two factors explain the lack of scholarly consensus on this subject: the highly contingent nature of the eect of electoral reform, and the extreme diculty of recovering plausibly causal estimates of this eect. Ward elections can only help minority candidates win oce when the minority group is suciently large and geographically concentrated, such that it constitutes a local majority. Identifying the cases where these important prerequisites are in place is beyond the scope of most datasets that have been available to date, but essential for a proper test of the reform. 4 Moreover, the vast majority of studies have attempted to identify the eects of at-large versus ward representation by comparing localities with each kind of system, or, at best, those that have switched to ward systems to those that have not. But even after controlling for any number of covariates, crucial unobserved dierences remain between areas with long histories under each rule, and those that choose to switch to wards are fundamentally dierent from those that do not. The result is a selection eect that can bias the estimated quantity of interest in either direction. As the political debate in California intensifies—and additional states such as Texas 4Some notable exceptions that consider the contingent eects of institutional arrangements include Marschall, Ruhil and Shah (2010) and Trounstine and Valdini (2008). 3 57 Author Manuscript This article is protected by copyright. All rights reserved. appear poised to follow California’s lead in undertaking reform 5—questions about the relative merits of at-large and ward systems for the interests of historically underrepresented groups take on a new urgency. The contribution of the present study is to provide the first causal, contingent, and policy-relevant estimates of the eect of conversion from at-large to ward representation on minority oceholding. To estimate these eects, we apply a novel research design that takes advantage of exogenous variation in electoral reform induced by the CVRA, as well as an original panel dataset that allows us to measure key predictors of the reform’s success. Our findings shed light on why the academic literature has for so long failed to arrive at a consensus—and on the conditions that advocates who hope to achieve real lasting change need to target in their eorts. Consistent with theoretical expectations, we find a dramatic positive eect of being forced to convert to ward elections under the CVRA on Latinos’ ability to get elected to California school boards among districts with a high level of residential segregation. Our analysis also uncovers a key moderator of the eect of reform that previous studies of at-large elections have ignored, though scholars have long recognized its importance for political mobilization: the size of districts, as measured by total school enrollment. In large and segregated districts, the eects of electoral reform are generally positive and steadily increasing in the size of the Latino community—a dierence that rises above one additional Latino oceholder for every three available seats—but may also be negative when the minority community is suciently small. When these conditions are not met, we consistently see null estimated eects. Additionally, we examine districts that voluntarily chose to convert from at-large to ward elections. Through the use of an instrumental variables framework, we show that conversions spurred by district-adjacent legal threats have a large and unconditional positive impact on the political fortunes of the Latino community. 5See https://www.texastribune.org/2018/12/11/richardson-isd-school-board-representation/ (last accessed 09/30/19). 4 58 Author Manuscript This article is protected by copyright. All rights reserved. 2 Theory Scholars of local and urban politics have deservedly devoted significant attention to the dierential eects of at-large and ward elections on oceholding among Blacks, Latinos, and women.6 The justification for their focus on descriptive representation—broadly conceived as the resemblance between the representative and the represented along some politically salient dimension, such as sex or ethnicity (Pitkin 1967)—is twofold. First, descriptive representa- tion is valuable in its own right, as it has been shown to increase underrepresented groups’ sense of political ecacy, trust in government, and legitimacy of the governing regime, and to provide role models for those groups (Phillips 1991, 1995, 1998; Mansbridge 1999; Dovi 2002). Second, descriptive representation may lead to more tangible gains, such as an in- crease in the share of public resources allocated to a given group, or an improvement in the quality of life of its members (Haider-Markel, Joslyn and Kniss 2000; Wald, Button and Rienzo 1996; Fraga, Meier and England 1986; Meier, Stewart and England 1991; Leal, Martinez-Ebers and Meier 2004; Marschall and Ruhil 2007; Browning, Marshall and Tabb 2003). How might ward elections increase descriptive representation of minorities compared to at-large systems? A key condition for this relationship to hold is that the voting population be segregated enough for the minority group to constitute a local majority in at least one ward, and that the political boundaries be drawn accordingly (Trounstine and Valdini 2008; Marschall, Ruhil and Shah 2010; Sass 2000). To understand why this is the case, consider the unit depicted in Figure 1. The diagram illustrates how votes in an electorate (above) are translated to seats in a governing body (below) under at-large and ward systems. In each of the three cases shown, the polity consists of two groups with opposing political interests: majority group A, which comprises 13/20 of the voting population, and minority group B, 6See Table 1 in Marschall, Ruhil and Shah (2010) for a useful review of the vast empirical literature on Black representation alone. 5 59 Author Manuscript This article is protected by copyright. All rights reserved. which comprises the remaining 7/20. In the first case, there is an at-large system: every member of the electorate gets to vote on every seat in the legislature as it becomes available. As long as Group A can field a candidate for every race, it will be victorious every time, leaving the sizable minority without any political representation. Moving to a ward system, as Cases II and III show, may improve electoral outcomes for Group B—conditional on how the members of this group are distributed. In Case II, the electorate is divided into four wards, each with its own seat in the legislature. Because the boundaries are drawn such that the minority population constitutes a local majority in Ward 1, Group B is able to capture one seat and increase its voice in the legislature. But as Case III shows, geographic segregation is a necessary condition for ward representation to yield electoral gains for this group. Here, Group B is distributed approximately equally throughout the wards, and remains a minority in each one. As a result, the problem present in the at-large system is replicated one level down, and wards yield no representational gains for the minority. Another key determinant of whether ward systems can improve electoral outcomes for underrepresented groups is district size—though there are compelling reasons to believe the eect may swing in either direction. On the one hand, the likelihood of finding willing and qualified candidates to run for oce increases with district size, and those candidates may be able to take advantage of greater resources and more sophisticated political organization. On the other hand, smaller districts may advantage political outsiders who have less money and experience, as they can garner support from their local communities through face-to-face contact rather than large-scale campaigns. And while scholars have gone so far as to caution against extrapolating from findings on large units to smaller ones (Welch 1990), no further empirical work to our knowledge has focused on this important contingency. For policymakers weighing the benefits of conversion from at-large districts to wards, process also matters. In the U.S., reform has generally occurred in one of two ways: localities have been forced to change their systems as the result of successful litigation against them, or they have chosen to do so voluntarily, either by popular vote or a unilateral decision of 6 60 Author Manuscript This article is protected by copyright. All rights reserved. Figure 1: Conversion of Votes to Seats, Wards vs. At-Large Districts Ward 1 Ward 2 Ward 3 Ward 4 Ward 1 Ward 2 Ward 3 Ward 4 (II) Ward, Segregated (III) Ward, Not Segregated(I) At-Large Ward 1: 3/5 A, 2/5 B Ward 2: 3/5 A, 2/5 B Ward 3: 3/5 A, 2/5 B Ward 4: 4/5 A, 1/5 B Ward 1: 1/5 A, 4/5 B Ward 2: 4/5 A, 1/5 B Ward 3: 4/5 A, 1/5 B Ward 4: 4/5 A, 1/5 B Overall: 13/20 A, 7/20 B Seat 1 Seat 2 Seat 3 Seat 4 Seat 1 Seat 2 Seat 3 Seat 4 Seat 1 Seat 2 Seat 3 Seat 4 Notes: Group A represented in dark gray and Group B represented in light gray. Squares represent the geo- graphical distribution of the voting population; circles represent seats on the legislative council. Proportion of the electorate comprised of Groups A and B is the same in each case. 7 61 Author Manuscript This article is protected by copyright. All rights reserved. the governing body. In the following section, we discuss the historical context of American local electoral reform, culminating in the adoption of California’s own Voting Rights Act. The CVRA led to a dramatic rise in both kinds of conversions: those resulting from legal action (or the threat thereof), which began to favor plaintis after the law’s passage, and those undertaken voluntarily as the issue gained salience around the state. In both cases, the law’s implementation and the way in which reform unfolded introduced some random variation in the likelihood that districts would change their systems, and we discuss how we exploit this in our empirical analysis. 3 Causal Identification Through the CVRA Historical Context Since the passage of the federal Voting Rights Act of 1965, at-large districts around the country have come under legal attack on the grounds of minority vote dilution (see Browning, Marshall and Tabb (1986)), but not all plaintis have been successful in federal court. The Supreme Court ruling in Thornburg v. Gingles in 1986 clarified the standards that a claimant must meet in order to demonstrate that at-large elections are responsible for a failure of representation, setting a high bar for plaintis alleging vote dilution in at-large elections (Trounstine and Valdini 2008; Epstein and O’Halloran 1999). Many suits filed under the federal Voting Rights Act since then have been unsuccessful as a result. Then, in 2001, California passed its own Voting Rights Act eliminating the Gingles requirements and imposing a much lower standard on plaintis: to win in court, they would only have to demonstrate the presence of “racially polarized voting” in the district. Moreover, unlike in cases filed under federal law, the CVRA required the district being sued to pay all legal fees, even if the two parties chose to settle out of court. Civil rights groups across California quickly recognized a powerful tool in the CVRA. Under this new law, suing an at-large district for minority vote dilution had relatively little 8 62 Author Manuscript This article is protected by copyright. All rights reserved. downside compared to federal cases: most of the time, the prospect of paying exorbitant legal fees convinced districts to convert at the mere threat of legal action, and if a case went all the way to trial, the low “racially polarized voting” standard almost assured victory for the plainti. The result was a mobilization of eorts around the state to initiate legal action in as many at-large districts as possible where there was a mismatch between the size of the Latino population and its representation in local government, a movement described as no less than “a quiet revolution” with the potential to transform “the literal face of California politics.”7 In 2002, the vast majority of California’s school districts had at-large elections (906 of 978 districts in our sample); by 2017, 138 of these districts had switched to ward systems, either voluntarily or under direct threat of legal action. For a complete picture of how the CVRA transformed California’s electoral landscape, we estimate the eects of two distinct treatments on Latino oceholding across the state: first, the direct eect of conversion to ward representation on districts threatened with litigation, and second, the indirect eect on districts that were incentivized to convert of their own accord by the new legal standard. Taken together, these estimands do more than measure the dierences between at-large and ward systems, which is where the vast majority of the academic literature stops; rather, they illuminate the change in minority representation we can expect from precisely the sort of mechanism that policymakers would use to induce institutional reform. The Eect of Conversion by Legal Threat: A Fixed Eects Approach The constraint upon the number of CVRA cases that civil rights groups could file was not the availability of at-large districts where a case could succeed, but rather the willingness 7Quoted in “Districts Abandoning At-Large Elections,”Education Week, http://www. edweek.org/ew/articles/2013/02/27/22schoolboards ep.h32.html (last accessed 09/30/19). 9 63 Author Manuscript This article is protected by copyright. All rights reserved. of potential plaintis to engage in legal action. Civil rights groups operating at the state or national level could not initiate suits themselves, but had to do so on behalf of residents of the jurisdiction that they were taking to court. Through a series of interviews with the lawyers centrally involved in the cases tried under the CVRA, we learned about the process driving the conversion of school districts to ward elections in California. First, they identified all of the at-large districts across the state where there was a sizable Latino population (as a general rule of thumb, at least fifteen percent) and a misalignment between the size of the Latino population and representation. Next, they arranged meetings with local organizations within these districts to educate citizens about at-large voting. They would ask whether community members felt that minority interests could be better represented in local government, and whether they thought their districts would benefit from ward representation; invariably, the answer was yes. They also hired statistical consultants to assess whether there was evidence of “racial polarization” in the selected districts, but given the low standard set by the CVRA and the demographic realities on the ground, this step did not eliminate many candidates. The most significant drop-o from initial identification to ultimate litigation occurred at the level of plainti recruitment: although the lawyers had no diculty demonstrating the value of legal action to the communities they approached, they struggled to convince specific individuals to shoulder the burden themselves. Importantly, the ability of civil rights lawyers to identify plaintis in targeted areas, with limited time and resources at their disposal, was not systematically related to the key political characteristics of those districts. To be sure, the people who stepped forward as plaintis do not represent a random sample of citizens. As Robert Rubin, a lawyer with the Lawyers’ Committee on Civil Rights (LCCR) who spearheaded CVRA litigation, stated, “you’re asking a member of a disenfranchised community to fight the establishment,” so the plainti would often be a retiree or someone not working for the city. However, conditional on being identified for legal action, the communities from which these plaintis were drawn closely resembled those where a plainti could not be found, and varied widely in “political 10 64 Author Manuscript This article is protected by copyright. All rights reserved. sophistication”: while some were highly organized, others were unprepared for a legal victory, and could not field a competitive candidate after winning their case. 8 Figure 2: Process of Legally Mandated Conversion from At-Large Districts to Wards under the CVRA Figure 2 summarizes all of the pathways to conversion as a result of legal action, and helps clarify how we define the treatment and control groups in our fixed eects analysis. 346 at-large districts met the LCCR’s criteria for identifying potential litigants. The criteria were simple: the district’s population had to be at least 15% Latino, the voting-eligible population had to be less than 60% Latino, 9 and the school board had to be composed 8Conversation with Robert Rubin. 9If the voting-eligible population was greater than 60% Latino, Latinos would be the majority bloc and therefore hurt by conversion to ward districts. We follow the LCCR’s 11 65 Author Manuscript This article is protected by copyright. All rights reserved. of proportionally fewer Latinos than the district’s population. The remaining 632 districts in the state were outside of their target group for one of three reasons: they already had ward elections by the time of the CVRA’s passage, they did not satisfy the demographic and political selection criteria described above, or they voluntarily converted without the lawyers’ intervention. Within the target group, four suits were successfully filed, with one going to court and three others settling, all in favor of the plainti. In an additional twenty cases, districts received a threat letter stating that they would soon be sued if they did not promptly change their systems. The threat was credible: organizations only sent such a letter if they had in fact identified claimants from the community who were prepared to move forward if necessary (see Appendix A.1 for an example of such a letter). Regardless of whether a school board was merely threatened with legal action or taken all the way to court, the final outcome was the same: the district converted to ward elections, either by vote or by obtaining a waiver from the state permitting them to convert by fiat. These cases, shown in dark gray in Figure 2 (and listed in full in Appendix A.4), constitute the treatment group in our analysis. The control group, shown in light gray, includes all of the districts that would have followed the same deterministic trajectory if not for the short-term inability to find a plainti. Within this sample of potentially treated districts, we estimate the two-way fixed eects regression model: Yit =0 +1 proportionLatinoit +2wardit +3(proportionLatino * ward)it+ Xit +i +t +it (1) where the outcome Yit is the proportion of seats up for election in school district i and election year t won by Latino candidates; proportionLatinoit is the proportion of the over-18 protocol in choosing a 60% threshold rather than a bare 50% majority in order to adjust for low turnout within this group. This does not dramatically change the sample, and the results are not sensitive to this choice. 12 66 Author Manuscript This article is protected by copyright. All rights reserved. population of the district that is both Latino and eligible to vote (a native-born or naturalized U.S. citizen); wardit is a binary indicator for having ward elections;Xit is a vector of financial, demographic, and socioeconomic controls (described more fully in the Data and Measurement section); and i and t are district and year fixed eects, respectively. The primary quantity of interest is the marginal eect of conversion estimated at a given level of Latino population, or 2 +3 proportionLatino, because if the reform did indeed improve minority electoral outcomes, then the Latino composition of school boards should rise with the size of the Latino population in the district. This is the modeling approach recommended by Engstrom and McDonald (1981) and used in other recent empirical studies on the subject (e.g., Trounstine and Valdini (2008), Meier and Rutherford (2014)). Our estimates recover the causal eects of conversion from at-large to ward representa- tion. The fixed eects account for any time-invariant dierences between treated and control districts that could bias the results. The assumption for identification is that there are also no unobserved time-variant sources of selection into treatment. We are confident that this is the case. The two sets of districts have statistically indistinguishable pre-treatment elec- toral outcomes even before applying any controls. However, to minimize the potential for confounding and to increase the precision of our estimates, we additionally control for a wide range of district-level demographic, financial, and socioeconomic characteristics, enumerated in the Data and Measurement section below. The Eect of Voluntary Conversion: An Instrumental Variables Approach Most of the CVRA’s eect was not through lawsuits. Fearing repercussions, most places that converted did so of their own accord. One factor that made districts more likely to con- vert, particularly early on, was the presence of legal action in the same county. According to one media report, for instance, a lawsuit in nearby Madera convinced a school superin- tendent in Fresno County to mandate reform across all the districts under his jurisdiction. 13 67 Author Manuscript This article is protected by copyright. All rights reserved. Moreover, he faced no political opposition. “I’ve had no chafing on the part of anybody,” he said. “They said, ‘It’s the right thing to do. Let’s do it.’” 10 Litigation nearby, it seems, served as an exogenous shock to some districts that were amenable to minority incorporation but unaware of the potential problems with at-large representation or lacking in political will to enact reform; alternatively, they were fearful of the costs of litigation but previously un- aware of the legal threat posed by the CVRA. In this analysis, we use legal action in another district in the same county as a binary instrument for the treatment of voluntary conversion to ward representation. We prefer a binary instrument of same-county membership to a continuous distance measure because CVRA cases are initiated in the county courts, and because the county is the next administrative unit above school districts in local educational governance; thus, we would expect information about legal action to diuse over county-level networks rather than uniformly over geographical distance. For the IV analysis to identify a causal eect of voluntary conversion, the exclusion restriction must be satisfied: nearby legal action must only aect Latino representation in a school district by increasing its propensity to convert to ward elections, and not by any other means. This rules out, for example, the possibility that nearby legal action mobilizes Latinos to push for greater representation under their current at-large system through candidate financing or voter turnout. We do not think this is likely to be the case. The rhetoric around the CVRA cast at-large elections as a first-order barrier to minority political access, and lawyers were actively seeking plaintis to take part in legal action; it is dicult to imagine these activities mobilizing community activism for something other than electoral reform. Nevertheless, any such spillover eects should bias our results downward, as they would reduce the contrast between electoral outcomes in at-large and ward districts. 10Los Angeles Times, “Madera Unified case is changing elections throughout California,” http://articles.latimes.com/2009/jan/04/local/me-madera4 (last accessed 09/30/19). 14 68 Author Manuscript This article is protected by copyright. All rights reserved. Data and Measurement Treatment We constructed a comprehensive dataset tracking the process of conversion from at- large to ward representation that was set in motion by the CVRA. To do so, we conducted an extensive search of state and local media accounts, school board minutes, and publicly available records of waiver requests from the California Department of Education. For every school district in California, we documented which electoral system was in place in every year from the CVRA’s passage in 2001 through 2016, as well as how every switch occurred: by specific type of legal action (court ruling, settlement, or threat letter) or voluntarily. We also used data that we obtained from the Lawyers’ Committee on Civil Rights (LCCR), the organization centrally involved in CVRA litigation, to reconstruct the sample it targeted for legal action according to its own stated criteria. This dataset contains three key variables: the total number of seats on each district’s school board, the number of those seats occupied by someone with a Latino last name, and the proportion of the district’s population that was Latino. We used these variables to define a subgroup of districts where there was a sizable Latino population (over 15%) that exceeded the proportion of Latino representatives on the school board, just as the LCCR had done when they identified potential sites for legal action. Outcomes Through the Education Governance and Accountability Project at The Ohio State Uni- versity, we have obtained the names and vote counts of every candidate who ran for a school board position in California from 2001 to 2016. We aggregate these observations to construct our primary outcome as the number of school board candidates with Latino last names who won oce as a proportion of the number of seats in the district up for election in that year. 11 11Our election data only gives us names, not ethnicities, of candidates, so we identified Lati- nos using the wru package in R (Imai and Khanna 2017). This package employs a Bayesian 15 69 Author Manuscript This article is protected by copyright. All rights reserved. To unpack the mechanism by which more Latinos may have won oce, we also construct two secondary outcome variables: one capturing the availability of Latino candidates, and another capturing their share of the vote. To measure the former, we compute the proportion of seats up for election in a given district-year that had at least one Latino candidate on the ballot. Constructing the outcome in this way tells us how many school board seats Latino candidates could have possibly won, given how many actually ran for oce. For example, if two Latino candidates ran in an at-large election in which three seats were vacant, no more than two-thirds of the school district’s seats could be filled by Latinos. If there were two dierent ward elections in a school district in a given year, and at least one Latino ran in one of them but not the other, no more than one-half of the district’s school board seats could be filled by Latinos. Our other secondary outcome of interest is simply the vote share received by all candidates with Latino last names in each election, averaged across all races within a district-year. prediction procedure that uses data from the U.S. Census to compute the probabilities that a person is of a given ethnicity, given his last name and geolocation at the county level. There is a valid concern that surname alone may fail to accurately reflect one’s heritage, for instance as a result of someone taking their partner’s last name in mixed-ethnicity mar- riages, but for the present purposes, we do not believe this poses an issue. As ?point out, their identification method is biased only if the individual’s surname is correlated with her location or personal attributes, including the rate of interracial marriage and the likelihood of changing her last name after marriage. So long as Latina women are no more or less likely to marry non-Latino men than non-Latina women are to marry Latino men, and so long as Latina women are no more or less likely to change their surname than non-Latina women after entering into a marriage with someone of a dierent ethnicity, the phenomenon of an individual changing their last name after marriage should only introduce random noise—but no bias—into our estimates. 16 70 Author Manuscript This article is protected by copyright. All rights reserved. Residential Segregation and Additional Controls We measure residential segregation using the index of dissimilarity between whites and Latinos, computed at the school district level. The dissimilarity index captures how evenly whites and Latinos are distributed across schools within a district, and is given by: 1 2 N i=1 wi W -li L (2) where wi and li represent the number of whites and Latinos in school i, respectively;W and L represent the total number of whites and Latinos in the district, respectively; and N represents the total number of schools in the district. Another intuitive interpretation of this measure is the proportion of Latinos that would have to move to a dierent school in order for the composition of each school to be identical to the composition of the district as a whole (Ananat 2011). In general, a low dissimilarity index is considered to be below 0.3; 0.3 to 0.6 is considered moderate; and above 0.6 is considered high (Massey and Denton 1993). We choose to measure segregation with the dissimilarity index in keeping with a large literature that has favored its use (Ananat 2011; Collins and Margo 2000; Cutler, Glaeser and Vigdor 1999; Cutler and Glaeser 1997; Massey and Denton 1993), but supplement our analyses with the Theil index, an alternative measure of residential segregation, collected at the school district level by the Stanford Education Data Archive (SEDA). 12 Finally, we use a vector of controls assembled from the U.S. Census and the California Department of Education. We select time-varying characteristics that are important corre- lates of Latino political participation and vote choice, including: the proportion of students who are Black, white, and Asian; financial characteristics, including property taxes collected, total current spending on instruction, and total educational revenues and expenditures, all scaled by enrollment, as well as enrollment itself; and socioeconomic factors, including me- dian income in the district among all residents and specifically among Latinos, the proportion 12Within our sample, the Theil index and the dissimilarity index are correlated at 0.79. 17 71 Author Manuscript This article is protected by copyright. All rights reserved. of students who receive free lunch and English Language Learner (ELL) services, the pro- portion of the district living below the poverty line as well as between 100 and 149 percent of the poverty line, the proportion of Latinos who have less than a high school education, who have completed high school, and who have attended some college, the unemployment rate among Latinos, and the proportion of Spanish speakers who speak English “very well” as opposed to “less than very well.” Finally, we include a control for the total number of school board members, since the representational consequences of winning a seat will vary with overall board size. 4 Results Validating the Selection of Control Units We empirically verify our qualitative evidence that within the targeted subgroup, districts treated with legal action were analogous in prior Latino electoral performance to those where a plainti did not step forward. In other words, we do not see any evidence that lawyers strategically pursued litigation in districts where Latinos were more politically organized or successful, or, conversely, where they were particularly underrepresented. Figure 3 shows that, for each of our three electoral outcomes, the treatment group—defined as any district that underwent conversion to ward elections through legal action at any point during the time series—is statistically indistinguishable from the control group—defined as all districts in the targeted sample that never underwent conversion over the same period. Here, we only include pretreatment data: as soon as a district converts to ward elections, it exits the sample. Table 1 summarizes our sample of districts, contrasting the main variables of inter- est across the three categories of treatment status: those districts that remained at-large throughout our entire sample period and never converted to ward elections, those that con- verted after experiencing some sort of legal action, and those that converted voluntarily 18 72 Author Manuscript This article is protected by copyright. All rights reserved. without any sort of legal coercion. For the latter two classes of school districts, the summary statistics were calculated using years before the treatment occurred. Figure 3: Comparing Pre-Treatment Electoral Outcomes Within Sample Targeted for Legal Action 0.0 0.1 0.2 0.3 0.4 0.5 2004 2008 2012 YearPre-treatment outcome Treated Control (a) Latinos elected as a propor- tion of open seats 0.2 0.4 0.6 0.8 2004 2008 2012 YearPre-treatment outcome Treated Control (b) Proportion of open seats with at least one Latino candi- date 0.1 0.2 0.3 0.4 0.5 2004 2008 2012 YearPre-treatment outcome Treated Control (c) Vote share to all Latino can- didates Notes: Loess-smoothed lines are fitted through the data, and 95% confidence intervals are shown in gray. 19 73 Author Manuscript This article is protected by copyright. All rights reserved. Table 1: Summary Statistics (Pre-Treatment) Never Converted Legal Conversion Voluntary Conversion Min. Max. Mean (sd) Min. Max. Mean (sd) Min. Max. Mean (sd) Latino Winners (Prop.) 0 1 0.159 (0.222) 0 0.417 0.122 (0.129) 0 0.625 0.153 (0.154) Latino Candidates (Prop.) 0 1 0.296 (0.309) 0 0.708 0.346 (0.206) 0 1 0.379 (0.288) Latino Voteshare (Prop.) 0 1 0.174 (0.212) 0 0.397 0.157 (0.104) 0 0.635 0.191 (0.151) Dissimilarity Index 0 0.354 0.046 (0.066) 0 0.415 0.118 (0.108) 0 0.284 0.053 (0.064) Enrollment 11 58,780 6,110 (8,724) 2,973 48,912 16,698 (10,993) 117 79,266 8,699 (12,524) Latino VEP (Prop.) 0.062 0.580 0.250 (0.110) 0.139 0.401 0.273 (0.077) 0.042 0.739 0.274 (0.104) Household Income ($) 32,940 108,837 59,692 (15,396) 41,342 88,169 60,620 (11,689) 10,531 91,249 54,063 (11,867) Size of School Board 3 10 5.124 (0.895) 4 7 5.750 (0.989) 3 9 5.336 (1.052) N (School Districts) 322 24 112 Notes: First two columns are constructed from a subset of school districts that were “eligible” to have legal action taken against them, as is reflected in the fixed eects analysis, below. The third column is a subset of all, rather than strictly “eligible,” districts that voluntarily converted to ward elections and reflects the thrust of the instrumental variables analysis.2074 Author Manuscript This article is protected by copyright. All rights reserved. Fixed Eects Analysis Table 2: Eect of Ward Elections on Proportion of Elected Board Members that Were Latino Dependent variable: Segregation District Size All Low High Low High (1) (2) (3) (4) (5) Ward elections -0.301 1.119 -0.708 0.559 -0.873 (switch by legal threat) (0.276) (0.554) (0.223) (0.366) (0.205) Proportion Latino -0.203 -0.159 -0.528 -0.183 0.880 (voting eligible) (0.211) (0.227) (0.630) (0.220) (0.758) Ward * proportion Latino 1.050 -2.753 2.538 -1.464 3.351 (0.986) (1.415) (0.829) (1.009) (0.596) Year FE Yes Yes Yes Yes Yes District FE Yes Yes Yes Yes Yes Controls Yes Yes Yes Yes Yes Observations 1,477 1,117 360 1,158 319 R2 0.620 0.643 0.587 0.649 0.607 Notes:p<0.05;p<0.01;p<0.001. For complete results, see Appendix Table B.1. There is no overall eect of conversion to ward districts on Latino oceholding (see col- umn 1 of Table 2). But this result masks important heterogeneities: namely, the reform leads to a closer alignment between the size of the Latino population and its descriptive rep- resentation in large and segregated districts. To construct these subgroups, we first compute the mean enrollment and dissimilarity index for each school district over the time series. We then define the threshold for inclusion in the “high” group on each condition based on the median value of each moderator among treated units. Consistent with theoretical expectations, the degree of Latino-white segregation is a key moderator of the eect of conversion to ward districts on Latino oceholding. In Figure 4, we plot the marginal eects of conversion (1 +2 proportionLatino from Equation 1) against the proportion of the over-18 population of the district that is both Latino and eligible to vote (a native-born or naturalized U.S. citizen), henceforth Latino VEP. Panel 21 75 Author Manuscript This article is protected by copyright. All rights reserved. (a) shows eects for the full targeted sample, and (b) and (c) disaggregate by low and high segregation subgroup, corresponding to the results reported in columns 1-3 of Table 2). The x-axis ranges over the observed distribution of Latino VEP in the data, and the histograms at top and bottom reflect these distributions in the treatment and control groups, respectively. In relatively integrated districts, there is no evidence that the reform increased minority representation. However, in segregated districts with large Latino populations, conversion had a large and positive eect on the likelihood of Latinos winning elections. For instance, in a district with a Latino VEP of 40%, conversion to ward elections increased the proportion of seats that are won by Latinos by 31 percentage points. Figure 4: Marginal Eects of Conversion to Ward Elections on Proportion of Elected Board Members that Were Latino, by Geographic Segregation 0 2 4 6 0.2 0.3 0.4 0.5 Distribution, treatment observations -0.5 0.0 0.5 1.0 0.2 0.3 0.4 0.5 Proportion Latino VEPMarginal effect, Proportion Latinos elected0 25 50 75 0.2 0.3 0.4 0.5 Distribution, control observations (a) All 0 2 4 6 0.2 0.3 0.4 0.5 Distribution, treatment observations -0.5 0.0 0.5 1.0 0.2 0.3 0.4 0.5 Proportion Latino VEPMarginal effect, Proportion Latinos elected0 20 40 60 80 0.2 0.3 0.4 0.5 Distribution, control observations (b) Low Segregation (Dissimilarity Index 0.29) 0 2 4 6 0.2 0.3 0.4 0.5 Distribution, treatment observations -0.5 0.0 0.5 1.0 0.2 0.3 0.4 0.5 Proportion Latino VEPMarginal effect, Proportion Latinos elected05101520 0.2 0.3 0.4 0.5 Distribution, control observations (c) High Segregation (Dissimilarity Index >0.29) Notes: Results correspond to those reported in columns 1, 2, and 3 of Table 2. 95% confidence intervals are shown in gray. The same pattern emerges when analyzing district size as a moderator of the eect of reform. While conversion to ward elections in small school districts has no eect on the likelihood of Latinos winning oce, there is a dramatic and precisely estimated positive eect in large districts that are composed of at least 30% Latinos (Figure 5). For example, when 22 76 Author Manuscript This article is protected by copyright. All rights reserved. 40% of the VEP is Latino, the proportion of school board seats won by Latino candidates in large school districts increases by 47 percentage points after switching to ward elections. Figure 5: Marginal Eects of Conversion to Ward Elections on Proportion of Elected Board Members that Were Latino, by District Size 0 2 4 6 0.2 0.3 0.4 0.5 Distribution, treatment observations -0.5 0.0 0.5 1.0 0.2 0.3 0.4 0.5 Proportion Latino VEPMarginal effect, Proportion Latinos elected0 25 50 75 0.2 0.3 0.4 0.5 Distribution, control observations (a) All 0 2 4 6 0.2 0.3 0.4 0.5 Distribution, treatment observations -0.5 0.0 0.5 1.0 0.2 0.3 0.4 0.5 Proportion Latino VEPMarginal effect, Proportion Latinos elected0 20 40 60 80 0.2 0.3 0.4 0.5 Distribution, control observations (b) Small Districts (Enrollment 13,700) 0 2 4 6 0.2 0.3 0.4 0.5 Distribution, treatment observations -0.5 0.0 0.5 1.0 0.2 0.3 0.4 0.5 Proportion Latino VEPMarginal effect, Proportion Latinos elected0 5 10 15 0.2 0.3 0.4 0.5 Distribution, control observations (c) Large Districts (Enrollment >13,700) Notes: Results correspond to those reported in columns 1, 4, and 5 of Table 2. 95% confidence intervals are shown in gray. How can we account for the absence of an eect of institutional change in small districts? One likely explanation is that sheer numbers are useful for propelling minority candidates to oce. Larger constituencies are more likely to yield at least one high-quality candidate, and they allow candidates to build broader coalitions, mobilize voters more eectively, and take advantage of greater resources. According to a study of city council elections across the state by GrassrootsLab, conversions failed to translate into Latino representation in large part because of a shortage of candidates with the means to run. 13 Consistent with this claim, when we replicate our analysis on the intermediate outcome of running for, not 13https://www.latimes.com/politics/la-pol-ca-voting-rights-minorities-california-20170409-story. html (last accessed 09/30/19). 23 77 Author Manuscript This article is protected by copyright. All rights reserved. winning, oce (Appendix Figure B.2), we see a similar pattern of results to our main findings: an increase in large districts commensurate with the size of the Latino population, and no change whatsoever in small districts. In large and segregated districts with suciently small Latino populations, the reform actually had a negative eect on Latino oceholding. As the third panels of Figures 4 and 5 show, at Latino VEP of 0.20, conversions in both high-dissimilarity and high-enrollment districts decreased the proportion of seats won by Latinos by 20 percentage points (p < .05). The negative eects are not surprising: not only do these districts lack a large enough Latino minority to constitute an influential voting bloc, but, as critics of the reform have argued, introducing an ethnic gerrymander may amplify voters’ perceptions that political conflict falls along this particular dimension. The result—increasingly racially polarized voting coupled with small numbers of Latino voters relative to other groups—may create new barriers to Latino electoral victories. As the histograms at the top of Figures 4 and 5 show, not many of the treated observations were both low on Latino VEP and high on segregation and/or enrollment, and therefore not many districts experienced negative treatment eects in practice; that said, reformers ought to carefully consider moving forward with conversion eorts under this set of adverse conditions. Our data allows us to explore the pathways by which districts elected more Latino can- didates to oce when the right conditions were in place: conversion to ward representation both encouraged Latino candidates to run for oce in more school board races, and increased the share of the vote that they collectively received. In Appendix Figures B.1 and B.2, we show dramatic eects of conversion on Latino candidacy in only large and segregated dis- tricts. For instance, a large district with 40% Latino VEP saw a 48 percentage point increase in the number of elections with at least one Latino candidate in the race (as a proportion of all elections in the district that year). These candidates were also able to pull in a larger share of the vote: they received 27 more percentage points in segregated districts with a 40% Latino VEP (Appendix Figure B.3), and 38 more percentage points in large districts with a 24 78 Author Manuscript This article is protected by copyright. All rights reserved. 40% Latino VEP (Appendix Figure B.4). Finally, we compare districts that saw an increase in Latino candidacy after conversion to those where the reform had no immediate eect. Overall, 65% of legally treated districts had more races with at least one Latino candidate after the reform than prior. As Appendix Table B.2 shows, these successful districts were significantly larger than other treated districts where Latino candidacy did not increase. Unsurprisingly, they also reflected a higher level of wealth and social capital: fewer students received free lunch and English language services and lived below the poverty line, and more of the Latino population was employed and had some college education. Robustness Checks Our key findings are robust to alternative measures of residential segregation and defini- tions of the low and high subgroups. In Appendix Figure B.5, we replicate Figure 4 using the Theil index instead of the dissimilarity index, recovering very similar estimates. Our find- ings, furthermore, do not hinge on the particular cutos that define subgroups on residential segregation and district size. For instance, when we define the high subgroups according to the top third rather than the top half of treated units, the pattern of results is unchanged, and the treatment eects increase in magnitude (see Appendix Figures B.6 and B.7). Recent methodological work identifies two potential pitfalls of interpreting coecients from a multiplicative interaction model such as the one we use for our two-way fixed eects specification (Hainmueller, Mummolo and Xu 2018). First, the model relies on a linearity assumption: the interaction eect is assumed to change at a constant rate with the mod- erator (in this case, Latino VEP). In Appendix Figures B.8-B.9, we relax this assumption, instead estimating a non-interacted model in each of three bins constructed based on the distribution of Latino VEP among treated units. We find that the eects of conversion are indeed generally increasing with Latino VEP, and statistically significantly dierent from one another; thus, our conclusions do not crucially depend on the linearity assumption. A 25 79 Author Manuscript This article is protected by copyright. All rights reserved. second potential danger is a lack of common support of the moderator across treatment and control groups, which can result in overextrapolation if the linearity assumption is not satisfied. To guard against this, we only report marginal eects in the region of Latino VEP where we can find both treated and control units—in practice, between the minimum and maximum values of Latino VEP in the treatment group. IV Analysis We isolate the causal relationship between voluntary conversions of school districts to ward voting and Latino political performance by utilizing an instrumental variables frame- work. We do this by treating the incidence of a legal action being initiated against a nearby school districtkc (within the same county c) as having no direct impact on the success of Latino candidates in districtic (satisfying the exclusion restriction) but as being a strong predictor of districtic’s decision to pursue its own electoral reform (satisfying the strong first stage requirement). Districtic might choose to convert after observing a legal threat being made against a neighboring school district for a number of reasons, the most salient being the fear of becoming a subsequent target for legal action. The CVRA was written in a way such that 1) litigation was virtually guaranteed to be successful in forcibly converting school districts, and 2) the cost of pursuing litigation would be essentially zero for the plainti due to the districts’ obligation to cover the plainti’s legal bills. Thus, for reform advocates, pursuing litigation was a win-win scenario: they were all but guaranteed success, their costs were covered, and the redistricting plan drawn up by the school district or county was subject to the courts’ approval before being implemented. On the other hand, if a district could manage to get ahead of what they (rightly) perceived as an impending wave of litigation, they could avoid all of the costs associated with legal action and retain control over their redistricting plans, so long as they satisfied outside legal observers. For those districts that might be especially fearful of a protracted, costly legal battle that could stoke backlash from white constituents, 26 80 Author Manuscript This article is protected by copyright. All rights reserved. the incentive to voluntarily and peaceably convert to ward elections designed to satisfy legal challengers would be particularly strong. As such, the eect of electoral reform on Latino representation in forcibly converted districts is likely to be considerably dierent than the eect in districts that voluntarily converted. Table 3 presents the results of the first and second stage regressions using the full sample of school districts (after removing districts that were ward for the entirety of our sample and districts that were directly threatened by civil rights groups via legal action), and the treatment eects are depicted visually in Figure 6. The instrument is a binary indicator of whether the school district resided in a county in which there occurred at least one legal action taken against another school district the year before or anytime prior, while the treatment is voluntary conversion to ward voting by non-targeted districts. The dependent variables remain the same three political outcomes from the previous section. The eect of voluntary conversion is striking. The first panel of Figure 6 shows that the number of school board seats won by Latinos increases 42 percentage points when the population is comprised of 40% voting-eligible Latinos. On average, this amounts to Latino candidates winning 29% of school board seats up for election under an at-large system versus Latino candidates winning 71% of seats under a ward system. This gain is explained both by an increase in Latino vote share and by the percentage of Latinos who run for election. Both increase by about 45 percentage points, going from 26% to 70% as a share of the vote on average and from 38% of elections having at least one Latino candidate to 83% of elections having at least one Latino candidate, on average. This positive impact of electoral reform becomes statistically significant when at least 40% of the voting eligible population is Latino. 27 81 Author Manuscript This article is protected by copyright. All rights reserved. Table 3: Legal Action Taken Against Nearby School District as an Instrument for Voluntary Conversion from At-Large to Ward Representation First Seats won Latino Latino stage by Latinos candidates vote share (1) (2) (3) (4) In-county legal threat 0.139 (0.029) Voluntary conversion -1.219 -0.524 -0.983 (0.666) (0.650) (0.567) Proportion Latino 0.364 0.548 0.350 (0.111) (0.123) (0.105) Voluntary * Proportion Latino 4.099 2.420 3.551 (1.883) (1.804) (1.650) Controls Yes Yes Yes Yes Observations 3,433 3,386 3,398 3,398 R2 0.036 0.211 0.390 0.333 F-statistic on instrument 130.6 Notes: Robust standard errors clustered by school district in parentheses. Estimated intercept and controls not reported.p<0.05;p<0.01;p<0.001 28 82 Author Manuscript This article is protected by copyright. All rights reserved. Figure 6: Marginal Eect of Voluntary Conversion to Ward Elections: Instrumental Variables Analysis 0 2 4 6 0.2 0.3 0.4 0.5 Distribution, instrument observations -0.5 0.0 0.5 1.0 0.2 0.3 0.4 0.5 Proportion Latino VEPMarginal effect, Prop. Latinos elected0 50 100 0.2 0.3 0.4 0.5 Distribution, control observations (a) Seats Won by Latinos 0 2 4 6 0.2 0.3 0.4 0.5 Distribution, instrument observations -0.5 0.0 0.5 1.0 0.2 0.3 0.4 0.5 Proportion Latino VEPMarginal effect, Prop. Latino candidates0 50 100 0.2 0.3 0.4 0.5 Distribution, control observations (b) Elections with at Least One Latino Candidate 0 2 4 6 0.2 0.3 0.4 0.5 Distribution, instrument observations -0.5 0.0 0.5 1.0 0.2 0.3 0.4 0.5 Proportion Latino VEPMarginal effect, Latino vote share0 50 100 0.2 0.3 0.4 0.5 Distribution, control observations (c) Latino Vote Share 29 83 Author Manuscript This article is protected by copyright. All rights reserved. 5 Conclusion The research presented in this paper showed, for the first time, that there is a causal link between electoral institutions and Latino political success. Specifically, the descriptive representation of Latinos is likely to improve significantly after moving from at-large to ward elections in school districts that are large and residentially segregated, and where the Latino population is of sucient size. When all of these conditions are met, the positive impact of reform is striking, exceeding one additional oceholder for every three available seats. This is in spite of the facts that: 1) the CVRA set a relatively low bar for demonstrating racially polarized voting, so the treated districts need not have had the most egregious representation gaps to begin with; 2) the federal Voting Rights Act predated the CVRA in proactively pursuing instances of minority vote dilution, so it already eliminated the most egregious oenders; 3) Latinos had to be politically mobilized and to run eective candidates to take advantage of the electoral reforms; and 4) most instances of electoral reform under the CVRA took place as voluntary conversions in which the school board or county—and not the state—controlled the redistricting maps. Given these countervailing conditions, it is impressive to see the size and significance of these contingent eects of electoral reform under the CVRA. On the other hand, when these important conditions are not in place, moving from at- large to ward elections can actually have null, or even negative, eects on Latino descriptive representation, with the additional downsides of imposing expensive legal fees, large trans- action costs, and often divisive political conflict on already overburdened districts. The presence of these contingencies may help explain why a large and active academic literature on the subject has produced so many conflicting findings, and opens avenues for future work. While our finding on the moderating eect of residential segregation is entirely consistent with theoretical predictions based on how the respective institutions aggregate votes, more research is needed to understand the mechanisms driving our novel finding on district size. Political networks, resources, and the sheer availability of viable candidates are all plausible 30 84 Author Manuscript This article is protected by copyright. All rights reserved. explanations that should be investigated further. Our findings also highlight that when undertaking electoral reform, process matters for outcomes. Whereas our analysis of legally mandated conversions uncovered positive and negative eects moderated by key conditions, districts that converted voluntarily saw un- conditionally positive eects at every step of the electoral process, from candidacy to vote share to the ultimate outcome of Latino oceholding. While one might expect, a priori, that districts that select into electoral reform are already more concerned about minority representation and would therefore exhibit smaller treatment eects, it seems instead that buy-in from institutional actors contributed dramatically to the reform’s eectiveness. Taken together, we hope our findings can inform the best way forward for reformers who aim to increase the voice of minorities in the political process. A useful takeaway is that, at least on the dimension of segregation, the federal VRA seems to have gotten it right. Though the CVRA was written with the best intentions in mind, it caused a number of low-segregation school districts to convert to ward elections that would not have done so under the federal standard, and they may have done so needlessly, and at great cost to them. Additionally, both the CVRA and the federal VRA miss one of the crucial conditions necessary for improvements in representation, independent of segregation: the size of the district. Even in places that are not particularly segregated, greater enrollment strongly predicts the likelihood of successful electoral reform, an insight that may extend to other forms of local government such as city councils and municipal boards. Overall, our findings on these contingencies and on the success of reforms undertaken voluntarily suggest that, rather than a broad, sweeping legal remedy, states should consider oering information, demographic analysis, and technical assistance for conversion targeted to precisely those districts where reform is likely to have the largest impact. Finally, our work has shed light on the limitations of institutional change, and on the persistent barriers to minority political access that require even greater investments than interventions such as the CVRA. Changing the way that votes are aggregated into seats may 31 85 Author Manuscript This article is protected by copyright. All rights reserved. lead to some easy representational gains under certain conditions, but it does not change more fundamental realities like the propensity of minority candidates to run for oce, the resources at their disposal when they do, and the levels of voter turnout and mobilization. As we show, districts that were able to take full advantage of the CVRA had relatively low levels of poverty and unemployment and high levels of English language proficiency and education, and reformers should target these crucial constraining factors alongside institutional change. There is still much work to be done to understand the relationship between electoral institutions and substantive representation. We did not find any eect of conversion on a number of educational, achievement, or finance outcomes (not reported). At least part of this finding may be a result of the relatively short period of policy implementation that we can possibly study at this time. It is likely that the graduation and drop-out rates of Latino students, for example, are slow-moving variables that benefit from policy changes only over the long-run. In future years, we should be able to better gauge the long-term policy consequences of improved descriptive representation on local school boards. 32 86 Author Manuscript This article is protected by copyright. All rights reserved. References Ananat, Elizabeth Oltmans. 2011. “The Wrong Side(s) of the Tracks: The Causal Eects of Racial Segregation on Urban Poverty and Inequality.”American Economic Journal: Applied Economics 3(2):34–66. Browning, Rufus P., Dale R. Marshall and David H. Tabb. 1986.Protest Is Not Enough: The Struggle of Blacks and Hispanics for Equality in Urban Politics. Berkeley: University of California Press. Browning, Rufus P., Dale R. Marshall and David H. Tabb. 2003.Racial Politics in American Cities. New York: Longman. Bullock, Charles S. and Susan A. MacManus. 1987. “Staggered Terms and Black Represen- tation.”The Journal of Politics 49(2):543. California Common Cause. 2017. “California Municipal Democracy In- dex.” Available at: https://www.commoncause.org/california/resource/ california-municipal-democracy-index/ (last accessed 09/30/19). Clark, Alistair and Timothy B. Krebs. 2012. Elections and Policy Responsiveness. In The Oxford Handbook of Urban Politics, ed. P. John, K. Mossberger and Susan E. Clarke. Oxford University Press. Cole, Leonard A. 1974. “Electing Blacks To Municipal Oce. Structural and Social Deter- minants.”Urban Aairs Review 10(1):17–39. Collins, William J. and Robert A. Margo. 2000. “Residential Segregation and Socioeconomic Outcomes: When Did Ghettos Go Bad?”Economics Letters 69(2):239–43. Cutler, David M. and Edward L. Glaeser. 1997. “Are Ghettos Good or Bad?”The Quarterly Journal of Economics 112(3):827–72. 33 87 Author Manuscript This article is protected by copyright. All rights reserved. Cutler, David M., Edward L. Glaeser and Jacob L. Vigdor. 1999. “The Rise and Decline of the American Ghetto.”Journal of Political Economy 107(3):455–506. Davidson, Chandler and Bernard Grofman. 1994.Quiet Revolution in the South: The Impact of the Voting Rights Act 1965–1990. Princeton, NJ: Princeton University Press. Davidson, Chandler and George Korbel. 1981. “At-Large Elections and Minority-Group Rep- resentation: A Re-Examination of Historical and Contemporary Evidence.”The Journal of Politics 43(4):982–1005. Dovi, Suzanne. 2002. “Preferable Descriptive Representatives: Will Just Any Woman, Black, or Latino Do?”American Political Science Review 96(4):729–43. Engstrom, Richard L. and Michael D. McDonald. 1981. “The Election of Blacks to City Councils: Clarifying the Impact of Electoral Arrangements on the Seats/Population Re- lationship.”The American Political Science Review 75(2):344–54. Epstein, David and Sharyn O’Halloran. 1999. “Measuring the Electoral and Policy Impact of Majority-Minority Voting Districts.”American Journal of Political Science 43(2). Fraga, Luis Ricardo. 2009. “Interests and Representation: Ethnic Advocacy on California School Boards.”Teachers College Record 111(3):659. Fraga, Luis Ricardo, Kenneth J. Meier and Robert E. England. 1986. “Hispanic Americans and Educational Policy: Limits to Equal Access.”The Journal of Politics 48(4):850–76. Haider-Markel, Donald P., Mark R. Joslyn and Chad J. Kniss. 2000. “Minority Group Interests and Political Representation: Gay Elected Ocials in the Policy Process.”The Journal of Politics 62(2):568–77. Hainmueller, Jens, Jonathan Mummolo and Yiqing Xu. 2018. “How Much Should We Trust Estimates from Multiplicative Interaction Models? Simple Tools to Improve Empirical Practice.”Political Analysis 27(2):163–192. 34 88 Author Manuscript This article is protected by copyright. All rights reserved. Imai, Kosuke and Kabir Khanna. 2017. “wru: Who Are You? Bayesian Predictions of Racial Category Using Surname and Geolocation.” Available at: https://github.com/ kosukeimai/wru (last accessed 09/30/19). Karnig, Albert K. and Susan Welch. 1982. “Electoral Structure and Black Representation on City Councils.”Social Science Quarterly 63(1):99–114. Leal, David L., Valerie Martinez-Ebers and Kenneth J. Meier. 2004. “The Politics of Latino Education: The Biases of At-Large Elections.”The Journal of Politics 66(4):1224–44. MacManus, Susan A. 1978. “City Council Election Procedures and Minority Representation: Are They Related?”Social Science Quarterly 59(1):153–61. Mansbridge, Jane. 1999. “Should Blacks Represent Blacks and Women Represent Women? A Contingent “Yes”.”The Journal of Politics 61(3):628–57. Marschall, Melissa J. and Anirudh V. S. Ruhil. 2007. “Substantive Symbols: The Attitudinal Dimension of Black Political Incorporation in Local Government.”American Journal of Political Science 51(1):17–33. Marschall, Melissa J., Anirudh V. S. Ruhil and Paru R. Shah. 2010. “The New Racial Cal- culus: Electoral Institutions and Black Representation in Local Legislatures.”American Journal of Political Science 54(1):107–24. Massey, Douglas S. and Nancy A. Denton. 1993.American Apartheid: Segregation and the Making of the Underclass. Cambridge, MA: Harvard University Press. Meier, Kenneth J. and Amanda Rutherford. 2014. “Partisanship, Structure, and Repre- sentation: The Puzzle of African American Education Politics.”The American Political Science Review 108(2):265–80. 35 89 Author Manuscript This article is protected by copyright. All rights reserved. Meier, Kenneth J., Eric Gonzalez Juenke, Robert D. Wrinkle and J.L. Polinard. 2005. “Struc- tural Choices and Representational Biases: The Post-Election Color of Representation.” American Journal of Political Science 49(4):758–68. Meier, Kenneth J., Joseph Stewart and Robert E. England. 1991. “The Politics of Bureau- cratic Discretion: Educational Access as an Urban Service.”American Journal of Political Science 35(1):155–77. Molina Jr., Angel Luis and Kenneth J. Meier. 2016. “Demographic Dreams, Institutional Realities: Election Design and Latino Representation in American Education.”Politics, Groups, and Identities pp. 1–18. Moncrief, Gary F. and Joel A. Thompson. 1992. “Electoral Structure and State Legislative Representation: A Research Note.”The Journal of Politics 54(1):246–56. Phillips, Anne. 1991.Engendering Democracy. Cambridge: Polity Press. Phillips, Anne. 1995.The Politics of Presence. Oxford Political Theory. New York: Claren- don Press. Phillips, Anne, ed. 1998.Feminism and Politics. Oxford Readings in Feminism. New York: Oxford University Press. Pitkin, Hanna Fenichel. 1967.The Concept of Representation.Berkeley: University of California Press. Polinard, Jerry L., ed. 1994.Electoral Structure and Urban Policy: The Impact on Mex- ican American Communities. Bureaucracies, Public Administration, and Public Policy. Armonk, NY: M.E. Sharpe. Robinson, Ted and Robert E. England. 1981. “Black Representation on Central City School Boards Revisited.”Social Science Quarterly 62(3):495–502. 36 90 Author Manuscript This article is protected by copyright. All rights reserved. Sass, Tim R. 2000. “The Determinants of Hispanic Representation in Municipal Govern- ment.”Southern Economic Journal 66(3):609–30. Stewart, Joseph, Robert England and Kenneth J. Meier. 1989. “Black Representation in Urban School Districts: From School Board to Oce to Classroom.”Western Political Quarterly 42(2):287–305. Trounstine, Jessica and Melody E. Valdini. 2008. “The Context Matters: The Eects of Single-Member versus At-Large Districts on City Council Diversity.”American Journal of Political Science 52(3):554–69. Wald, Kenneth D., James W. Button and Barbara A. Rienzo. 1996. “The Politics of Gay Rights in American Communities: Explaining Antidiscrimination Ordinances and Poli- cies.”American Journal of Political Science 40(4):1152–78. Welch, Susan. 1990. “The Impact of At-Large Elections on the Representation of Blacks and Hispanics.”The Journal of Politics 52(4):1050–76. Welch, Susan and Albert K. Karnig. 1978. “Representation of Blacks on Big City School Boards.”Social Science Quarterly 59(1):162–72. 37 91 by Jessica Trounstine Department of Political Science Vanderbilt University 2024 Edition DEMOCRACY REFORM PRIMER SERIES A Practical Research Guide District vs At-Large Elections 92 The University of Chicago Center for Effective Government was founded in 2019 at the University of Chicago’s Harris School of Public Policy to help solve the problems of government ineffectiveness with a multi-faceted theory of action. The Center organizes its work and activities around three key areas—ideas, education, and engagement—and builds bridges across differences between scholars, students, practitioners, leaders, journalists, and advocates. Through robust, innovative programming, the Center works to strengthen institutions of democracy and improve government’s capacity to solve public problems. About the Center for Effective Government About the Democracy Reform Primer Series Narrowing the gap between research and public dialogue, the University of Chicago Center for Effective Government's Democracy Reform Primers responsibly advance conversations and strategy about proposed changes to our political institutions. Each Primer focuses on a particular reform, clarifies its intended purposes, and critically evaluates what the best available research has to say about it. The Primers do not serve as a platform for either authors or the Center to advance their own independent views about the reform; to the contrary, they serve as an objective and authoritative guide about what we actually know—and what we still don’t know—about the likely effects of adopting prominent reforms to our political institutions. In some instances, the available evidence may clearly support the claims of a reform’s advocates. In other instances, it may cut against them. And in still others, the scholarly literature may be mixed, indeterminate, or altogether silent. Without partisan judgment or ideological pretense, and grounded in objective scholarship, these Primers set the record straight by clarifying what can be said about democracy reforms with confidence and what requires further study. 193 Jessica Trounstine earned her Ph.D. in Political Science from UC San Diego in 2004 and is the Centennial Chair and Professor of Political Science at Vanderbilt University. She previously served as the Foundation Board of Trustees Presidential Chair of Political Science at UC Merced and as assistant professor of politics and policy at Princeton University. She is the author of two award-winning books, Segregation by Design: Local Politics and Inequality in American Cities (Cambridge University Press) and Political Monopolies in American Cities: The Rise and Fall of Bosses and Reformers (University of Chicago Press), and numerous articles and book chapters. Professor Trounstine's work studies the process and quality of representation in American democracy. She is focused on the ways in which formal and informal local political institutions generate inequalities. Professor Trounstine's scholarship is mixed-method; reliant on historical analysis, case studies, experiments, and large-n quantitative analyses. She has served as a consultant for the U.S. Department of Justice, city governments, and various community organizations; and serves on numerous editorial and foundation boards. About the Author CEG Faculty Affiliate Anthony Fowler is a Professor at the Harris School of Public Policy at the University of Chicago. His research applies econometric methods for causal inference to questions in political science, with particular emphasis on elections and political representation. Fowler is currently the Co-editor in Chief of the Quarterly Journal of Political Science, and the co-author (with Ethan Bueno de Mesquita) of Thinking Clearly with Data: A Guide to Quantitative Reasoning and Analysis (Princeton University Press, 2021). Fowler earned his Ph.D. in government from Harvard University and completed undergraduate studies at the Massachusetts Institute of Technology. About the Series Editor 294 Promise of the Reform At-large elections are discriminatory1. Adopting district elections will improve descriptive representation for marginalized groups, produce councilors who are closer to voters, and generate political outcomes that are more likely to address the needs of neighborhoods. Summary 395 •District elections increase descriptive representation for racial/ethnic minorities when the group is moderately sized and segregated. •District elections can allow neighborhoods to protect their interests. •At-large elections encourage councils to consider a citywide perspective. Key Takeaways from the Research •Are local legislative districts subject to the problems of gerrymandering that have been identified at higher levels of government? •Does election type have policy consequences for local governments? •Are marginalized communities’ substantive interests best served by at-large or district elections? Important Questions We Need Better Answers To 496 Local governments in the United States feature a wide variety of governing institutions. Some cities have elected mayors, others do not. Some cities hold elections concurrently with state and national elections, others do not. This brief explores another type of variation – the way in which local legislators are elected. Cities in the United States tend to elect their city councils using two electoral system types: single member districts or at-large elections. When councilors are elected by district, the city is divided into geographic areas of roughly equal population size that elect a single member to the city council in a plurality or majoritarian contest. An at-large system is one in which members of the city council are selected by the entire city electorate. In most cities this means that voters are offered a slate of candidates and are allowed to vote for the same number of candidates as there are seats available. In some cases, cities allow casting multiple votes for the same candidate (e.g., if three council seats are up for election, the voter can cast three votes for whichever candidates they prefer). Cities often have majoritarian requirements such that if a candidate does not receive 50% of the vote, she is forced into a run-off election. Some cities designate seats or residency requirements for at-large positions turning the election into a series of single member “ The adoption of at-large systems typically flowed from efforts to dilute Black voters’ power in Southern cities after the Civil War and the municipal reform movement that sought to promote a citywide perspective in governance.” contests, while other cities vote for only one at-large member in any given election. A small but growing number of cities use mixed-systems, electing some council members by district and others at-large. Most research on the effects of district/at-large elections is focused on general purpose local governments, e.g., cities. Cities (along with towns, villages, and boroughs) are incorporated places that are governed by elected officials, and typically have the power to raise and spend revenue, regulate the uses of land, and enforce the law. Cities are chartered by the state in which they are located, and the rules governing incorporation vary from state to state. The determination of how legislators are elected occurs when cities are incorporated. Many smaller cities are general law cities. Their governance structure, powers, and functions are set by state statute. Most mid- sized and larger cities are governed by a constitution that is approved (and amended) by the voters of the city. These charter cities have home-rule. That is, charter cities are free to enact laws without attaining state permission. They have full authority to set their own tax rates and decide how and where to provide services. Importantly, they are able to determine various institutional features, such as how the city council is elected. Charter cities can and do change the way that their legislators are elected. Introduction 597 Introduction “ The number of cities using pure at-large systems has declined over the past 20 years, but at-large elections remain a common feature in city politics.” Up until the late 1800s, cities almost universally had district (or ward) elections.2 The adoption of at-large systems typically flowed from two sources: efforts to dilute Black voters’ power in Southern cities after the Civil War and the municipal reform movement that sought to promote a citywide perspective in governance and limit the power of parochial neighborhood concerns.3 According to one commentator, “It is contended that at best [the district system] lends itself to the election of members who will pay more attention to the needs of their district than to the larger interests of the city as a whole; and that the concentration of the worst elements of the city’s population in some wards makes inevitable the election of a number of very objectionable members.”4 At-large systems spread rapidly throughout the United States in the first decades of the 20th century. The number of cities using pure at-large systems has declined over the past 20 years, but at-large elections remain a common feature in city politics. The passage of the 1965 Voting Rights Act (and the 2002 California Voting Rights Act) laid the groundwork for legal contestation of at-large systems. Several court cases, including the landmark Thornburg v. Gingles,5 have found at-large systems to discriminate against Black voters by diluting their voting power under certain conditions. Voting must be racially polarized (where different racial groups vote cohesively for different candidates) and the racial group that is in the minority must be sufficiently large and compact to plausibly constitute a majority of voters in a single district. Over the last several decades, the court has required many legislative bodies to adopt district elections when these conditions are met. Nonetheless, today about 68% of cities elect their members at-large.6 The way in which members of the local legislature are elected affects who wins election to office, who is represented in the political system, and what policies are made. Generally, at-large elections shift representation toward voters rather than residents. In a districted system, regardless of the level of turnout in an area, the area receives representation on the council. In at-large systems, the voting majority selects the winners. Theoretical work by one group of researchers showed that at-large systems will be used as a disenfranchising mechanism when the minority group is small.7 When the minority group is closer to 50% of the population, the majority will prefer gerrymandered districts to maintain power. 698 Who Wins One of the most persistent findings by scholars of urban politics is that single member district elections increase descriptive representation of historically underrepresented racial and ethnic groups on city councils.8 This effect has been found to be particularly strong for African Americans.9 Districts have also been found to be beneficial to Latinos.10 These statistical findings have been supported by extensive case study and historical research as well.11 In sum, “the effect of …districts is unequivocally…greater equity.”12 These consequences typically result when the minority group in question is large enough and sufficiently concentrated (e.g., segregated) to elect representatives in a districted system.13 However, one study also showed that at-large systems disadvantage Black candidates because they are more cognitively taxing for voters.14 “ Single member district elections increase descriptive representation of historically underrepresented racial and ethnic groups on city councils.” The literature on the representation of women finds mixed effects for single member districts. Most research has concluded that districts are either meaningless15 or disadvantageous for women candidates.16 However, one paper found that women benefit from district elections. It argued that this is because at-large elections are typically more competitive and require more campaign funds and larger mobilization efforts for candidates.17 Scholars have also sought to understand how electoral institutions affect women of color given conflicting institutional effects. Generally, scholars find that Black men, but not Black women are advantaged by districts.18 799 Policy Consequences “ Increases in minority representation leads to higher test scores for minority students, likely because of increased funding and the hiring of more minority principals.” Furthermore, there is evidence that electing minorities in district elections produces better substantive outcomes for minority residents compared to legislatures in which minorities are elected to the council in at-large elections.24 More specifically, a study found that in at-large systems increasing the number of Black or Latino school board members has no effect on the hiring of black administrators but find a powerful effect in districted systems. The paper argued that the explanation for the difference is that in the at-large setting Black and Latino administrators must attend to the preferences of the median (usually white) voter and as a result are more constrained in producing benefits for minority communities. In districted systems, where logrolling frequently prevails, minority communities appear to be better served. This is contrary to theoretical findings by another researcher, who demonstrated that policy outcomes are more likely to favor minority preferences under at-large systems, even though districts are more likely to produce descriptive representation.25 Holding constant Black representation, another study found that at-large systems produce better representation for Black residents.26 Because district systems tend to elect more officials of color, they also tend to make policy that is more responsive to minority interests. Higher proportions of minority councilors are associated with increases in minority appointments to the bureaucracy19 and increased focus and/or spending on minority policy issues.20 Several scholars find that Black school board members are associated with more Black administrators and teachers; and that a higher proportion of Black teachers is associated with a higher proportion of Black students in gifted classes, as well as higher grades and test scores among Black students.21 Others produce similar findings for Latinos.22 One study found that increases in minority representation leads to higher test scores for minority students, likely because of increased funding and the hiring of more minority principals.23 8100 These results echo findings in the Congressional politics literature that explores the policy consequences of majority-minority districts and shows that such districts can undermine minority interests.27 This kind of outcome is produced when the minority community is packed into a small number of districts such that their representatives are typically on the losing side of legislative votes. As a result, some scholars argue that minority interests will be better served when minority voters are able to impact the election of a larger number of legislators, even if this means they will not be descriptively represented.28 We know virtually nothing about the present state of gerrymandering/district line drawing in local elections. Most municipalities hold non-partisan elections and those with partisan elections are frequently dominated by a single party. This means that it is unlikely that partisan gerrymandering shapes decisions about district lines. Given that district maps are adopted by the sitting legislature, it is likely that any bias in line drawing will tend to advantage incumbents and entrench existing hierarchies along race and class lines. So, it may be that the lessons learned from analyses of state and federal policy outcomes will be relevant at the local level. But new research is required to determine whether or not this is the case. “ Electing minorities in district elections produces better substantive outcomes for minority residents compared to legislatures in which minorities are elected to the council in at-large elections.” Policy Consequences 9101 “ At-large elections tend to promote a citywide perspective on policymaking, as opposed to a neighborhood perspective.” “ Changing from at-large to district systems depresses the total amount of housing that cities build.” What we do have evidence of in local politics is the degree to which neighborhood versus citywide interests are represented under different election systems. At-large elections tend to promote a citywide perspective on policymaking, as opposed to a neighborhood perspective. One study found that district councils are more factionalized, with more geographically focused divisions and councilors playing a greater ombudsman role compared to at- large councils.29 Another paper found that changing from at-large to district systems depresses the total amount of housing that cities build. Anti-development preferences are generally widely shared, but minority neighborhoods which tend to lack political power in at-large systems, are often saddled with new development. The implementation of districts, and thus a greater ability to dictate policy outcomes, means that minority neighborhoods are better able to resist development – and so, overall new housing declines.30 Researchers found that districted cities tend to have larger budgets (spending, debt, and taxation), likely as a result of legislator deference to neighborhood priorities (e.g., log-rolling). Policy Consequences However, another paper found that districts do not always have this effect. It showed that neighborhoods were routinely disregarded in favor of citywide needs in controversial decisions in one districted city (Los Angeles).31 Another piece of research also found that election type does not affect aggregate policy outcomes including per capita taxes and the share of taxes raised by sales tax.32 Thus, it remains something of an open question whether election type has a causal effect on policy outcomes, and for whom these outcomes are beneficial. 10102 Conclusion It is clear that district elections increase the likelihood that Black and Latino legislators will be elected when the population of Black and Latino residents is large enough to elect a councilor in a single member district and is residentially segregated. “ Having more Black and Latino legislators is correlated with city policy that reflects the interests of Black and Latino residents.” The median city council has six members, and so, this means that Black and Latino populations must typically constitute at least 10% of the population and live in neighborhoods with heavy concentrations of Black/Latino residents to see an effect from districts. Women appear to have a slight advantage in district elections as well, likely because of the lower cost of campaigning and lower levels of competitiveness in elections that cover a limited geographic area. Scholars have also provided evidence of an association between descriptive representation and substantive representation. That is, having more Black and Latino legislators is correlated with city policy that reflects the interests of Black and Latino residents. However, we do not have clear causal evidence that election type (district vs. at-large elections) generates policies that are preferred by marginalized communities. We have stronger evidence that districts do offer the opportunity for neighborhoods to protect their interests. Generally though, more research is needed on the effects of election type on policy outcomes and whose interests are served by these outcomes. Finally, it remains uncertain whether municipal legislative districts suffer from the problems of gerrymandering that plague district line-drawing at higher levels of government. 11103 1 NAACPLF, “At-large voting frequently asked questions,” 2024 https://www.naacpldf.org/wp- content/uploads/At-Large-Voting-Frequently- Asked-Questions-1.pdf 2 Fairlie, John, “American Municipal Councils,” Political Science Quarterly, 19(2) (1904):234-251 https://doi. org/10.2307/2140282 3 Liazos, Ariane, Reforming the City: the Contested Origins of Urban Government, 1890-1930 (New York, Columbia University Press, 2020) 4 Fairlie, American Municipal Councils 5 Thornburg v. Gingles: 478 U.S. 30, 1986 6 International City/County Management Association, “2018 Municipal Form of Government Survey, Summary of Results,” 2018 https://icma.org/sites/ default/files/2018%20Municipal%20Form%20of%20 Government%20Survey%20Report.pdf 7 Trebbi, Francesco, Philippe Aghion, and Alberto Alesina, “Electoral Rules and Minority Representation in U.S. Cities,” Quarterly Journal of Economics, 123(1) (2008): 325-357 8 Bullock, Charles S., III, and Susan A. MacManus, “Testing Assumptions of the Totality-of-the- Circumstances Test: An Analysis of the Impact of Structures on Black Descriptive Representation,” American Politics Quarterly 21, (3) (1993): 290–306 Karnig, Albert, and Susan Welch, “Sex and the Ethnic Differences in Municipal Representation.” Social Science Quarterly 60, (3) (1979): 465–81 Mladenka, Kenneth R., “Barriers to Hispanic Employment Success in 1200 Cities. Social Science Quarterly 70, (2) (1989): 391–407 9 Arrington, Theodore, and Thomas Gill Watts, “The Election of Blacks to School Boards in North Carolina,” The Western Political Science Quarterly 44, (4) (1991): 1099–1105 Davidson, Chandler, and Bernard Grofman, eds., Quiet Revolution in the South: The Impact of the Voting Rights Act, 1965–1990 (Princeton, Princeton University Press, 1994) Welch, Susan, “The Impact of At-Large Elections on the Representation of Blacks and Hispanics,” Journal of Politics 52 (1990): 1050–76 10 Bullock, Charles, and Susan MacManus, “Structural Features of Municipalities and the Incidence of Hispanic Councilmembers,” Social Science Quarterly 71, (4) (1990): 665–81 Collingwood, Loren and Sean Long, “Can Districts Promote Minority Representation? Assessing the Effects of the California Voting Rights Act,” Urban Affairs Review 57, (3) (2021) Davidson, Chandler, and George Korbel, “At-Large Elections and Minority-Group Representation: A Re-Examination of Historical and Contemporary Evidence,” The Journal of Politics 43, (4) (1981): 982–1005 Heilig, Peggy, and Robert Mundt, “Changes in Representational Equity: The Effect of Adopting Districts,” Social Science Quarterly 64, (2) (1983): 393–97 Leal, David, Valerie Martinez-Ebers, and Kenneth J. Meier, “The Politics of Latino Education: The Biases of At-Large Elections,” The Journal of Politics 66, (4) (2004): 1224–44 Polinard, Jerry, Robert Wrinkle, and Tomas Longoria, “The Impact of District Elections on the Mexican American Community: The Electoral Perspective,” Social Science Quarterly 72, (3) (1991): 608–14 Taebel, Delbert, “Minority Representation on City Councils: Impact of Structure on Blacks and Hispanics,” Social Science Quarterly 59, (1) (1978): 142–52 11 Bridges, Amy, Morning Glories: Municipal Reform in the Southwest (Princeton, Princeton University Press, 1997) Rice, Bradley Robert, Progressive Cities (Austin, University of Texas Press, 1977) Endnotes 12104 12 Mundt, Robert, and Peggy Heilig, “District Representation: Demands and Effects in the Urban South,” The Journal of Politics 44, (4) (1982): 1035–48 13 Abott, Carolyn and Asya Magazinnik, “At-Large Elections and Minority Representation in Local Government,” American Journal of Political Science 64, (3) (2020): 717-33 Marschall, Melissa J., Anirudh V. S. Ruhil, and Paru R. Shah, “The New Racial Calculus: Electoral Institutions and Black Representation in Local Legislatures,” American Journal of Political Science 54, (1) (2010): 107–24. Trounstine, J. and M. E. Valdini, “The context matters: The effects of single-member versus at- large districts on city council diversity,” American Journal of Political Science 52, (3) (2008): 554-69 14 Crowder-Meyer, M., Gadarian, S. K., Trounstine, J., & Vue, K., “A different kind of disadvantage: Candidate race, cognitive complexity, and voter choice,” Political Behavior, 42(2) (2020), 509–530 https://doi.org/10.1007/s11109-018-9505-1 15 Alozie NO, Manganaro LL., “Womens’ council representation - measurement implications for public-policy,” Political Research Quarterly 46, (2) (1993): 383-98 Bullock CS, MacManus SA., “Municipal electoral structure and the election of councilwomen,” The Journal of Politics 53, (1) (1991): 75-89 16 Darcy, Robert, Susan Welch, and Janet Clark, “Women Candidates in Single-Member and Multi Member Districts: American State Legislative Races,” Social Science Quarterly, 66 (4) (1985): 945–53 Hogan, Robert E., “The Influence of State and District Conditions on the Representation of Women in State Legislatures,” American Politics Research, 29 (2001): 4–24. King, James D., “Single-Member Districts and the Representation of Women in American State Legislatures: The Effects of Electoral System Change,” State Politics and Policy Quarterly 2 (2002): 161–75 Matland, Richard E., “How The Electoral System Has Helped Women Close the Representation Gap in Norway," In Closing the Gap: Women in Nordic Politics, ed. Lauri Karvonen and Per Selle (Dartmouth Press, London, 1995) Matland, Richard E., and Deborah Dwight Brown, “District Magnitude’s Effect on Female Representation in U.S. State Legislatures,” Legislative Studies Quarterly 17, (4) (1992): 469–92 Norris, Pippa, Politics and Sexual Equality (Boulder, Reinner, 1985) Rule, Wilma, “Parliaments of, by, and for the People: Except for Women?” In Electoral Systems in Comparative Perspective: Their Impact on Women and Minorities, ed. Wilma Rule and Joseph F. Zimmerman (Westport, CT, Greenwood, 1994) 15-30 Schwindt-Bayer, Leslie A., and William Mishler, “An Integrated Model of Women’s Representation,” The Journal of Politics 67 (May) (2005): 407–28 Welch, Susan, and Donley Studlar, “Multimember Districts and the Representation of Women: Evidence from Britain and the United States,” The Journal of Politics 52, (2) (1990): 391–412 Endnotes 13105 Endnotes 17 Crowder-Meyer, M., Gadarian, S. K., & Trounstine, J., “Electoral institutions, gender stereotypes, and women’s local representation,” Politics, Groups, and Identities, 3(2) (2015), 318–334 https://doi.org/10.10 80/21565503.2015.1031803 18 Darcy, Robert., Charles D. Hadley, and Jason F. Kirksey, “Election Systems and the Representation of Black Women in American State Legislatures.” Women and Politics, 13 (1993): 73–89 Herrick, Rebekah, and Susan Welch, “The Impact of At-Large Elections on the Representation of Black and White Women.” National Political Science Review 3 (1992): 62–77 Karnig, Albert, and Susan Welch, “Sex and the Ethnic Differences in Municipal Representation,” 1979 Rule, Wilma. 1992, “Multimember Legislative Districts: Minority and Anglo Women’s and Men’s Recruitment Opportunity,” In United States Electoral Systems: Their Impact on Women and Minorities, ed. Wilma Rule and Joseph F. Zimmerman (New York, Greenwood Press, 1992) 57-72 Trounstine and Valdini, “The context matters,” 2008 19 Kerr B and Mladenka KR., “Does politics matter- --a time-series analysis of minority employment patterns,” American Journal of Political Science, 38(4) (1994): 918-43 Mladenka KR., “Blacks and Hispanics in urban politics,” American Political Science Review, 83(1) (1989): 165-91 Sass, Tim and Stephen Mehay, “Minority Representation, Election Method and Policy Influence,” Economics and Politics, 15 (3) (2003): 225-341 20 Karnig, AK and Susan Welch, “Black Representatives and Urban Policy” (Chicago, University of Chicago Press, 1980) 21 Meier KJ, and England RE., “Black representation and educational-policy--- Are they related?” American Political Science Review, 78(2) (1984): 392—403 Stewart J, England RE, and Meier KJ., “Black representation in urban school districts---from school board to office to classroom,” The Western Political Quarterly, 42(2) (1989): 287--305 22 Leal DL, Martinez-Ebers V, and Meier KJ., “The politics of Latino education: the biases of at-large elections,” The Journal of Politics, 66(4) (2004): 1224—44 Polinard JL., Wrinkle RD, Longoria T, and Binder N., Electoral Structure and Urban Policy: The Impact on Mexican American Communities (Armonk, NY, M.E. Sharpe, 1994) 23 Kogan, Vladimir, Stephane Lavertu, and Zachary Peskowitz, “How Does Minority Political Representation Affect School District Administration and Student Outcomes?” American Journal of Political Science, 65(3) (2021): 699-716 24 Meier KJ, Juenke EG, Wrinkle RD, and Polinard JL., “Structural choices and representational biases: the postelection color of representation,” American Journal of Political Science, 49(4) (2005):758--68 25 Wittman, Donald, “District versus at-large voting: Why district voting results in worse policy for minorities,” European Journal of Political Economy 81 (2024) https://doi.org/10.1016/j. ejpoleco.2024.102502 14106 Endnotes 26 Sass and Mehay, “Minority Representation,” 2003 27 Brace, Kimball, Grofman, Bernard, and Handley, Lisa, “Does redistricting aimed to help blacks necessarily help Republicans?” The Journal of Politics, 49 (1987): 169–185 Epstein, David, and O’Halloran, Sharyn, “Measuring the electoral and policy impact of majority- minority voting districts,” American Journal of Political Science, 43 (1999): 367–395 Lublin, David Ian, The Paradox of Representation (Princeton University Press, Princeton, 1997) 28 Cameron, Charles, David Epstein, and Sharyn O’Halloran, “Do Majority-Minority Districts Maximize Substantive Black Representation in Congress?” American Political Science Review, 90(4) (1996): 794-812 Swain, Carol, Black Faces, Black Interests: The Representation of African Americans in Congress (Cambridge, MA, Harvard University Press, 1993) Thernstrom, Stephan and Abigail Thernstrom, America in Black and White: One Nation, Indivisible (Simon and Schuster, 1997) Wittman, “District versus at-large voting,” 2024 29 Bledsoe, Timothy and Susan Welch, Urban Reform and its Consequences: A Study in Representation (Chicago: University of Chicago Press, 1988) 30 Hankinson, Michael and Asya Magazinnik, “The Supply Equity Trade-Off: The Effect of Spatial Representation on the Local Housing Supply,” The Journal of Politics, 85(3) (2021): 1033-1047 31 Burnett, Craig and Vladimir Kogan, “Local Logrolling? Assessing the Impact of Legislative Districting in Los Angeles,” Urban Affairs Review 50(5) (2014) https://doi.org/10.1177/1078087414522408 32 Tausanovitch, Chris and Christopher Warshaw, “Representation in Municipal Government,” American Political Science Review, 108(3) (2014): 605-641 15107 effectivegov.uchicago.edu Scan QR code to view all Primers in the 2024 series 108 No. 14 The Trade-Offs between At-Large and Single-Member Districts October 2018 Cong Huang Hobby School of Public Affairs University of Houston chuang23@central.uh.edu Richard Murray Hobby School of Public Affairs Department of Political Science University of Houston rmurray@central.uh.edu Scott Hofer Hobby School of Public Affairs Department of Political Science University of Houston sjhofer@uh.edu 109 Hobby School of Public Affairs White Paper Series___________________________________________________ 1 The Trade-Offs between At-Large and Single-Member Districts Scott Hofer, University of Houston Cong Huang, University of Houston Richard Murray, University of Houston This report provides a review of the existing literature about the trade-offs between at- large and single-member districts at the local offices in the United States. Overall, at- large districts were the most common form of representation in most local offices until the 1960s and the passage of Voting Rights Act in 1965. Since the 1960s single-member districts have been the method of choice for most local elections because they enable smaller, geographically situated communities to send their own representatives to larger legislative assemblies. The history behind these two electoral systems corresponds with their respective pros and cons. In general, at-large elections are found to improve diversity in gender representation on city councils with more female councilors being elected. On the other hand, single-member districts benefit the representation of some racial minority groups, including African Americans and Latinos. But the positive potential depends on context: (1) the concentration, (2) the size, and (3) the polarization of the vote. Specifically, the advantages of single-member districts are minimal in a largely homogenous community or in a community where underrepresented groups are not concentrated geographically. African Americans are found to be overrepresented on school boards with at-large elections when African Americans occupy a smaller part of the population. Keywords: at-large districts, single-member districts, gender representation, racial representation. Background The literature on structures of elections focuses on local levels in the United States because there is little variation in representation structures in higher offices (Davidson 1979; Meier and Stewart 1991; Rocha 2007; Zax 1990). The United States does have single member (hereafter: SM) districts in the 43 states that have two or more members in the House of Representatives, while U.S. senators are elected at-large (hereafter AL) in all 50 states, as are almost all executive officers at the state level. That being the case, the best opportunity to study the effects on of electoral systems are in the thousands of local offices in the U.S., specifically school districts, county commissions and city councils. Overall, at-large districts were the most common form of representation in most local offices until the 1960s and the passage of Voting Rights Act (VRA) in 1965 (Davidson 1992). The history behind SM and AL districts systems correspond with their respective pros and cons. While context is key, the following section is a broad overview of the existing literature about the trade-offs in both systems. Then we move to the discussion of more nuanced findings pertaining to gender and racial representation in these two systems. 110 At-Large and Single-Member Districts 2 At-large elections have been employed when ruling majorities attempt to emphasize the corporate identity of particular jurisdictions and to suppress partisan or ethnic factionalism. The basic idea being that those elected to AL districts will be more likely to work toward the best result for the whole community rather than pander to the specific demands in parts of the community. Work in political science broadly illustrates that substantive representation is most common in AL systems for the wealthiest and most connected in the community (Enns and Wlezien 2011; Gilens and Page 2014; Meier et al. 2005). Additionally, AL systems have the benefit of increasing the diversity in gender representation with more women being elected in these systems (Trounstine and Valdini 2008). However, people of color are less likely to be elected in AL systems because the votes of racial minorities are diluted in elections that cover a broader area (Trounstine and Valdini 2008). The importance of AL systems for diverse cities has been the focus of lawsuits and VRA compliance. In more homogenous communities, the difference between AL and SM districts are less pronounced in terms of racial representation, while the overall trends of representation patterns along gender and economic lines remain. Since the 1960s SM districts have been the method of choice for most local elections because they enable smaller, geographically situated communities to send their own representatives to larger legislative assemblies. SM systems provide the benefits of localized democracy. In cases of city councils and school boards, elected members in SM systems might only represent a small neighborhood which allows legislators to be intimately aware of the issues of the local community. This allows the elected member to focus on the needs of their localized constituency rather than the interests of all. In diverse places, especially where diversity is in highly segregated communities, the SM systems promote diversity with increasing minority representation (Leal, Martinez-Ebers, and Meier 2004; Meier et al. 2005; Trounstine and Valdini 2008; Zax 1990). Yet, this diversity in racial representation is likely to be contrasted with the loss of gender representation and the tendency for fewer women to be elected in these systems. As noted, SM systems have become more popular in the United States after passage of the VRA as a mechanism to increase representation of racial minorities (Davidson 1992). In some cases, SM systems were implemented due to lawsuits such as Thornburg v. Gingles in 1986 alleging the AL systems in place unduly discriminated against cohesive groups of people of color to participate equally in the process by electing disproportionately white officials (Kosterlitz 1986). Finally, a small, but growing number of communities have incorporated a mixed approach that combines AL and SM systems. The Houston City Council is an example of these mixed bodies. While there is variation in the impacts of mixed system (they should be thought of as a continuum between AL and SM) the conclusion in the literature is that mixed systems typically provide benefits similar to SM districts. These mixed systems and modified AL systems provide descriptive representation similar to SM districts (Brockington et al. 1998; Karnig and Welch 1982; Welch 1990). Given the proportion of SM to AL districts (11 to 5) on the Houston City Council,1 it is especially likely to produce representation outcome similar to purely SM district systems. 1 https://www.houstontx.gov/council/. 111 Hobby School of Public Affairs White Paper Series___________________________________________________ 3 The Voting Rights Act and Electoral Structures In the United States, AL elections were popular for local elections; especially as a mechanism to ensure that a bloc-voting white majority could deny black citizens the opportunity to choose representatives of their choice in local governments. In 1965, mass politics changed the landscape of racial diversity and racial representation through the VRA. Language allowing judicial review of minority vote dilution efforts in places with a history of disenfranchising minority voters initially helped push communities away from AL systems. However, court rulings undermined this language in the late 1970s. By 1980, the courts had established that racial minorities must prove that a challenged election structure was designed or maintained intentionally to dilute their voting power. However, the passage of the VRA of 1982 changed this standard was from racial intent to vote dilution in practice, making minority lawsuits more likely to succeed. The 1986 Thornburg v. Gingle ruling created a quicker and easier process for providing a remedy for vote dilution, resulting in widespread changes from AL elections to SM elections, through both litigation and legislation (Davidson 1992; Kosterlitz 1986). Although the courts would later reverse course on some aspects of the VRA, the legacy of the rulings in the 1980s has become the status quo. Electoral Systems and Gender Representation One major focus in the relevant literature is on the role of gender representation in AL versus SM structures. Table 1 summarizes the trade-offs between these two systems regarding gender representation in particular. While SM districts are typically seen as a remedy to a lack of diversity, AL districts are the most likely to produce female elected members (Trounstine and Valdini 2008). Unlike racial diversity, gender diversity among the population is stable across geography. The existing scholarship on gender representation overwhelmingly suggests that SM districts either lead to fewer women being elected on city councils (Darcy, Welch, and Clark 1985; Hogan 2001; Matland 1995; Matland and Brown 1992; Rule 1994; Schwindt-Bayer and Mishler 2005; Welch and Studlar 1990) or no effect (Alozie and Manganero 1993; Bullock and MacManus 1991). This result may make more sense in terms of the non-exclusive relationship between race and gender (Githens and Prestage 1977). For instance, nuanced analysis of intersectionality found that black women tend to be advantaged by AL elections in cities while black men are disadvantaged by this structure (Darcy, Hadley, and Kirksey 1993; Rule 1992). However, there is no influence of SM versus AL districts for Mexican American women or men (Karnig and Welch 1979). A more recent study (2008) by Trounstine and Valdini that focused on more than 7,000 United States cities also found that the impact of either AL or SM district on gender representation is much more significant regarding white female and black male representation than was the case for Latinas or black women. 112 At-Large and Single-Member Districts 4 Table 1. The Trade-Offs between At-Large and Single-Member Districts – Gender Representation Electoral Systems Effects Literature Single-member districts (1) Have positive effect on the representation of African American men in city councils. Karnig and Welch (1979) (2) Have negative effect on the representation of African American women in city councils. Herrick and Welch (1992) (3) Have no effect on the representation of Mexican American women or men in city councils. Karnig and Welch (1979) At-large districts (1) Promote diversity in gender representation on city councils. Trounstine and Valdini (2008) (2) Benefit African American female candidates. Darcy, Hadley, and Kirksey (1993) (3) African American male candidates are disadvantaged by this structure. Rule (1992) Note: Regarding mixed systems, they are considered to provide descriptive representation similar to single- member districts (see e.g., Brockington et al.). Same for Table 2. Electoral Systems and Racial Representation The historical transformation at the local level from AL to SM has had the expected impact. Majority-minority districts became very popular in the 1990s and the representation of underrepresented groups began to improve (Cameron, Epstein, and O’Halloran 1996). Table 2 lists the trade-offs on racial representation specifically. Numerous studies put forward that SM districts have positive effect on the representation of some racial groups in city councils, including African Americans and Latinos (Arrington and Watts 1991; Bullock and MacManus 1990; Davidson and Grofman 1994; Polinard, Wrinkle, and Longoria 1991; Welch 1990). If equity in representation is the goal, in a diverse community, then SM districts are overwhelmingly cited as a better mechanism. 113 Hobby School of Public Affairs White Paper Series___________________________________________________ 5 Table 2. The Trade-Offs between At-Large and Single-Member Districts – Racial Representation Electoral Systems Effects Literature Single-member districts (1) Improve diversity in racial representation on city councils. Cameron, Epstein, and O’Halloran (1996) (2) Have positive effect on the representation of African Americans and Latinos in city councils. Davidson and Grofman (1994) (3) These effects are conditional on the context. The benefits are minimal in a largely homogenous community or in a community where underrepresented racial groups are not geographically concentrated. Trounstine and Valdini (2008) At-large districts (1) Lead to over-representation of minority voters where racial minority groups are geographically concentrated or where democratic voters make up a larger proportion of the population. Arrington and Watts (1991) (2) Lead to over-representation of African Americans on school boards where African Americans make up a smaller proportion of the population. Meier and Rutherford (2014) Despite the positive potential, the improved representation of historically underrepresented groups depends on context (Trounstine and Valdini 2008). The concentration (Sass 2000), size (Bullock and MacManus 1990; Leal, Martinez-Ebers, and Meier 2004), and polarization (Brace et al. 1988) of the vote are the key variables to consider. For instance, in a largely homogenous community or in a community where underrepresented groups are not concentrated geographically, the benefits of SM districts are minimal (Trounstine and Valdini 2008). Conversely, AL districts lead to over-representation of minority voters under some circumstances where Democratic voters make up a larger proportion of the population and racial minority groups are concentrated in the area (Arrington and Watts 1991). Similarly, African Americans are found to be overrepresented on school boards with AL elections when African Americans occupy a smaller part of the population (Meier and Rutherford 2014). 114 At-Large and Single-Member Districts 6 Houston Specific Outlook Given the trends found in scholarship on SM versus AL districts, we can make some educated predictions about the impact of these districts specifically for the City of Houston. Houston is a geographically large city with an incredibly diverse racial makeup; however, this racial diversity is also highly segregated by neighborhood (Houston Chronicle 20152). These two patterns make Houston an ideal location for a mix of AL and SM districts. The result of the concentrated pockets of different ethnic and racial groups in these districts will allow greater descriptive representation of the diverse communities within the city in the SM districts while the AL members must appeal to the broader diverse electorate of the entire city. References Alozie, N., and Manganaro, L. 1993. “Women’s Council Representation: Measurement Implications for Public Policy.” Political Research Quarterly 46(2): 383-398. Arrington, T., and Watts, T. G. 1991. “The Election of Blacks to School Boards in North Carolina.” The Western Political Science Quarterly 44(4): 1099-1105. Brace, K., Grofman, B., Handley, L., and Niemi, R. 1988. “Minority Voting Equality: The 65 Percent Rule in Theory and Practice.” Law and Policy 10(1): 43-62. Brockington, D., Donovan, T., Bowler, S., and Brischetto, R. 1998. “Minority Representation under Cumulative and Limited Voting.” Journal of Politics 60(4): 1108-1125. Bullock, C., and MacManus, S. 1990. “Structural Features of Municipalities and the Incidence of Hispanic Councilmembers.” Social Science Quarterly 71(4): 665-681. Bullock, C. S. III, and MacManus, S. A. 1991. “Municipal Electoral Structure and the Election of Councilwomen.” Journal of Politics 53(1): 75-89. Cameron, C., Epstein, D., and O’Halloran, S. 1996. “Do Majority-Minority Districts Maximize Substantive Black Representation in Congress?” American Political Science Review 90(4): 794-812. Darcy, R., Hadley, C. D., and Kirksey, J. F. 1993. “Election Systems and the Representation of Black Women in American State Legislatures.” Women and Politics 13: 73-89. Darcy, R., Welch, S., and Clark, J. 1985. “Women Candidates in Single-Member and Multimember Districts: American State Legislative Races. Social Science Quarterly 66(4): 945-953. 2 https://www.chron.com/news/houston-texas/article/Houston-is-both-one-of-the-most-diverse-and-most- 6236793.php. 115 Hobby School of Public Affairs White Paper Series___________________________________________________ 7 Davidson, C. 1979. “At-Large Elections and Minority Representation.” Social Science Quarterly 60(2): 336-338. Davidson, C. 1992. “The Voting Rights Act: A Brief History.” In B. Grofman and C. Davidson, eds., Controversies in Minority Voting: The Voting Rights Act in Perspective, 7-34. Washington, D.C.: Brookings Institution Press. Davidson, C., and Grofman, B. (eds.) 1994. Quiet Revolution in the South: The Impact of the Voting Rights Act, 1965-1990. Princeton, NJ: Princeton University Press. Enns, P. K., and Wlezien, C. (eds.) 2011. Who Gets Represented? New York, NY: Russell Sage Foundation. Gilens, M., and Page, B. I. 2014. “Testing Theories of American Politics: Elites, Interest Groups, and Average Citizens.” Perspectives on Politics 12(3): 564-581. Githens, M., and Prestage, J. (eds.) 1977. A Portrait of Marginality: The Political Behavior of American Women. New York, NY: McKay. Herrick, R., and Welch, S. 1992. “The Impact of At-Large Elections on the Representation of Black and White Women.” National Political Science Review 3: 62-77. Hogan, R. E. 2001. “The Influence of State and District Conditions on the Representation of Women in US State Legislatures.” American Politics Research 29(1): 4-24. Karnig, A., and Welch, S. 1979. “Sex and the Ethnic Differences in Municipal Representation.” Social Science Quarterly 60(3): 465-481. Karnig, A., and Welch, S. 1982. “Electoral Structure and Black Representation on City Councils.” Social Science Quarterly 63(1): 99-114. Kosterlitz, M. J. 1986. “Thornburg v. Gingles: The Supreme Court’s New Test for Analyzing Minority Vote Dilution.” Catholic University Law Review 36(2): 531-563. Leal, D., Martinez-Ebers, V., and Meier, K. 2004. “The Politics of Latino Education: The Biases of At-Large Elections.” Journal of Politics 66(4): 1224-1244. Matland, R. E. 1995. “How the Electoral System Structure Has Helped Women Close the Representation Gap.” In L. Karvonen and P. Selle, eds., Closing the Gap, Women in Nordic Politics, 281-312. London, UK: Dartmouth. Matland, R. E., and Brown, D. D. 1992. “District Magnitude’s Effect on Female Representation in U.S. State Legislatures.” Legislative Studies Quarterly 17(4): 469-492. Meier, K. J., and Stewart, J. Jr. 1991. “Cooperation and Conflict in Multiracial School Districts.” Journal of Politics 53(4): 1123-1133. 116 At-Large and Single-Member Districts 8 Meier, K. J., Juenke, E. G., Wrinkle, R. D., and Polinard, J. L. 2005. “Structural Choices and Representational Biases: The Post-Election Color of Representation.” American Journal of Political Science 49(4): 758-768. Meier, K. J., and Rutherford, A. 2014. “Partisanship, Structure, and Representation: The Puzzle of African American Education Politics.” American Political Science Review 108(2): 265- 280. Polinard, J., Wrinkle, R., and Longoria, T. 1991. “The Impact of District Elections on the Mexican American Community: The Electoral Perspective.” Social Science Quarterly 72(3): 608-614. Rocha, R. R. 2007. “Black-Brown Coalitions in Local School Board Elections.” Political Research Quarterly 60(2): 315-327. Rule, W. 1992. “Multimember Legislative Districts: Minority and Anglo Women’s and Men’s Recruitment Opportunity.” In W. Rule and J. F. Zimmerman, eds., United States Electoral Systems: Their Impact on Women and Minorities, 57-72. New York, NY: Greenwood Press. Rule, W. 1994. “Parliaments of, by, and for the People: Except for Women?” In W. Rule and J. F. Zimmerman, eds., Electoral Systems in Comparative Perspective: Their Impact on Women and Minorities, 15-30. Westport, CT: Greenwood. Sass, T. R. 2000. “The Determinants of Hispanic Representation in Municipal Government.” Southern Economic Journal 66(3): 609-630. Schwindt-Bayer, L. A., and Mishler, W. 2005. “An Integrated Model of Women’s Representation.” Journal of Politics 67(2): 407-428. Thornburg v. Gingles, 478 US 30. 1986. Trounstine, J., and Valdini, M. E. 2008. “The Context Matters: The Effects of Single-Member versus At-Large Districts on City Council Diversity.” American Journal of Political Science 52(3): 554-569. Welch, S. 1990. “The Impact of At-Large Elections on the Representation of Blacks and Hispanics.” Journal of Politics 52(4): 1050-1076. Welch, S., and Studlar, D. 1990. “Multimember Districts and the Representation of Women: Evidence from Britain and the United States.” Journal of Politics 52(2): 391-412. Zax, J. S. 1990. “Election Methods and Black and Hispanic City Council Membership.” Social Science Quarterly 71(2): 339-355. 117 Wards, At-Large Systems and the Focus of Representation in Canadian Cities Author(s): Royce Koop and John Kraemer Source: Canadian Journal of Political Science / Revue canadienne de science politique, September 2016 septembre, Vol. 49, No. 3 (September 2016 septembre), pp. 433-448 Published by: Canadian Political Science Association and the Société québécoise de science politique Stable URL: https://www.jstor.org/stable/26291378 JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a widerange of content in a trusted digital archive. We use information technology and tools to increase productivity and.facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available athttps://about.jstor.org/terms Canadian Political Science Association and are collaborating with JSTOR to digitize, preserve and extendaccess to Canadian Journal of Political Science / Revue canadienne de science politique This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 20:12:16 UTC All use subject to https://about.jstor.org/terms 118 Wards,At-Large Systems and the Focus of Representation in Canadian Cities Royce Koop University ofManitoba John Kraemer Queen's University Ward politics tends to keep issues on a local rather than on a city wide basis.When a city undertakes to reconstitute its political system by dropping ward in favour of at-large elections,the change is fondamental,not procedural;elected officials find that they must represent the city,not the ward. This quotation from the classic text on city politics by Banfield and Wilson (1963:52)illustrates a foundational point ofagreement between opponents in the ongoing debate over the use of ward versus at-large systems of elec tion for city councillors.The point of agreement is that the system of elec tion influences the subsequent representational perceptions and behaviours ofcouncillors.The use of wards incentivizes a representational preoccupa tion with neighbourhoods and geographically defined areas within cities.In contrast,at-large systems lead councillors to diversify their representational activities in order to represent what they perceive to be the best interests of their cities as wholes (see Berger,2004:28),which may in fact cloak repre sentational preoccupations with privileged city interests such as land-based private business interests (Elkin,1987:36-60).To employ the terminology ofthe literature on representational attitudes,the unit ofelection is expected to shape the representationalfocus ofcouncillors (which is defined simply as the population councillors choose to represent),with ward councillors focused on certain geographic segments of the city and at-large councillors focused on their cities as wholes (on the focus of representation,see,for example,Eulau and Karps,1977). Royce Koop,Department of Political Studies,University of Manitoba,532 Fletcher Argue,Winnipeg MB R3T 2N2,royce.koop@umanitoba.caJohnKraemer,Department of Development,Faculty of Arts and Science,Queen's University,94 University Avenue,Kingston ON K7L 3N6,j.kraemer@queensu.ca Canadian Journal ofPoliticalScience/Revuecanadienne de sciencepolitique 49:3 (September /septembre 2016)433-448 doi:10.1017/S0008423916000512 ©2016 Canadian Political Science Association (l'Association canadienne de science politique) and/et la Société québécoise de science politique This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 20:12:16 UTC All use subject to https://about.jstor.org/terms 119 434 Royce Koop and John Kraemer The representational consequences of electoral institutions are well accep ted by advocates of both ward and at-large systems of elections;indeed,it is these consequences that make up a significant portion of both sides'argu ments in favour of their preferred systems.Beginning in the late eighteenth century,the US municipal reform movement advocated at-large systems of election explicitly to rid cities ofwhat members viewed as party-associated corruption and the shortsighted parochialism of councillors focused on par ticular geographic segments of their cities.The use of at-large systems,it was thought,would root out neighbourhood machines and force councillors to take a wider,citywide view in their representational activities;in other words,the change to at-large systems was thought to lead to a necessary and desirable alteration to councillors'foci of representation. In contrast,defenders ofward systems note that at-large systems privi lege particular segments of cities and allow councillors to ignore entire tracts of their cities,and can point to judicial decisions ruling that at large systems discriminate against minority communities (Dhindsa,2012; Packer,1982;Scarrow,1999:557;Smith and Stewart,1998;Trounstine and Valdini,2008).These defenders see the perpetuation of both non-par tisan local politics and at-large systems as an outcome ofthe self-interest of sitting elected officials rather than any idealistic focus of representation (Magnusson,1983:10).The goal of ward systems of elections is therefore explicitly to incentivize representational focus on certain geographic sec tions of the city in order to ensure that all citizens,including underprivile ged sections ofthe city,are represented at city hall.The substantial literature focused on the descriptive representation of African Americans,Hispanics and women on US city councils supports this particular consensus,as at large systems are understood to produce poor representational outcomes for these groups (Bullock and MacManus,1991;Herrick and Welch, 1992;Sass,2000;Taebel,1978;Tolley,2011;Welch,1990;but also see Clark and Krebs,2012). Other studies directly test the relationship between elective institutions and foci of representation.Early work such as the study by Eulau and Prewitt (1973)found that the vast majority of councillors surveyed in American cities held a citywide representational focus;however,Eulau and Prewitt do not control for unit of election,and the vast majority ofcoun cillors surveyed were elected in at-large systems. In contrast,Welch and Bledsoe (1988)provide evidence that institu tions do shape representational foci.They draw on a survey of American councillors across a range of institutional settings and find that councillors are most likely to report either a citywide or geographic community focus of representation as "Very Important"(79%and 49%respectively).Further more,Welch and Bledsoe demonstrate that differences in representational foci are clearly related to institutional differences.At-large units ofelection are positively related to citywide foci,but are negatively related to This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 20:12:16 UTC All use subject to https://about.jstor.org/terms 120 Abstract.Proponents ofboth wardand at-large systems agree that these systems ofelection play a role in shaping the representational foci (that is,who representatives seek to represent)of city councillors and,in so doing,affect the quality of local democracy.Canadian cities employ both ward and at-large systems ofelection,and therefore provide an opportunity to explore the relations hip between elective systems and focus ofrepresentation.We draw on data derived from both 52 interviews with and a survey of councillors in Canadian cities to test the proposition that cities' systems of election influence the representational foci of councillors.We find strong evidence that ward systems are related to arepresentational focus on geographically defined neighbourhoods, whereas councillors in at-large systems report prioritizing representation of their cities as wholes. Résumé.Les promoteurs des deux modes de scrutin par quartier et de type at large conviennent que ces systèmes électoraux jouent un rôle dans l'établissement des objectifs de représentation des conseillers municipaux (c.-à-d.les groupes que les représentants cherchent à représenter)et que,ce faisant,ils ont une incidence sur la qualité de la démocratie locale.Les villes canadiennes appli quent aussi bien des modes de scrutin par quartier et at large et offrent donc la possibilité d'exa miner la relation entre systèmes électifs et objectifs de représentation.Nous nous appuyons sur des données issues d'un sondage et de 52 entrevues auprès de conseillers de villes canadiennes pour vérifier l'hypothèse selon laquelle les systèmes électoraux influent sur les objectifs de représentation des conseillers.De solides éléments probants révèlent que les modes de scrutin par quartier sont liés à une représentativité axée sur des districts délimités géographiquement, tandis que dans les modes de scrutin at large les conseillers indiquent accorder une priorité à la représentation de leurs villes considérées comme des ensembles. neighbourhood orientations.These studies of representational foci are related to the broader literature outlined above,which is concerned with the linkage between elective institutions on the one hand and democratic responsiveness on the other (Brack et al.,2012;Clark and Krebs,2012; Pilet et al.,2012). In this study,we add empirical evidence from Canada to the existingacademicliterature,which demonstrates the importance of elective institu tions to shaping councillors'representational perceptions.This research builds upon a vibrant homegrown literature that explores and assesses the benefits and drawbacks of both at-large and ward systems of election in Canadian cities (Berger,2004;Dhindsa,2012;Mellon,1993).We draw on 52 original interviews and a survey of Canadian councillors to test the proposition that electoral institutions shape councillors'representationalfocus.We address the following research question:To what extent do the presence of ward and at-large units of election in Canadian cities influence the adoption by councillors of representational foci on specific geographicareaswithincitiesversustheircitiesaswholes? This research is grounded in an institutionalist framework. Institutionalist approaches explain the behaviours of political actors by exploring the roles ofinstitutions in shaping both the perceptions and beha viours ofthose actors (Thelen,1992).In this study,city councillors as representativesaresituatedininstitutionalcontexts(particularly related to the This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 20:12:16 UTC All use subject to https://about.jstor.org/terms 121 436 Royce Koop and John Kraemer system of election present)that produce strong incentives for certain repre sentational foci over others.Canada is a particularly useful laboratory in order to study the influence of institutions on representational foci ofcoun cillors because,like the US,city government in Canada is characterized by diversity in the institutions used to elect councillors (Sancton,2011:ch.9).In addition to their election in wards or at-large,councillors may,for example,represent very populous or sparsely populated wards.Further, some councillors act as representatives within the shadow ofamalgamations that created "super cities,"such as Toronto and Montreal,whereas other councillors,such as those in Vancouver,act as representatives within insti tutional settings characterized by local governments that are much smaller in scope (Sancton,2008).Further,elective institutions have consequences for democratic outcomes of crucial importance,particularly turnout, which have been understudied in the Canadian context (but see Couture et al.,2014). Both the framework and the relevant literature generate two easily tes table hypotheses.First, Hj:Councillors elected in wards are more likely to have a geographic representational focus. And,conversely, H2:Councillors elected in at-large systems are more likely to have a citywide representational focus. This study adds further empirical clarification for the role of institu tions in shaping both the representational perceptions and behaviours of city councillors.This is achieved in a non-traditional context for the study of city councillors,Canada,and therefore expands this topic of study to different comparative contexts.Furthermore,this study expands our understanding of institutions and the roles of councillors in local demo cracy:in contrast to members of Parliament,there is very little understand ing ofthe day-to-day perceptions and activities ofCanadian city councillors (Mévellec,2009;Taylor and Eidelmann,2010).While this paper does not report on the activities of councillors,it elucidates an important psycholo gical orientation of councillors that affects their subsequent behaviours. We first describe the interviews and survey that make up the data for this multi-method project.The analysis tests the above hypotheses,first, through a bivariate analysis,then in a multivariate analysis where several control variables are included.The conclusion ends by outlining the paper's contribution to both the literature and the ongoing debate on ward and at-large systems of representation in Canada and elsewhere. This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 20:12:16 UTC All use subject to https://about.jstor.org/terms 122 Wards,At-Large Systems and the Focus ofRepresentation 437 Methodology and Case Selection A multi-method research design involving interviews with and a survey of municipal councillors in Canada was developed for this project.A multi method design was used to leverage the unique advantages of both small n and large-n approaches.Very little has been written about processes of representation in Canadian cities;as a result,a qualitative approach employ ing interviews was appropriate to generate theory regarding this phenome non (see,for example,Lijphart,1971:685).The data collected through these interviews,however,were not representative of all Canadian council lors.Furthermore,hypotheses concerning the influence of elective institu tions on councillors'focus of representation could not be adequately tested with only interview data.Accordingly,a survey was employed to achieve these goals. The multi-method approach employed is a sequential exploratory multi method design (Creswell,2003:211).Such a multi-method approach is desi gned in such a way that qualitative data collection and analysis takes places prior to quantitative data collection.The advantages of such an approach for this research are threefold.First,phenomena uncovered during the explora tion ofa small number ofcases can inform the design ofthe subsequent quan titative inquiry.Second,causal relationships suggested through qualitative data analysis can be further explored and,more importantly,tested through the use of large-n data in subsequent quantitative analysis.Finally,triangula tion—comparing results from different data sources or by using different methods of data collection (Mays and Pope,2000:51)—can strengthen the validity of findings in qualitative research. Interviews with councillors were conducted both in person and by tele phone throughout 2012 and 2013.The interview schedule was semi-struc tured in order to ensure comparable data were produced but also so that new themes could be explored.The danger of an unstructured interview format is that data cannot be compared;this potential problem was addressed by the addition of a structured component to the interviews (Small,2009). Councillors were asked to speak to a number of themes but,in accordance with the semi-structured approach,were given wide latitude to expand on their answers and take the interviews in unanticipated directions.All inter views were recorded and subsequently transcribed. Councillors were invited to participate in order to achieve variation on a number of institutional and ecological variables related to the cities theyservedwithin.All 179 sitting councillors from Montreal,Ottawa,Toronto and Greater Vancouver (Burnaby,Coquitlam,Delta,Richmond,Surrey and Vancouver)were invited to participate in the interviews for this study.1 Fifty-two councillors agreed to be interviewed for the project,for a response rate of 29 per cent.Variation on all key variables—including individual level variables such as gender and time spent in office—was achieved. This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 20:12:16 UTC All use subject to https://about.jstor.org/terms 123 438 Royce Koop and John Kraemer The second phase of data collection involved an online survey of all sitting councillors in Canadian cities with populations over 20,000 resi dents.The survey was conducted in February 2014;589 out of 1841 coun cillors responded to the invitation to participate,for a total response rate of 32 per cent.An online survey format was selected because the vast majorityofcouncillorsinCanadiancitieswithpopulationsover20,000 provideemailaddressesoncitywebsites.The potential for bias to result from under-reporting of councillors without email addresses or online access, which is a potential problem associated with online surveys,was therefore ruled out (Van Selm and Jankowski,2006:436). Councillors were asked a range of questions about their representatio nal perceptions and practices.2 During the interviews for this project,repre sentational foci were probed through asking three types ofquestions.First, direct questions were designed to elicit groups that councillors were parti cularly focused on (for example,"Who do you seek to represent as a coun cillor?Do you see yourself as representing any particular group of people within your city?").Second,one question related to conflict between ward and citywide interests was employed to determine the extent to which ward councillors prioritized the needs of their wards versus those of their cities (for example,"Has there ever been a time when what's best for the city and what's best for the ward came into conflict?How did you deal with that?").Finally,open-ended policy questions were designed to elicit the policy priorities of councillors,which could then later be coded as primarily geographic or citywide in nature. The responses to these questions informed three sets ofquestions in the survey that explored councillors'representational foci.The first was a direct question asking councillors how important their representation ofparticular groups was,including "a geographical area of the city or a neighbourhood" and "the city as a whole."Second,a conflict question asking "When voting, do you feel you should be more concerned with the needs ofyour ward/dis trict or the needs of the city as a whole"was put to councillors elected in wards.Finally,a closed policy question was formulated by combining res ponses to the opened-ended question asked in the interviews:"In your opinion,how important are each of the following policy areas to you as a council member?"In this and other closed questions asked in the survey, the options available to respondents were generated inductively from semi-structured questions asked during the interviews with councillors that came beforehand. While we had three measures of representational foci available to us from the survey,we have selected the direct measure for analysis only for three reasons.First,the conflict question in the survey was asked only of councillors elected in wards and so could not be used to address the research questions in this paper.Second,it was not always clear in coding responses to our policy questions which answers indicated a This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 20:12:16 UTC All use subject to https://about.jstor.org/terms 124 Wards,At-Large Systems and the Focus ofRepresentation 439 geographic focus and which indicated a citywide focus.Finally,the use of the direct measure only was conducive to addressing our research questions within the context of a relatively short article. Analysis The research question for this paper is to what extent does the presence of ward and at-large units ofelection in Canadian cities influence the adoption by councillors of representational foci on specific geographic areas within cities versus their cities as wholes.This section provides two examples of these different foci and tests the two hypotheses presented in the first section. The interviews for this project revealed examples of both geographic and citywide representational foci.The interviews also provide anecdotal evidence in support of our two hypotheses.When asked whom they sought to represent,councillors elected in wards tended to mention citizens within the geographic confines oftheir wards whereas councillors elected at large tended to mention all residents of their cities.There was,however, important nuance evidence in councillors'replies to these questions. In the 2004 report of the Vancouver Electoral Reform Commission, Justice Berger wrote,"I have come down in favour of wards because... they provide for neighbourhood representation and accountability"(Berger,2004:49).Councillors who were elected in wards often affirmed Berger's argument by reporting geographic foci of representation.Councillor Adam Vaughan,elected in a Toronto ward,provides a good example ofa councillor with a local geographic focus."My job is to represent the residents who live in the neighbourhoods that comprise Ward 20 that I represent.And myjob is to make sure that they get city services they have been promised and have paid for and created through their taxes but also to make sure the city services they need or want are delivered to them...in the best way possible."In contrast,councillors may also take a citywide focus of representation.The Vancouver Electoral Reform Commission found local support for councillors adopting this representational focus,reporting, "Supporters of the at-large system argue that it results in the election of city councillors with a citywide vision.This was,aside from the more general assertion that the present system was working well,the most force fully advanced endorsement of the present system"(Berger,2004:28).Councillor Bruce Hayne,who was elected at large in the British Columbia city of Surrey,emphasizes just such a citywide focus. "[I represent]all of the city as a whole,absolutely.We have to make deci sions that are absolutely in the best interest of all of Surrey." Geographic and citywide foci ofrepresentation are not mutually exclu sive;councillors may not view themselves as concerned with the interests of This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 20:12:16 UTC All use subject to https://about.jstor.org/terms 125 440 Royce Koop and John Kraemer the city as a whole over particular geographic areas or vice versa.Instead, councillors often explained how they reconcile their representational concern for these two units of representation.Councillor Frank Di Giorgio from Toronto,for example,exhibits concern for both units,but, despite his being elected in a ward,ultimately he identifies a citywide focus as his primary focus of representation."First of all,you're an elected representative of people in an area in the city so you represent the interests of that area.But you also serve as a member of a larger group, council,which presumably has the interests of the city at large at heart. So,notwithstanding that you represent the interests of a particular sub group,I believe that...your main responsibility is to look after the interests of the larger group." Councillor Rick Chiarelli from Ottawa similarly balances the needs of his ward and the city as a whole.For Chiarelli,representation of ward inter ests takes place through his own work on citywide issues,and his primary representational focus,in contrast to that of Di Giorgio,is undoubtedly on his own ward:"I think my role is to protect the interests of my ward resi dents and to promote individual strategic initiatives within the ward.... So in every major city issue,there will be nuances that impact each ward differently.So I have to be on top of that.But [I]also find out where indi vidual ward initiatives fit within the city's overall plans and make sure they get [addressed]in a way that has the best chance of succeeding." Councillor Peter Chernushenko,also from Ottawa,identifies a clear geographic focus ofrepresentation,but presents this as a logical implication ofhis residence in the area rather than as a result of any electoral imperati ves."It's only natural that I would be aware ofand have a soft spot for the issues that are closest to home.That's the neighbourhood,the street that I like to shop on,so I want it to do well.That's the route that I cycle on fre quently or the park that my kids play at.There will naturally be a desire to defend,protect and improve those sorts of things." Similarly,even when councillors are elected at large,they may report a representational preoccupation with geographic areas they have a particular connection to,especially those in which they reside.Councillor Sylvia Bishop from Delta provides an example of such a representative."I live in Tsawwassen so I get the mood of Tsawwassen just by being out in the community.To get the same feeling from North Delta is more difficult because I'm not there every day....[But]at the end of the day,I think the role is to represent the whole community." Even Councillor Hayne,who identified a citywide representational focus within the context of Surrey's at-large system,alluded to a particular representational connection to his own geographic community."I think everybody has some affinity to the geography of where they live.I live in Cloverdale so I tend to be around Cloverdale a lot,shop around Cloverdale.That sort of thing."These results suggest that while at-large This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 20:12:16 UTC All use subject to https://about.jstor.org/terms 126 Wards,At-Large Systems and the Focus ofRepresentation 441 Figure 1 Representational Foci of Councillors Elected in Wards and At-Large QElected in Ward E3 Elected At-Large n =283 n =32 n =291 n =134 Focus on Geographic Area Focus on City as Whole systems may provide incentive for a citywide focus of representation,coun cillors may still find themselves focusing their representational efforts on certain geographic areas of their cities,which reinforces some critiques of the at-large system (see,for example,Dhindsa,2012). As noted,our interviews hinted at the presence of a relationship between electoral institutions and representational foci,which would confirm the finding of Welch and Bledsoe (1988)from their survey of American councillors.However,these findings are not representative and,furthermore,it is possible that self-serving bias may result from coun cillors'own accounts of their representational foci.We therefore further explore our research question through the use of our survey dataset,speci fically /-tests,to determine whether differences in representational foci are statistically significant between councillors elected in wards and at large and subsequently in a multivariate logistical regression model. Figure 1 summarizes the percentages of councillors who report a geo graphic or citywide focus of representation as "Very Important"divided into two groups,those who were elected in wards and those who were elected at-large.There are significant differences in the extents to which councillors elected in wards and at large rate geographic and citywide foci ofrepresen tation as very important to them.Seventy per cent of councillors elected in wards report a geographic focus,whereas only 21.6 per cent of councillors elected at large report such a focus.The difference between the two groups is 48.4 per cent,and the difference is statistically significant in a /-test (p = 0.00,two-tailed).In addition,95 per cent of councillors elected at large n =283 n =32 Focus on Geographic Area QElected in Ward S Elected At-Large 95.0 n =291 n =134 Focus on City as Whole This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 20:12:16 UTC All use subject to https://about.jstor.org/terms 127 442 Royce Koop and John Kraemer reported a citywide focus ofrepresentation,whereas 72.6 per cent of coun cillors elected in wards reported such a focus.The difference is 22.4 per cent,and the difference is again statistically significant in a ?-test (p = 0.00,two-tailed). While these bivariate tests substantiate both hypotheses,it is possible that other variables are related to the distinction between geographic and citywide foci of representation.It is therefore necessary to include several control variables to test the strength of ward and at-large systems on focus of representation.This is achieved through the estimation of two binary logistical regression models with several control variables included as independent variables in addition to an indicator for unit of election to explore the main effects of interest.All control variables have been found in previous studies to affect both representational attitudes and behaviour. These control variables can be separated into three groups:institutional, ecological and individual level. Three institutional variables are included.First,Population by Chamber Size is a measure ofrepresentational load derived by dividing the city popu lation by the size of the council chamber (Downey and Williams,1998). Second,councillors'salary is one indicator ofprofessionalized legislative ins titutions,with higher pay associated with both full-time employment and pro fessional rather than amateur councillors (Squire,2007). Third,we argue that the presence ofparties or slates in city elections is an institutional characteristic of those cities that may act to shape council lors'representational perceptions.Parties and slates add both structure and a need for co-operation to councillors'elective and representational beha viours.Furthermore,local voters may use party labels as heuristics in guiding their vote choice (Cutler and Matthews,2005:363-64),rendering the representational behaviours of individual councillors less important for re-election than in cities where parties or slates are characteristics of local democracy.There is substantial diversity in the presence of parties and slates in Canadian cities;among the largest cities in this sample,for example,Montreal and Vancouver are characterized by entrenched party systems,whereas Toronto is not. One ecological variable is included:the population ofthe city in which councillors are elected.Large and small cities impose different representa tional requirements and obligations on representatives and so are expected to lead to different representational perceptions on the part of councillors. (Ellickson and Whistler,2001;Johannes,1984;Thomas,1992).Stewart (1997)makes this point in a Canadian context,finding that population is related to citizen engagement and thus the expectations of local representa tives.Accordingly,a variable is included indicating the population of the city councillors are elected in (by 1000). Finally,five individual-level variables are included in the model.First, councillors'gender is included as a variable since gender influences overall This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 20:12:16 UTC All use subject to https://about.jstor.org/terms 128 Wards,At-Large Systems and the Focus ofRepresentation 443 Table 1 Influence of Ward Election on Likelihood to Adopt Geographic Area and City as Whole Focus of Representation (Binary Logistical Regression) Model 1 Model 2 Geographie Area Focus City as Whole Focus B S.E.Sig.B S.E.Sig. Institutional Elected in ward Pop.by chamber (by 1000) Councillor salary Party or slate 2.21 0.30 0.00 -2.68 0.73 0.00 0.02 0.02 0.38 0.01 0.02 0.54 0.00 0.00 0.62 0.00 0.00 0.16 -0.54 0.33 0.12 -0.23 0.36 0.51 Ecological Population (by 1000)0.00 0.00 0.53 -0.00 0.00 0.91 Individual Gender Years in office Style 0.15 0.27 0.57 0.06 0.02 0.01 -0.17 0.08 0.04 0.41 0.24 0.09 0.19 0.31 0.54 0.01 0.02 0.61 0.11 0.09 0.23 Static ambition Progressive ambition 0.82 0.38 0.03 Constant N Pseudo R2 -1.52 0.53 0.01 381 0.30 3.36 0.87 0.00 374 0.21 Note:Bolded values are significant at the P <0.05 level. representational approaches (Richardson and Freeman,1995).Second,the total years served in office by councillors is included.Third,the represen tational style of councillors is included.Style is understood in the literature on representational attitudes to address the extent to which representatives consider themselves delegates—faithful servants ofconstituents'opinion— or trustees who are tasked with using their judgment in approaching repre sentational challenges (Eulau et al.,1959).Finally,two variables measuring the ambition of councillors are included.Static ambition refers to council lors who intend to remain in their current positions,whereas progressive ambition refers to councillors who wish to advance to higher office (Herrick,2011).Table 1 illustrates two logistical regression models with the binary res ponse of "Very Important"for geographic areas and cities as wholes inclu ded as dependent variables.Model 1 uses geographic focus as the dependent variable,whereas Model 2 employs a citywide focus as the dependent variable.Descriptions of all variables included can be found in appendix 1. Both models perform reasonably well in explaining variation in the dependent variables,although the first model performs better at explaining B Institutional Elected in ward 2.21 Pop.by chamber (by 1000)0.02 Councillor salary 0.00 Party or slate -0.54 Ecological Population (by 1000)0.00 Individual Gender 0.15 Years in office 0.06 Style -0.17 Static ambition 0.41 Progressive ambition - Pnnctnnt —1 S9 S.E.Sig.B S.E. 0.30 0.00 0.02 0.38 0.00 0.62 0.33 0.12 0.00 0.53 0.27 0.57 0.02 0.01 0.08 0.04 0.24 0.09 0 53 noi -2.68 0.73 0.01 0.02 0.00 0.00 -0.23 0.36 -0.00 O.OO 0.19 0.31 0.01 0.02 0.11 0.09 0.82 0.38 3 36 0.87 This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 20:12:16 UTC All use subject to https://about.jstor.org/terms 129 444 Royce Koop and John Kraemer variation in focus on geographic areas (31%)than the second model per forms in explaining variation in citywide focus (21%). This multivariate analysis confirms the substantiation of both hypothe ses.The presence of wards (and,conversely,at-large systems)has a statis tically significant effect on the likelihood of councillors adopting either geographic or citywide foci of representation.Furthermore,both coeffi cients are in the anticipated direction,with a positive coefficient indicating that wards increase the likelihood of geographic focus and a negative coef ficient indicating that at-large systems increase the likelihood of citywide focus.The finding of a relationship between elective institution and focus of representation withstands the inclusion of control variables. It is worth pointing out the presence ofthree control variables that also exercised statistically significant effects on the dependent variables.First is years in office,which is related to a higher likelihood ofadopting a geogra phic focus of representation.Second is representational style.Given the coding of this variable (see appendix 1),the positive direction on this coef ficient suggests that a willingness to follow one's conscience in making representational decisions is also related to increased focus on geographic areas.Finally,progressive ambition is related to a citywide focus of repre sentation.It is likely that councillors who envision themselves in higher office expand their overall focus of representation to include new groups and neighbourhoods in cities to court future electors.This intriguing preli minary finding is further interrogated below. Conclusion This paper explored the role of elective institutions—particularly the use of ward and at-large systems of election—in shaping the representational foci of councillors in Canadian cities.This research provides substantiation for the view that elective institutions are crucial to shaping the representational foci of city councillors.Furthermore,this understanding is derived from an analysis ofcouncillors in a democracy in which the representational percep tions and behaviours of city councillors have not often been studied.The addition ofthis comparable case reiterates the finding that elective institu tions matter for councillors'foci of representation,with consequences for their representational priorities and behaviours.Further research could expand this finding to other states and institutional settings in order to enhance our understanding of the role ofinstitutions in shaping representa tives perceptions and behaviours. The multivariate research in this article also suggests a further avenue of research:the role of ambition in shaping the representational foci of councillors.It was found that progressive ambition—aspiring to higher office—is related to a citywide focus of representation among Canadian This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 20:12:16 UTC All use subject to https://about.jstor.org/terms 130 Wards,At-Large Systems and the Focus ofRepresentation 445 councillors.Councillors who hope to run for higher office as mayor or perhaps to sit in the provincial or federal legislatures as elected officials are more likely to focus their representational activities on their cities as wholes than are councillors who lack such ambition.The explanation underlying this is likely that rational councillors are laying the groundwork for higher office while still in their present positions by reaching out to communities beyond individual wards or geographic areas and that this is reflected in councillors'own representational foci.Future research could explore the interaction of institutions and individual ambition in shaping representational perceptions and behaviours in city politics. Notes 1 The greater number of cities from British Columbiawas included in order to explore the effects of the at-large systems on representation in that province. 2 Both the interview schedule and the survey are available from the corresponding author upon request. References Banfield,Edward and James Wilson.1963.City Politics.Cambridge MA:Harvard University Press. Berger,Thomas.2004.A City ofNeighbourhoods:Report ofthe 2004 Vancouver Electoral Reform Commission. Brack,Nathalie,Oliver Costa and Conceicao Teixeira.2012."Attitudes towards the Focus and Style of Political Representation among Belgian,French and Portuguese Parliamentarians."Representation 48 (4):387-402. Bullock,Charles and Susan Macmanus.1991."Municipal Electoral Structure and the Election of Councilwomen."The Journal ofPolitics 53 (1):75-89. Clark,Alistair and Timothy Krebs.2012."Elections and Policy Responsiveness."In The Oxford Handbook of Urban Politics,ed.Peter John,Karen Mossberger and Susan E. Clarke.New York:Oxford University Press. Couture,Jérôme,Sandra Breux and Laurence Bherer.2014."Analyse écologique des déterminants de la participation électorale municipale au Québec."Canadian Journal ofPolitical Science Al (4):787-812. Creswell,John.2003.Research Design:Qualitative,Quantitative,and Mixed Methods Approaches.New York:Sage. Cutler,Fred and Scott Matthews.2005."The Challenge of Municipal Voting:Vancouver 2002."Canadian Journal ofPolitical Science 38 (2):359-82. Dhindsa,Rebecca.2012."The At-Large System and Unequal Representation:An Analysis of Surrey Council."The Seed 13:10-16. Downey,Terrence and Robert Williams.1998."Provincial agendas,local responses:The 'common sense'restructuring of Ontario's municipal governments."Canadian Public Administration 41 (2):210-38. Ellickson,Mark and Donald Whistler.2001."Explaining State Legislators'Casework and Public Resource Allocations."Political Research Quarterly 54 (3):553-69. Elkin,Stephen.1987.City and Regime in the American Republic.Chicago IL:University of Chicago Press. This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 20:12:16 UTC All use subject to https://about.jstor.org/terms 131 446 Royce Koop and John Kraemer Eulau,Heinz and Kenneth Prewitt.1973.Labyrinths ofdemocracy;adaptations,linkages, representation,and policies in urban politics.Indianapolis IN:Bobbs-Merrill. Eulau,Heinz,John Wahlke,William Buchanan and LeRoy Ferguson.1959."The Role ofthe Representative:Some Empirical Observations on the Theory of Edmund Burke."The American Political Science Review 53 (3):742-56. Eulau,Heinz and Paul D.Karps.1977."The Puzzle of Representation:Specifying Components of Responsiveness."Legislative Studies Quarterly 2 (3):233-54. Herrick,Rebekah.2011.Representation and institutional design.Lanham MD:Lexington Books. Herrick,Rebekah and Susan Welch.1992."The Impact ofAt-Large Representation on the Representation of Black and White Women."In Ethnic Politics and Civil Liberties, ed.Lucius Barker.New Brunswick NJ:Transaction. Johannes,John.1984.To Serve thePeople:Congress and Constituency Service.Lincoln NE: University of Nebraska Press. Lijphart,Arend.1971."Comparative Politics and the Comparative Method."The American Political Science Review 65 (3):682-93. Magnusson,Warren.1983."Introduction:The Development of Canadian Urban Government."In City Politics in Canada,ed.Warren Magnusson and Andrew Sancton.Toronto:University ofToronto Press. Mays,Nicholas and Catherine Pope.2000."Qualitative research in health care:Assessing quality in qualitative research."British Medical Journal.50-52. Mellon,Hugh.1993."Reforming the Electoral System of Metropolitan Toronto:Doing Away with Dual Representation."Canadian Public Administration 36 (1):38-56. Mévellec,Anne.2009."Working the Political Field in Stormy Weather:A Mayor's Role in the Quebec Municipal Mergers."Canadian Journal ofPolitical Science 42 (3):773-92. Packer,Mark.1982."Tracking the Court through a Political Thicket:At-Large Election Systems and Minority Vote Dilution."Journal of Urban and Contemporary Law 23 (1):227-59. Pilet,Jean-Benoit,Andre Freire and Olivier Costa.2012."Ballot Structure,District Magnitude and Constituency-Orientation of MPs in Proportional Representation and Majority Electoral Systems."Representation 48 (4):359-72. Richardson,Lilliard and Patrick Freeman.1995."Gender Differences in Constituency Service among State Legislators."Political Research Quarterly 48 (1):169-79. Sancton,Andrew.2008.The Limits of Boundaries:Why City-Regions Cannot be Self Governing.Montreal and Kingston:McGill-Queen's University Press. Sancton,Andrew.2011.Canadian Local Government:An Urban Perspective.Toronto: Oxford University Press. Sass,Tim.2000."The Determinants of Hispanic Representation in Municipal Government." Southern Economic Journal 66 (3):609-30. Scarrow,Harold.1999.The impact of at-large elections:Vote dilution or choice dilution? Electoral Studies 18 (4):557-67. Small,Mario.2009.'"How many cases do I need?'On science and the logic ofcase selection in field-based research."Ethnography 10 (1):5-38. Smith,Patrick J.and Kennedy Stewart.1998.Making Local Accountability Work in British Columbia.Government-commissioned report for British Columbia Vancouver BC.: Ministry ofMunicipal Affairs and Housing,Institute of Governance Studies Squire,Peverill.2007."Measuring State Legislative Professionalism:The Squire Index Revisited."State Politics and Policy Quarterly 1 (2):211-27. Stewart,Kennedy.1997."Measuring Democracy:The Case of Vancouver."Canadian Journal ofUrban Research 6 (2):160-82. Taebel,Delbert.1978."Minority Representation on City Councils:The Impact of Structure on Blacks and Hispanics."Social Science Quarterly 59 (1):142-52. This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 20:12:16 UTC All use subject to https://about.jstor.org/terms 132 Wards,At-Large Systems and the Focus ofRepresentation 447 Taylor,Zack and Gabriel Eidelman.2010."Canadian Political Science and the City:A Limited Engagement."Canadian Journal ofPolitical Science 43 (4):961-81. Thelen,Kathleen,and Steinmo,Sven.1992."Historical Institutionalism in Comparative Politics."In Structuring Politics:Historical Institutionalism and Comparative Analysis, ed.Kathleen Thelen,Sveb Steinmo,and Frank Longstreth.Cambridge:Cambridge University Press. Thomas,Sue.1992."The Effects ofRace and Gender on Constituency Service."The Western Political Quarterly 45 (1):169-80. Tolley,Erin.2011."Do Women 'Do Better'in Municipal Politics?Electoral Representation across Three Levels of Government."Canadian Journal of Political Science 44 (3): 573-94. Trounstine,Jessica and Melody E.Valdini.2008."The Context Matters:The Effects of Single-Member versus At-Large Districts on City Council Diversity."American Journal ofPolitical Science 52 (3):554-69. Van Selm,Martine and Nick Jankowski.2006."Conducting Online Surveys."Quality and Quantity 40 (3):435-56. Welch,Susan.1990."The Impact ofAt-Large Elections on the Representation ofBlacks and Hispanics."The Journal ofPolitics 52 (4):1050-76. Welch,Susan,and Timothy Bledsoe.1988.Urban reform and its consequences:A study in representation.Chicago IL:University of Chicago Press. Appendix 1:Descriptive Statistics and Survey Questions for Independent Variables Table 2 Descriptive Statistics Variable Obs Dependent Variables Geographic area focus 555 City as whole focus 545 Institutional Elected in ward 568 Population by chamber (by 1000)589 Councillor Salary (by 1000)519 Ran with party or slate 547 Ecological Population (by 1000)589 Individual Gender 539 Years in office 527 Style 567 Static ambition 433 Progressive ambition 433 M SD Min Max 0.57 0.78 0.50 0.41 0.73 12.40 39.16 0.22 0.44 13.01 24.63 0.41 0 1.68 11.50 0 216.44 391.81 20.08 0.30 7.71 4.78 0.55 0.23 0.46 6.79 1.53 0.50 0.42 1 37 7 1 1 jraphic area focus as whole focus onal ed in ward lation by chamber (by 1000) icillor Salary (by 1000) with party or slate :cal lation (by 1000) ial ler s in office 555 0.57 545 0.78 568 0.73 589 12.40 519 39.16 547 0.22 589 216.44 539 0.30 527 7.71 0.50 0 0.41 0 0.44 0 13.01 1.6! 24.63 11.5( 0.41 0 391.81 20.0S 0.46 0 6.79 1 This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 20:12:16 UTC All use subject to https://about.jstor.org/terms 133 448 Royce Koop and John Kraemer Elected in Ward.Survey question:"Are you elected at large or in a ward/ division?" Councillor Salary.Survey question:"What is your annual gross salary as a council member?" Party or Slate.Survey question:"In the last election,did you run as a can didate with a particular party or slate?" Years in Office.Survey question:"Including the current year,how many years in total have you served as a council member?" Style.Survey question:"When you think there is a conflict between what you think is best and what you think people want,do you think you should follow your conscience or follow what the people want?" Responded answered on a 7-point scale where "1"meant "Always what the people want"and "7"meant "Always conscience."The wording for this question is from Herrick (2011). Progressive and Static Ambition.Survey question:"Which of the follow ing best describes what you plan to do be doing five years from now?"The available responses include "Remain on council,""Hold a higher elected office,""Hold a lower elected office,""Hold no elected office,""Don't know."This variable was recoded into two dichotomous variables.For the variable indicating progressive ambition,"Hold a higher elected office"was recoded into "1"whereas all other responses were recoded into "0."For the variable indicating static ambition,"Remain on council,"was recoded into "1"whereas all other responses were recoded into "0."The wording for this question is from Herrick (2011:156). This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 20:12:16 UTC All use subject to https://about.jstor.org/terms 134 The Context Matters: The Effects of Single-Member versus At-Large Districts on City Council Diversity Jessica Trounstine Princeton University Melody E. Valdini Portland State University Scholars continue to debate the degree to which electoral institutions matter for representation. The literature predicts that minoritiesbenefitfromdistrictswhilewomenbenefitfromat-largeelections.Themechanismsbywhichinstitutionsaffectthe ability of traditionally underrepresented groups to win seats have been understudied. Using an analysis of over 7,000 cities and interviews with city councilors, we find that compared to at-large systems, district systems can increase diversity only when underrepresented groups are highly concentrated and compose a substantial portion of the population. In addition, we find that the electoral system has a significant effect on representation only for African American male and white female councilors; the proportion of African American women and Latina councilors is not affected by the use of either district or at-large systems. Extensiveresearchhasbeendevotedtounderstand- ingthecontinuingunderrepresentationof women and people of color in legislatures. At the city level scholars have found mixed results for the effect of single- member district elections in increasing descriptive repre- sentation. Particularly in places where citywide elections were implemented to dilute the vote strength of racial groups, districts have been seen as a key factor in increas- ing racial and ethnic diversity. Alternatively for women, districtshavebeenfoundtobedetrimentaltotheelection of female councilors. Scholars have proposed numerous, contradictoryexplanationsforthesefindings.Forminori- ties the focus has been on residential segregation and size of the group, while women are said to benefit from the multicandidate setting of at-large elections. For women of color these explanations are in direct conflict. This ar- ticle contributes to this large literature by exploring the mechanisms by which institutions affect the representa- tionof differentgroups,concurrentlytestingthesegrega- tion and group size hypotheses and taking into account the joint relationship between race and gender. Many of our findings confirm conventional wisdom, but advance ourknowledgeinthisareabyofferingempiricalestimates Jessica Trounstine is assistant professor of politics and public affairs, Princeton University, Woodrow Wilson School, 303 Robertson Hall, Princeton, NJ 08544 (jessica@trounstine.com). Melody E. Valdini is assistant professor of political science, Portland State University, Division of Political Science, Hatfield School of Government, P.O. Box 751, Portland, OR 97207 (mev@pdx.edu). TheauthorswouldliketothankZoltanHajnalforthegeneroususeofhisdata.TheywouldalsoliketothankChrisAchen,ChuckCameron, DavidLewis,DougMassey,NolanMcCarty,andMelodyRoseforextremelyhelpfulcommentsondraftsofthearticle,andGretchenKafoury for her very helpful advice on securing interviews with city council members. Benjamin Tagoe provided excellent research assistance. of the effect of different demographic contexts in varied institutional environments. As the Supreme Court anticipated in the landmark case Thornburg v. Gingles (1986), we demonstrate that compared to citywide elections, districts increase repre- sentation when a group is geographically concentrated andmoderatelysized.Further,wefinddistrictsonlyben- efit black men. That is, the positive effect of districts is conditional on the context. Districts can increase oppor- tunities for representation, but in some cases districts are nothelpful.Onlyrarelydodistrictshaveasubstantialim- pact. Taking advantage of variation among city institu- tional structures, council composition, and demograph- ics, we use quantitative and qualitative methods to study these relationships. We analyze data from surveys of city clerksandelectionresultsfrommorethan7,000citiesand connect this analysis to the experience of local legislators through interviews with city councilors. While the existing literature on underrepresentation is vast, our article makes several contributions to the un- derstanding of the relationship between electoral institu- tions and representation. First, we offer a methodologi- cal contribution. While no model is perfect, our analyses American Journal of Political Science,Vol. 52, No. 3, July 2008, Pp. 554–569 C2008, Midwest Political Science Association ISSN 0092-5853 554 135 THE CONTEXT MATTERS 555 improve on previous research by taking into account the large number of cities with no female or minority coun- cilors, allowing us to make more precise predictions. We usetobitmodelstopredict,first,thelikelihoodthatacity will elect any women or people of color and then, to esti- mate the proportion of female and minority councilors. Second, while existing research on electoral systems and underrepresentation has tested the effects of either seg- regation or group size, we are the first to include both variables in our analysis. Further, much of the work on the effect of districts studies councils at or before 1990 and/or is limited to a small sample of cities; we use re- cent data on a large number of cities to analyze patterns across time and place. Finally, most previous research as well as the Thornburg v. Gingles (1986) decision assumes that the effect of electoral systems on the election of peo- ple of color is constant across gender. For example, the justices refer to the effect of districts on the representa- tion of “minority groups”or “black citizens,”but there is no discussion of the possibility that electoral institu- tions work differently for men as opposed to women of color.Thefourthcontributionofourarticleistoquestion this assumption, and, although we have limited data, we present evidence that the effect of electoral institutions is significantly different for men versus women of color. Even after decades of progress there remain sub- stantial disparities in the representation of black/African American, Latino/Hispanic, and women city council members compared to their population proportions. 1 Theaveragecityinourdatasethasapopulationthatis8% AfricanAmerican,7.6%Latino,and52%femalewhilethe averagecitycouncilhasamembershipthatis4.8%African American, 2.3% Latino, and 20.5% female. Yet, there is wide variation among municipalities and across time. A clearquestionemerges:whydosomecitiesdobetterthan others at electing women and people of color? Single-Member Districts: An Institutional Solution? One of the most persistent findings by scholars of urban politics is that single-member district elections increase descriptiverepresentationof underrepresentedracialand ethnic groups on city councils. 2 This effect has been found to be particularly strong for African Americans 1We use the terms black/African American and Hispanic/Latino in- terchangeably. Due to data limitations we are not able to study the effect of electoral institutions for Asian Americans. 2Descriptive representation and substantive representation are not interchangeable. See Guinier (1992) and Tate (2003) for in-depth discussions. (see, for example, Arrington and Watts 1991; Bullock andMacManus1990;DavidsonandGrofman1994;Poli- nard,Wrinkle,andLongoria1991;Welch1990).3 Districts have also been found to be beneficial to Latinos (e.g., Davidson and Korbel 1981; Heilig and Mundt 1983; Leal, Martinez-Ebers, and Meier 2004; Taebel 1978). 4 These statistical findings have been supported by extensive case study and historical research as well (Bridges 1997; Rice 1977). In sum, the literature concludes that “the effect of ...districtsisunequivocally ...greaterequity”(Mundt and Heilig 1982, 1035). The literature on the representation of women finds precisely the opposite effect for single-member districts. While there are some exceptions, the vast majority of the research has concluded that districts are either meaning- less(AlozieandManganero1993;BullockandMacManus 1991) or disadvantageous for women candidates (see, for example,Darcy,Welch,andClark1987;Hogan2001;King 2002; Matland 1995; Matland and Brown 1992; Norris 1985;Rule1994;Schwindt-BayerandMishler2005,Welch and Studlar 1990). Because race and gender are not mutually exclu- sive categories, a handful of scholars have also sought to understand how electoral institutions affect women of colorgiventhattheyfaceapotentialdoubledisadvantage (Githens and Prestage 1977) and conflicting institutional effects.Existingresearchfindsthatblackwomenaremost likely to be elected in state multimember districts (anal- ogous to at-large elections in cities) while black men are disadvantagedbythisstructure(Darcy,Hadley,andKirk- sey1993;Rule1992).Similarly,HerrickandWelch(1992) andKarnigandWelch(1979)findthatblackmen,butnot blackwomen,areadvantagedbydistricts.Further,Karnig and Welch (1979) find no effect of districts for Mexican American men or women. This suggests that the effect of districts should be conditional on the characteristics of the group as well as the candidate in question. In addition to academic work, the process of vote dilution and the effect of institutional structures on rep- resentation have been the subject of intense legal analy- sis. The United States Supreme Court held in Thornburg v. Gingles (1986) that in challenging at-large or multi- member districts minority plaintiffs must demonstrate (among other things) that the group in question is suffi- cientlylargeandcompactenoughtoconstituteamajority of a single-member district. While these criteria are con- sistentlyusedinlegalandscholarlywork,therehavebeen 3Others find that districts are not superior (e.g., Bullock and Mac- Manus 1993) or that the effect of districts has substantially weak- ened over time (Welch 1990). 4Others find that at-large systems offer better representation for Latinos (e.g., Mladenka 1989) 136 556 JESSICA TROUNSTINE AND MELODY E. VALDINI no studies that have determined whether or not districts servetoincreaserepresentationwhentheseconditionsare met at the local level. We begin to do so here. CitiesintheUnitedStatestendtoelecttheircitycoun- cils using two electoral system types: single-member dis- tricts or at-large elections. When councilors are elected by district, the city is divided into geographic areas of roughly equal population size that elect a single member to the city council in a plurality or majoritarian contest. An at-large system is one in which members of the city council are selected by the entire city electorate. In most cities this means that voters are offered a slate of candi- dates and are allowed the same number of votes as there areseatsavailable.Citiesoftenhavemajoritarianrequire- ments such that if a candidate does not receive 50% of the vote she is forced into a run-off election. Some cities designateseatsorresidencyrequirementsforat-largepo- sitionsturningtheelectionintoaseriesofsingle-member contests,whileothercitiesvoteforonlyoneat-largemem- ber in any given election. A small but growing number of cities use mixed systems, electing some council members by district and others at-large. Scholars have found that these mixed systems as well as modified at-large systems that employ different vote count procedures lead to de- scriptive representation at levels closer to single-member districts(Brockingtonetal.1998;KarnigandWelch1982; Welch 1990). The number of cities using pure at-large systems has declined over the past 20 years, but at-large elections remain a common feature in city politics. The majorityofcitiesinourstudyelecttheirmembersat-large. In order for district elections to increase the propor- tion of councilors relative to the population size of an underrepresented group, previous literature has posited that three factors might come into play: concentration, size,andpolarizationof thevote.First,thegroupmustbe geographicallyconcentratedtotakeadvantageofdistricts (Sass2000;VedlitzandJohnson1982).If groupmembers are spread throughout the city so that they do not com- pose a simple majority of any one district, presumably districts would not increase representation of the group compared to an at-large system. The size of the group (of voting age citizens) should alsoimpacttheefficacyofdistricts(Braceetal.1988;Bul- lock and MacManus 1990; Grofman and Handley 1989; Leal, Martinez-Ebers, and Meier 2004). At minimum, if the group represents less than one-half of the population needed to elect a single council seat, districts are unlikely toensuregreaterrepresentationthanat-largesystems.Al- ternatively, if a group composes a majority of the city population in a majoritarian, at-large system, the group maybeabletowinallof thecouncilseats.Districtsmight evendecreasethegroup’srepresentationonthecitycoun- cil.Wepredictthatgeographicallyconcentrated,midsized groups will benefit most from district elections. Finally, these expectations rely on an assumption of polarized voting. The group must vote in a substantial bloc for candidates who are members of the group, and other groups must be substantially unwilling to vote for members of the group (Brace et al 1988; Davidson and Korbel1981;EngstromandMcDonald1982).Ifeitherone of these does not hold, it is unclear whether the electoral systemwillhaveanydirecteffectongrouprepresentation. Polarized voting affects different racial and ethnic minorities to different degrees. For instance, the more heterogeneous the group is, the less likely they may be to vote as a bloc, which is particularly important for Latino communities (Pachon 1999). For this reason we expect the effect of districts to be less pronounced for Latinos relative to African Americans. However, we still expect districts to have some impact. Research has found that Latinos share a significant number of characteristics that encourage ethnically based voting, including discrimina- tion, immigrant experiences, Latin American heritage, and Spanish language (see Barreto 2004 for a literature review). Further, due to the lack of partisan identifica- tion in most city-level contests, candidate characteristics like race, ethnicity, or gender may take on additional im- portanceforvoterslookingforinformationalcues(Bobo 1988; Kaufmann 2004; Popkin 1991; Tate 2003; Valdini 2006). Nearly all of the research on racially polarized voting hasconcentratedonstate-andfederal-levelelections(see Hutchings and Valentino 2005 for a review). At the lo- cal level, Hajnal and Trounstine (2005) found that blacks and Latinos tended to vote most cohesively for the same candidate. Across 10 of the United States’largest cities, 74% of blacks and 72% of Latinos voted for the group’s majority preferred candidate. This was compared to 67% of whites voting for the group’s first choice. Additionally they find a significant racial/ethnic divide in votes for the winning candidate, with a 39 percentage point gap be- tweenwhitesandAfricanAmericans,anda20percentage point gap between whites and Latinos. These figures bol- ster our prediction that districts will be most helpful for African Americans. The logic of concentration and group size also works toexplainpotentialdifferentialeffectsofdistrictsforpeo- pleofcolorandwomen.Womenarerarely(ifever)highly concentrated in a community. The same can be said for group size and the representation of women. Because womenarenearlyalwaysbetween48%and52%ofacom- munity’s population, we cannot expect that they will be aided by districts. Furthermore, there is little evidence of gender-polarizedvoting.Anumberof studieshavefound 137 THE CONTEXT MATTERS 557 that voters evaluate female candidates drawing on gen- dered stereotypes (e.g., Dolan 2004; Huddy and Terkild- sen 1993; McDermott 1997; Valdini 2006) and that these stereotypescanaffectperceptionsaboutcandidates(Koch 2000)andvotechoice(Brown1994;Brown,Heighberger, and Shocket 1993; Sanbonmatsu 2002). However, stereotyping only equates to polarized vot- ingwhentherearegendereddifferencesinthejudgments of voters. Some research has determined that women are morelikelytopreferfemalecandidatesandmentoprefer malecandidates(Sanbonmatsu2002).But,otherscholars argue that there is little evidence of gender group con- sciousness (Conover 1988; Gurin 1985) and that women are equally if not more unlikely to vote for female can- didates as men (Darcy and Schramm 1977; Karnig and Walter1976).Intheaggregate,thegendergap(whileper- sistent)tendstobesmallwithregardtosupportforparties and candidates (see Norrander 2003 for a review). So although it is likely that women are treated differ- entlyfrommeninelections,itisunclearhowthesediffer- encesshouldinteractwithinstitutionalvariation.Accord- ingtothecriteriaspecifiedinThornburgv.Gingles (1986), womenareunlikelytobenefitfromdistricts.Scholarshave suggested, among other reasons, that women might do better in multimember elections (such as at-large sys- tems) because the competition is not zero-sum, mean- ing that voters need not choose women at the expense of men (Karnig and Welch 1979; Matland and Brown 1992; Matland and Studlar 1996). On the other hand, scholars have not proposed that the zero-sum calculation applies toracialandethnicminorities.5 Thisimpliesthattheelec- toral structure is predicted to affect racial and ethnic mi- norities in a different way and for different reasons than women. So how should our expectation change when we are talking about women of color; do the predictions for multimember elections only apply to white women? There is some evidence that racial bloc group vot- ing does not apply to women of color, particularly when menofcolorarealsorunning.McClain,Carter,andBrady (2005) find that black women have a harder time gain- ing the support of race-based organizations compared to blackmen,andPhilpotandWalton(2007)findthatblack women are the strongest supporters of black female can- didates.Giventhatourconcentrationandsizehypotheses dependonpolarizedvoting,wemightnotexpectdistricts to help black women. On the other hand, some scholars havefoundthatblackwomenandLatinasare better repre- 5Clearly more research should be done to determine the extent to whichzero-sumcalculationsapplytodifferentgroups.Itispossible that even in at-large settings such a calculation could be invoked, particularly when cities use designated post systems or staggered elections. sented than white women (Darcy and Hadley 1988; Gar- ciaBedolla,Tate,andWong2005;Montoya,Hardy-Fanta, and Garcia 2000). Garcia Bedolla, Tate, and Wong (2005) explain this finding as potentially resulting from block group voting. This would also be supported by Philpot and Walton’s (2007) finding that black men tend to be strongersupportersofblackfemalecandidatesthanwhite women or white men. Further, a number of studies have found that race trumps gender in determining voting be- havior and attitudes (Gay and Tate 1998; Lien 1998) and thatthegendergapisessentiallythesameacrossracialand ethnic groups (Welch and Sigelman 1992). In sum, while weexpectwhitewomentobenefitfromat-largeelections, andblackmentobenefitfromdistricts,therearenoclear hypothesesthatemergeforblackwomenandLatinaswith regard to the effect of institutional structure. Testing the Effects of the Electoral System on Representation To understand more about why single-member districts help certain underrepresented groups and not others, we begin by testing the relationship between electoral struc- ture and diversity in cities. Our data come from surveys by the International City/County Manager’s Association (ICMA) conducted in 1986, 1992, 1996, and 2001. The ICMA survey is mailed to city clerks in approximately 7,500 cities including all municipalities with more than 2,500 residents. The average survey response rate for the years we analyze is 63%. 6 Thesurveyprovidesdemo- graphic information about council members and insti- tutional variables for the cities. To control for city-level demographicsweuse1990censusdataforall1986obser- vations, 2000 census data for all 2001 observations, and linearly interpolate values for 1992 and 1996. In total we have 7,174 unique cities in the data set. 6Determining the effect of response rates to the ICMA is difficult because no other source contains institutional data for the same time period; but we can use data from the 1987 Census of Govern- ments (COG) as a comparison for our main independent variable. The proportions of councilors elected in each type of system are similar in the two data sets. In the 1986 ICMA data 72.3% of cities elected councilors at-large, 11.6% used districts, and 16.2% used a mixed system. The COG reports 74.2% of cities elected councilors at-large, 13.5% used districts, and 12.3% used a mixed system. It does appear that western and southern cities are underrepresented intheICMAcomparedtothecensus.Thecontrolvariablesincluded in our analyses should mitigate the effect of this underrepresenta- tion; nonetheless we add the caveat that our findings are most di- rectlyapplicabletothetypesofcitiesincludedintheICMAsample. Weighting by region does not change our conclusions. Summary statistics for all variables are available from the authors. 138 558 JESSICA TROUNSTINE AND MELODY E. VALDINI In addition to the statistical analysis, we present re- sponses from interviews of current city councilors from a sample of cities with mixed electoral systems. 7 These interviews served a number of purposes in our investi- gation. First, they helped us to identify the contextual factors that interact with institutions and affect the elec- tion of women and minority councilors. Secondly, they provide useful examples of our empirical findings. Fi- nally, they offer face validity of our findings from people working in city politics. As we had hoped, all of the in- terviewees in our sample were familiar with both types of electoral systems and made a decision to run in one type rather than the other. Of the 174 councilors serv- ing in 2006, 98 (or 56%) were female and/or persons of color. We randomly selected one-third of these members foraninterview.Elevencouncilorschosetoparticipatein a phone interview in which we asked respondents open- ended questions regarding the effect of electoral institu- tionsforelectingwhitewomen,womenofcolor,andmen of color. 8 In the quantitative analysis our dependent variables are the proportion of city councils that are black, Latino, and female. Unfortunately, the ICMA survey data do not specify the race of women councilors or the gender (or ethnic background) of those in the included racial cate- gories.Whileitwouldbeidealtoaugmentourdiscussion of womenof colorwithICMAdata,wecannot.However, using data from a different source for 1986 we are able to perform a separate analysis of the effect of districts for blackwomenversusblackmenandLatinasversusLatinos. In all of the analyses our primary independent variable is thepercentageofcouncilorselectedbydistrictineachcity. The majority of cities in our data set have a city council that is either elected wholly by districts or at-large, but somehavemixedsystemsinwhichaportionof thecoun- cil is elected by district and a portion elected at-large. To capture this variation we use a continuous version of the variable. We add to these regressions a number of other insti- tutionalvariablesthathavebeenlinkedtominoritycoun- cil representation either directly or indirectly through turnout and mobilization effects. These controls include nonpartisanversuspartisanelections,mayorcouncilver- 7Thecitiesarethe10largestmixedsystemcities:Houston,Philadel- phia, Charlotte, Jacksonville, Indianapolis, Boston, Washington DC, Denver, Nashville, and New Orleans. 8Weinterviewedsixwhitewomen,twoAfricanAmericanmen,one African American woman, one Latino, and one Latina. Prior to each interview we requested permission to record and quote each councilor. We received consent from all but one council member who is not quoted by name in this manuscript. Transcripts are available from the authors upon request. suscouncilmanagersystems,9 thesizeofthecitycouncil,10 thepresenceof termlimits,andadummyvariablenoting whethercityelectionsareheldconcurrentlywithnational elections.11 Because some city institutions are subject to closer scrutiny as a result of the Voting Rights Act (VRA) and our primary independent variable (elections by dis- trict)mayinfactbetheresultofchallengesbroughtunder the VRA, we include a dummy variable indicating juris- dictions required to secure preclearance as per Section 5. We include citywide socioeconomic variables to account forthepossibilitythatfemaleorminoritypresenceonthe council is linked to wealthier or more educated commu- nities.12 We control for potential region effects and the racial and ethnic makeup of the city population. Latinos have lower citizenship rates and younger populations than whitesandAfricanAmericans,perhapslimitingtheirabil- ity to affect election outcomes (Jones-Correa 1998). To control for this we include a measure of the total propor- tion of the city population that are noncitizens and the proportionthatis18andolder.Toaccountforliberallean- ingcommunitiesthatmightbemorelikelytoelectwomen and minorities, we include a measure of the countywide vote for the Democratic presidential candidate in 1988 andadummyvariableforcentralcities.Tocontrolforthe likelyrelationshipbetweentimeandourindependentand dependent variables, we include year fixed effects (with 2001 as the base category). Finally, in all models we in- clude the population proportions of African Americans, Latinos, and Asians in each city. 13 As was true with our dependentmeasureofracialandethnicrepresentation,we 9Using a more nuanced version of this variable allowing for mayor council systems with a city manager makes no difference to the results. 10Scholarshavearguedthatinsmallercouncilsthevalueofeachseat is greater and therefore less likely to be represented by minorities or women (see Welch and Karnig 1979). 11Ideally we would have also included controls for the city’s vote- countprocedure,butthesedataarenotcollectedbyICMA,andthe large size of the data set precluded us from collecting it. 12Unfortunately, group-specific versions of these demographic variables are not available from the 1990 census so we were unable to test the alternative argument that group resources determine representation (see, for example, Cole 1974; Karnig 1979). 13Ideallytheseproportionswouldbeinreferencetothepopulation of citizens over the age of 18. However, the census did not provide data for citizens over the age of 18 by race/ethnicity for our en- tire time period. We tested alternative formulations of population measures assuming in 2000 a constant citizenship rate across age groups and in 1990 a constant citizenship rate across racial and ethnic groups. The alternative specification made little difference to the results and is available from the authors. 139 THE CONTEXT MATTERS 559 cannot account for racial and ethnic group heterogeneity in these models. 14 Likemostresearchonthistopic,werestrictourresults to cities with substantial minority populations. Rather than select an arbitrary minimum for the size of groups, we allow the threshold to vary by city depending on the size of the city council. An observation is included if the group in question composes at least one-half of the per- centagethatasinglecouncilseatrepresents.15Ouranalysis assumesthatblackresidentswillbethestrongestsupport- ers of black candidates and Latino residents for Latino candidates.Thelargerthecitycounciltheeasieritshould be for any group to win representation. Using a varying threshold takes this into consideration. The mean number of council seats is six, so on av- erage a city is included if the underrepresented group is at least 8% of the city’s total population. We apply this selection criterion regardless of the electoral system em- ployed. When we test the hypothesis that the size of the groupmattersfortheeffectivenessoftheelectoralsystem, werelaxthisselectioncriterionandrestricttheanalysisto citiesthathavenonzeropopulationsofthegroupinques- tion. This allows us to directly test the assumption that a groupwillbenefitmostfromdistrictswhenitspopulation islargerthanone-halfofthepercentageofasinglecouncil seat but less than a majority of the total population. Duetotheextremelylargenumberof citiesthathave nofemaleorminoritycouncilors,weusearandom-effects tobit model to estimate the effects of districts on council representation [yi∗=xi +εi∗,whereyi =yi∗if yi∗>0 &yi =0ifyi∗≤0].16 The model, a maximum-likelihood estimation censored at zero, combines the logic of probit and multiple regression to estimate both the probabil- ity of a council having any female or minority members, andgiventhis,predictstheexpectedproportionoffemale councilors and councilors of color. DodistrictsincreasetheproportionofAfricanAmer- ican,Latino,andwomencouncilors?Theresultsdisplayed in Table 1 confirm that district elections continue to aid minority members in getting elected and are a nominal 14Scholars find that assuming ethnic or racial group homogeneity severelybiasesestimatesofrepresentation(DeSipio1996;Sass2000; Tam 1995), but we have no fix for this problem. 15We repeated the analyses using a 5% threshold of the group in question instead of allowing the threshold to vary based on coun- cil size (available upon request from the authors). The results are extremely similar and our conclusions hold in all cases. 16The likelihood function for each unit is computed using the Gauss-Hermite quadrature. The estimates were stable in multiple tests.AlternatespecificationsusingatobitmodelwithHuber/White clustered standard errors are nearly identical. We further tested weighted models to correct for heteroskedasticity and got similarly strong results. detrimenttowomen.Avarietyofsimulationshelptoclar- ify the relationships between district elections and repre- sentation. We predict the marginal effect of the electoral systemontheproportionofwomenandminoritycouncil members when moving from a system in which a major- ity of the council is elected at-large to a system in which a majority of the council is elected by district, holding all other variables constant at their mean values. First, we predict the effect of districts on the probability of a city having any members of the underrepresented group on the council. Then, we predict the effect of districts on the expected proportion of female and minority council members, weighted by the probability that this value is positive.17 For African Americans, having a majority of coun- cil members elected by district increases the probability of electing any African Americans to the city council by more than 10 percentage points, from 73% in at-large cities to 84% in district cities. 18 The expected propor- tion of African American councilors increases by about five percentage points under districts, from 13% to 18%. Because the average city in our data set has six council members, in order for a group to gain an additional seat districts need to provide about a 16-point advantage. In our model, districts clearly fall short of this threshold for African Americans. The key factor in increasing African American rep- resentation is the proportion of the city that is black. For Latinos, districts have a weaker effect on representation. For both at-large and district systems the probability of having any Latino councilors at all is low; 27% in at-large systems and 33% in district systems. When this is taken into consideration the relationship between district sys- temsandtheexpectedproportionoftheLatinocouncilors islimitedtoabout11 2 percentagepoints,goingfrom4.1% in at-large systems to about 5.5% in district systems. The Latino population in a city plays a key role in the election of Latino council members. Nearly equal in effect is the percentage of the city that has the rights of citizenship. Echoing the results of our regression analyses, nearly every interviewee in our sample agreed that district sys- tems were better than at-large systems for electing peo- ple of color. For example, Councilman Jamie Isabel, an African American member on the Nashville City Coun- cil, explained,“It’s happened again and again where African Americans can’t get enough votes to win at- large.”Susan Burgess, a white woman serving in an at-large seat in the Charlotte City Council, echoed his 17Effects on uncensored observations are also available from the authors. 18Predictions calculated using Stata/SE 9.2 mfx command. 140 560 JESSICA TROUNSTINE AND MELODY E. VALDINI sentiments, stating that districts are “absolutely”better than at-large seats for electing people of color. Coun- cilwoman Rosemary Rodriguez, a Latina serving on the Denver City Council, believes so strongly in the posi- tive consequences of district elections for increasing the representation of people of color that she worked for electoral reform for other local offices in her city. She explains:“I actually persuaded the legislature to adopt single-member districts for Denver for a majority of the school board seats so that we could try to achieve His- panic representation. And ever since that bill was passed, TABLE 1 Tobit Regression on the Percentage of Minority and Female Council Members % Black % Latino % Women Coefficient St Err Coefficient St Err Coefficient St Err % District 0.06∗∗0.01 0.05∗∗0.02 −0.02∗∗0.01 Demographics % Latino 0.18∗∗0.07 1.68∗∗0.08 −0.00 0.03 % Black 0.85∗∗0.03 0.17∗0.09 0.12∗∗0.02 % Asian −0.41∗∗0.21 0.53∗∗0.16 0.08 0.07 % Women 0.43∗∗0.16 −0.53∗0.32 −0.16 0.10 Total Pop (mil) 0.02 0.03 0.02 0.05 0.05∗0.03 %Poor 0.29∗∗0.10 −0.00 0.18 −0.07∗0.04 Med. Income (ths)−0.00 0.00 0.00 0.00 −0.00 0.00 % Coll. Grad 0.23∗∗0.07 0.26∗∗0.13 0.21∗∗0.03 % Noncitizens −0.05 0.15 −1.26∗∗0.14 0.00 0.07 % Pop Over 18 0.02 0.11 −0.02 0.15 0.18∗∗0.05 Democratic Vote −0.14∗∗0.05 0.15∗0.09 −0.01 0.02 Institutions Term Limits −0.00 0.01 0.03 0.02 0.02∗∗0.01 Nonpartisan 0.00 0.01 −0.01 0.03 −0.00 0.01 Mayor Council −0.00 0.01 0.02 0.02 −0.00 0.00 Council Size 0.00 0.00 0.01∗∗0.00 0.00∗∗0.00 Concurrent −0.01 0.02 0.02 0.02 0.01∗0.01 VRA 0.00 0.01 0.04 0.03 Geography Central City 0.07∗∗0.01 0.08∗∗0.02 0.02∗∗0.01 West 0.08∗∗0.03 0.05∗0.03 0.10∗∗0.01 Northeast 0.02 0.02 −0.03 0.04 0.02∗∗0.01 Midwest −0.00 0.01 0.00 0.04 0.04∗∗0.01 1986 −0.07∗∗0.01 −0.03 0.03 −0.07∗∗0.01 1992 −0.02∗∗0.01 −0.01 0.02 −0.02∗∗0.01 1996 −0.01 0.01 −0.00 0.03 0.07∗∗0.01 Constant −0.37∗∗0.15 −0.46∗0.26 0.03 0.06 N 3042 2749 11537 Wald 2 1670.83∗∗1258.24∗∗1668.51∗∗ ∗p <.10,∗∗p <.05. Source: International City/County Manager’s Association (ICMA) surveys of 1986, 1992, 1996, and 2001. we have had a Hispanic member elected to the school board.” The Effect of Institutions for Women For women, the results in Table 1 suggest that the prob- ability of a council having at least one female councilor is high: about 83% in at-large systems and about 80% in district systems, with the expected proportion of female 141 THE CONTEXT MATTERS 561 councilors going from 20% in at-large cities to 18% in cities with district elections. None of the other variables perform particularly well predicting women’s represen- tation. The model suggests that increasing the propor- tion of women in a city negatively impacts the chance of having women on the council. Women are also aided by larger city councils. To test whether or not this finding reflects the benefit of increased district magnitude that other scholars identify, we interact this variable with a dummy variable indicating whether the majority of the councilorsareelectedat-largeorbydistrict.Ourfindings reflect the conventional wisdom (e.g., Alozie and Man- ganaro 1993); women do better with larger city councils, butinat-largecitiesthiseffectismuchmorepronounced. Once the interaction is included, the independent effect of at-large elections actually disappears. This offers indi- rect support for the argument that women benefit from a nonzero-sum setting. The opinions of our interviewees reflect these mud- dled findings of the effects of at-large versus district elections for women. Councilperson Carol Boigon felt strongly that at-large seats are better for electing women candidates. She explained that in her council,“the two at-large seats run at the same time—no differentiation occurs—onerace,twotopvotegettersgetseated.Sothere were seven men and me. So you win by a plurality, which reallystrengthensthehandofwomen.”Councilmembers JamieIsabel,GloriousJohnson,SusanBurgess,andAnna Verna,ontheotherhand,allfeltthatdistrictsarethebet- ter choice for increasing the number of women in office. However, when asked why they felt that district elections are better for electing women, every respondent gave a different answer. CouncilwomanJohnsonexplaineddistrictswerebet- terbecauseoftheabilityofwomentobeactiveandknown within their districts, stating that the women currently serving in district seats on her council “have ...clout when it comes to that district because they have been knownsincetheywerechildren.”CouncilwomanBurgess suggestedthatdistrictsarebetterforwomenbecausethey are less competitive. She explained,“Once you win a dis- trictelection,manytimesthedistrictrepresentativesdon’t even have competition in their subsequent elections. At- large is always competitive. Very tough races, to be truth- ful and we have only one woman and three men.”Three other council members all argued, however, that it was a toss-up and/or that the election of women depended on factors specific to each electoral contest, not the electoral system. Councilwoman Sanders of Indianapolis stated, “Idon’t know that there’s really much difference [be- tweenat-largeversusdistrictelections],atleastnotinmy experience.” Clearly, the variety of responses and opinions given on this subject is quite different from the nearly uniform responses given on the effects of district elections for the election of people of color. The wide range of responses regarding the effect of institutions on the representation ofwomenisnottoosurprisinggivenourstatisticalresults. Wefoundthatwomenarenegativelyaffectedbydistricts, but the results were small, with the predicted proportion of women increasing only about 2% in at-large cities. In sum, after controlling for a variety of factors, it appears that districts have a limited but distinctly positive effect on increasing representation for underrepresented racial and ethnic groups and a small negative effect for women that appears to be driven by the multimember nature of at-large elections. The Intersection of Race and Gender Giventhatracialandethnicminoritygroupsseemtoben- efit from districted systems while women seem to do bet- ter in at-large systems, how do black women and Latinas fareinthesecities?Ourmaindatasetdoesnotprovidethe racialandethnicbackgroundbygenderofcitycouncilors; however, the United States Census of Governments col- lected these statistics in one year that matches our data— 1986.Inthisyearabout16%ofwhiteandblackcouncilors and about 18% of Hispanic councilors were women. For the following analyses we use as dependent variables the proportion of the city council that is black women, black men,Latinas,Latinos,andwhite(non-Hispanic)women. We include all of the controls described above. As above we only include cities with substantial minority and fe- male populations. 19 The results in Table 2 suggest differential effects of districtsforblackwomenandLatinas.Wheredistrictelec- tions have no significant effect on increasing the propor- tion of councilors who are black women, black men get a significant boost from this institutional structure. In fact allofthepredictedincreaseinrepresentationfoundinTa- ble1isattributabletoblackmen.Thepredictedprobabil- ityofacouncilhavinganyblackwomenisabout13%,and the expected proportion of black women is about 1.6% regardless of the electoral system. The probability of a council having any black men is much higher, about 53% 19We chose not to select cities on the combined basis of race and gender (e.g., only including cities with a substantial population of blackwomen)becausepreviousscholarshiphasindicatedthatrace is a more important predictor of vote choice than gender. Thus we assume that the presence of black men and Latinos are important for the election of black women and Latinas. 142 562 JESSICA TROUNSTINE AND MELODY E. VALDINI TABLE 2 Tobit Regression on the Percentage of Council Members of Color by Gender % Black Women % Black Men % Latinas % Latinos % White Women Coefficient St Er Coefficient St Er Coefficient St Er Coefficient St Er Coefficient St Er % District −0.00 0.03 0.08∗∗0.02 0.04 0.09 −0.01 0.05 −0.02 0.01 Demographics % Latino 0.29 0.28 0.08 0.15 0.97∗∗0.27 1.63∗∗0.15 −0.15∗∗0.05 % Black 0.54∗∗0.10.58∗∗0.05 0.67∗∗0.33 0.27 0.19 −0.08∗∗0.04 % Asian 1.36∗∗0.63 −0.03 0.36 0.43 0.60.21 0.36 −0.09 0.13 % Women −0.19 0.57 0.49 0.31 −0.02 1.22 0.58 0.78 −0.36∗0.19 Total Pop (mil)−0.06 0.11 0.01 0.05 0.07 0.13 −0.04 0.08 0.06 0.05 %Poor 0.01 0.36 0.42∗∗0.19 0.67 0.60.25 0.36 −0.03 0.07 Med. Income (ths)−0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.00 0.00∗0.00 % Coll. Grad 0.42∗0.23 0.01 0.13 0.06 0.54 0.13 0.32 0.19∗∗0.05 % Noncitizens −0.61 0.64 −0.07 0.34 −0.70.46 −1.21∗∗0.26 0.00 0.14 %PopOver18 −0.09 0.45 −0.07 0.24 0.25 0.79 0.17 0.47 0.32∗∗0.09 Democratic Vote −0.03 0.15 −0.20∗∗0.08 −0.35 0.35 0.15 0.18 0.03 0.03 Institutions Term Limits 0.12∗∗0.06 −0.04 0.04 −0.04 0.11 0.08 0.06 0.04∗∗0.02 Nonpartisan 0.01 0.04 0.00 0.02 0.04 0.12 −0.04 0.06 0.00 0.01 Mayor Council −0.03 0.03 −0.03∗∗0.02 −0.13 0.09 0.03 0.04 −0.02∗0.01 Council Size 0.01 0.01 0.00 0.00 0.03∗0.02 0.02 0.01 0.01∗∗0.00 Concurrent 0.03 0.05 0.00 0.03 0.06 0.08 0.04 0.05 0.01 0.01 VRA −0.03 0.04 −0.01 0.02 −0.03 0.10 0.04 0.06 Geography Central City 0.12∗∗0.04 0.08∗∗0.02 0.01 0.09 0.10∗0.05 0.04∗∗0.01 West −0.14 0.11 0.03 0.05 0.03 0.10 0.09 0.06 0.12∗∗0.01 Northeast 0.06 0.06 0.00 0.03 −0.12 0.19 −0.15 0.11 0.01 0.01 Midwest 0.00 0.05 −0.02 0.02 −0.15 0.20 −0.02 0.10 0.04∗∗0.01 Constant −0.35 0.55 −0.38 0.3 −1.44 1.03 −1.45∗∗0.63 −0.13 0.12 N 893 893 698 698 3563 Wald 2 62.12∗∗303.92∗∗30.31 321.97 ∗∗430.62∗∗ ∗p <.10,∗∗p <.05. Source: United States Census of Governments 1986. in at-large councils and nearly 70% for district councils. The expected proportion goes from 8% in at-large cities to 14% in districted cities. For Hispanics the story is different. The electoral structure has no significant effect on the proportion of the council that is Latino or Latina. However, Latinos are muchmorelikelytoberepresentedoncouncils.Theprob- ability of having any Latinos on the council is about 21% andtheexpectedproportionabout3.5%,whiletheproba- bilityofhavingLatinasonthecouncilisabout4%andthe expectedproportionlessthan1%.Finally,intheseresults itappearsthatthepositiveeffectof at-largeelectionsisall goingtowardwhitewomen,althoughthecoefficientdoes not quite reach statistical significance. The probability of a council having any white women increases from 64% underdistrictsto67%inat-largecities,withtheexpected proportion increasing from 12% to 13%. In sum, black men and white women are the only groups in our analysis that are substantively and signif- icantly affected by electoral institutions, and the biggest benefit of the system appears to be increasing the prob- ability of having any black men or white women, rather than the proportion. Concentration of Population Matters The reason that the electoral system may have a relatively small overall effect for racial and ethnic minorities may 143 THE CONTEXT MATTERS 563 lieinpopulationsizeandresidentialconcentrationof the groups.Wenowlooktoseewhetherdistrictshavealarger effect if these factors are taken into consideration. As op- posed to women, African American and Latino voters can be heavily concentrated. The theory that concentra- tiondrivestherelationshipbetweendistrictelectionsand representation of racial and ethnic minorities has a sub- stantial number of subscribers (Brace et al. 1988; David- son and Korbel 1981; Engstrom and McDonald 1982; Mladenka 1989; Vedlitz and Johnson 1982). Yet there have been few attempts to actually test this claim directly (Sass 2000 is an exception). We use 1990 and 2000 census data on concentration in 331 metropolitan areas to do so here.20 Demographersrelyonavarietyofdifferentmeasures of racial and ethnic concentration and segregation calcu- latedusingdemographicdatacollectedatthecensus-tract level (Massey and Denton 1988). One measure is the iso- lation index, which ranges from 0 to 1 and represents the probability that group members will meet members of their own group in their census tract. A score of .6 for AfricanAmericansmeansthattheaverageAfricanAmer- ican lives in a census tract that is 60% black. This mea- sure has the benefit of being sensitive to a group’s size in addition to the distribution of the group throughout a community. It would be impossible to have a high isola- tion score unless a group composes a substantial portion of the total community. Both factors are likely impor- tant for a group to transform membership into voting strength. Inordertoanalyzetheeffectofconcentration,wesplit our data into four samples based on the isolation index for each group and run the same tobit models presented aboveforeachquartile.21 Wehypothesizethatthebenefits of district elections should be most likely if a group can reasonably generate a voting majority in some neighbor- hoods. In other words we predict increasingly significant effectsastheisolationindexincreases.Thefollowinganal- yses are restricted to cities with nonzero populations of AfricanAmericansandLatinos.Wepresentonlythevari- ablesofinterest,butthemodelsincludeallofthecontrols listed in Table 1. The full models are available from the authors. 20Thecensusdidnotproduceconcentrationdataatthemunicipal- ityleveluntil2000,andthenonlyforlargecities.Toshowtheeffects of concentration across time for as many cases as we can, we rely on the MSA-level statistics, but because of the potential mismatch between MSA- and city-level concentration, we rerun the analysis using 2000 data at the city level for 596 cases. 21Weelectednottopresentaninteractionmodelbecausetheeffect is nonlinear. We had enough data to estimate the effect in a split sample allowing the coefficients to vary. The results in Table 3 are clear. Only when a group is concentrated will districts promote increased descrip- tive representation on the council. For African Ameri- cans, the effect of districts goes from being negative at very low levels of concentration to significantly positive at high levels. Districts have the largest effect for cities in the third quartile, where moving from an at-large sys- tem to a district system increases the estimated probabil- ity of electing an African American council member by about 10 percentage points, from 14% to 24%. This is a powerful effect compared to the first quartile, where dis- trictsdecreasedboththeprobabilityofhavinganyAfrican American councilors (from 7% to 3%) and the expected proportion from .9% to .3%. When the isolation index is very high for African Americans the effect of districts becomes insignificant. This could indicate the decreased importance of the electoral system when a group makes up a majority of the electorate. The size of the black pop- ulation is most dominant in the first and fourth models, suggesting that African American council representation in cities at the two ends of the isolation spectrum is best predicted by the size of the minority group itself. We repeated this analysis with the 1986 census data using the proportion of black men and the proportion of black women as dependent variables. As expected the re- sultsholdsystematicallyfortheelectionofblackmen,but not black women. For black men districts have a negative effect in the first quartile and an increasingly powerful effect in the second through fourth quartiles. For black women the electoral institution has no effect in the first through third quartiles, but districts are extremely pow- erful and positive in the fourth quartile. When blacks compose a majority of a city’s population, districts—not at-large elections—help black women. As shown in the bottom half of Table 3, for Latinos, the effect of the interaction between districts and con- centration is even more striking. The effect of districts is small and highly insignificant in the first and second quartiles.Theeffectinthethirdquartileissubstantialbut not quite statistically significant. Unless Latinos are ex- tremely concentrated, districts make little difference for representation. In the fourth quartile, the impact is large. DistrictsincreasetheprobabilityofelectingLatinostothe council to 98% from 75% under at-large systems. 22 The predicted proportion of Latinos on the council increases by more than 25 percentage points from 19% in at-large cities to 48% in district systems. 23 22WecouldnotrunthesemodelsonLatinosandLatinasseparately because of a lack of data. 23Using city-level isolation measures from 2000 for large cities, the resultsareextremelysimilarthoughnotexactlythesame.Forblacks 144 564 JESSICA TROUNSTINE AND MELODY E. VALDINI TABLE 3 TobitRegressiononProportionofAfricanAmericanCouncilMembers,Controllingfor Concentration 0 ≤isolation <.25 .25 ≤isolation <.50 .50 ≤isolation <.75 .75 ≤isolation ≤1 Coefficient StErr Coefficient StErr Coefficient StErr Coefficient St Err % District −0.13∗∗0.05 0.04∗∗0.02 0.08∗∗0.02 0.06 0.04 % Black 3.15∗∗0.48 1.29∗∗0.08 1.22∗∗0.06 1.31∗∗0.10 VRA 0.14∗∗0.06 −0.03∗∗0.02 0.06∗∗0.03 −0.90 73.3 Constant −0.92 0.60 −0.05 0.22 0.13 0.26 1.44 0.68 N 1373 2247 2735 908 Wald 2 94.15∗∗680.00∗∗902.92∗∗383.65∗∗ TobitRegressiononProportionofLatinoCouncilMembers,ControllingforConcentration 0 ≤isolation <.25 .25 ≤isolation <.50 .50 ≤isolation <.75 .75 ≤isolation ≤1 Coefficient StErr Coefficient StErr Coefficient StErr Coefficient St Err % District 0.03 0.11 −0.00 0.04 0.06 0.05 0.32∗∗0.16 % Latino 4.63∗∗0.94 1.53∗∗0.17 1.53∗∗0.15 2.14∗∗0.40 % Noncitizens −4.98∗∗1.67 −1.22∗∗0.35 −0.97∗∗0.23 −0.64 0.67 Constant −2.80 1.73 −0.35 0.48 −0.53 0.46 1.61 1.19 N 4542 1966 968 133 Wald 2 163.28∗∗241.99∗∗427.35∗∗198.88∗∗ ∗p <.10,∗∗p <.05. Source:U.S.CensusBureau1990&2000;InternationalCity/CountyManager’sAssociation(ICMA)surveysof 1986,1992,1996,and2001. Thereareatleasttworeasonsthattheeffectofdistricts occurs only at the highest levels of isolation for Latinos. First, Latinos have lower levels of turnout compared to their population proportions than African Americans or whites. This may indicate that Latinos need to be a larger share of the population before they can affect electoral outcomesasagroup.Secondly,theseresultsmayindicate a lower level of polarized voting and bloc group voting until Latinos are highly concentrated and a large portion ofthepopulation.Itisplausiblethatethnicgroupidentity would be associated with a higher degree of segregation and a larger minority group presence. In general though, thesedataprovidestrongsupportforourhypothesisthat thebenefitofsingle-memberdistrictelectionsforminor- ity groups operates through patterns of concentration. Furthermore, the negative results for African Americans in the first quartile of isolation suggest that underrepre- sented groups may fare better in at-large elections when they are highly dispersed throughout the community. the effect of districts is negative in the first quartile, small and insignificantinthesecondquartile,andincreasinglypositiveinthe third and fourth quartiles. For Latinos the effect is negative in the first quartile, nearly zero in the second and fourth quartiles, and very powerful in the third quartile. This is precisely the conclusion drawn by one of our interviewees, Councilman Felix Arroyo, the first Latino elected to the Boston City Council. Councilman Arroyo statedthathechosetorunforanat-largeseatratherthan the district seats also available because “it is very difficult to win if you are a person of color by district except for two districts which are actually communities of colors.” Further,heexplainedthatbecauseofthedemographicsof thecity,theat-largeseatwasbetterforelectingLatinosin Boston,“becausemostoftheLatinocommunityisspread across the city, as well as the immigrant community and the progressive groups.”For Arroyo, the lack of concen- tration of his primary constituency means that districts do not offer him the best opportunity for election. Councilor Carol Boigon, a white woman serving on the Denver City Council, also emphasized the power of district elections when groups are concentrated. She ex- plains: ...in a district seat...some of the ethnic con- centrations have an opportunity to be repre- sented ....That’stheadvantageIsee.Wehavetwo seats that could reliably elect a black council per- son, the 8 th and 11 th. And those of us who feel 145 THE CONTEXT MATTERS 565 that it’s important to have diverse voices at the table and who are not black, you would say “Why would you run from one of those seats then?”I wouldn’t. CouncilwomanBoigon’spointisclear:AfricanAmerican council members benefit electorally in areas of the city that have large concentrations of black voters. Size of the Group Interacts withDistricts We hypothesized that the effect of districts should inter- act with group size, being most effective when groups are moderately sized. Our use of the isolation variable in theprevioussectiontestedthisindirectly,butbecausethe indexcombinesconcentrationwithgroupsize,inthissec- tionwetestthishypothesisdirectlybysplittingoursample by the size of the underrepresented group. 24 We divided ourdataintothreesubsamplesforAfricanAmericansand Latinos.Thefirstsampleincludescitiesthathaveminority populations greater than zero, but less than the one-half of the percentage that a council seat represents. The sec- ond sample includes cities with minority groups equal to or larger than one-half of the percentage the council seat represents,butlessthanamajorityofthepopulation.The third sample includes cities where the group in question composes a majority of the population. The results confirm our expectations. Districts mat- ter most for groups that are a moderate proportion of the population. For very small and very large groups the electoral system has no significant effect on representa- tion in the models. Rather than present these as regres- sion results, we have included a graph of the benefit of districts compared to at-large systems for African Amer- icans and Latinos depending on the size of the black or Latino population. The results in Figure 1 suggest that districts have a muchstrongereffectforAfricanAmericansthanLatinos, as would be expected given the differences in concentra- tionandblocvotingbetweenthegroups.Forbothgroups the only statistically significant differences between elec- toral systems occur when the group is moderately sized. 24Asinglemodelincludingtheinteractionsbetweendistricts,group size, and dummy variables for subsamples also generated signifi- cantresults.Thereisnosignificantdifferencebetweenat-largeand district systems when a group is very small; increasing the group population increases the proportion of minority group members anddistrictsenhancethisresult.Wepresentthesplitsampleanalysis because the results are easier to interpret. In no case does the electoral system bring a group to representational parity, but in cities where there are very largepopulationsofAfricanAmericansandLatinos,there is virtually 100% probability of at least a single council memberbeingAfricanAmericanorLatino.Thissuggests that African Americans and Latinos are breaking into the politicalsystemwhentheycommandasubstantialvoting bloc. Inourinterviewsanumberofcouncilorsemphasized groupsizeincombinationwithconcentrationintheirdis- cussion of the superiority of district for electing people of color. Councilman Jamie Isabel, an African American memberontheNashvilleCityCouncil,explainedthatdis- trictsarebetterbecauseat-largesystems “dilutethevotes.” Similarly,CouncilorSusanBurgess,awhitewomanonthe Charlotte City Council, stated that ...the reason is because we have drawn our dis- tricts to make sure we have minority represen- tation. Three [out of seven] of our districts are majority-minority. And there’s always a minor- ity elected there. We have had difficulty electing minoritiesat-large,evenwhentheyareextremely qualified...We’ve had awesome African Ameri- cancandidateswhohavetriedtogofromdistricts to at-large and lost citywide. Councilperson Joanne Sanders, a white woman serving in the Indianapolis council (which is consolidated with the county), nearly repeated the sentiments of Council- woman Burgess, stating that because of the demographics of the county, I think that the district level was better for peo- ple of color ...we still have heavily black areas, in our communities, where it’s easily a seventy- thirty Democrat district. So for people of color that’smucheasierthantryingtoruncounty-wide where some of the outlying areas are predomi- nantlywhite.Although,theblackpeoplewhohave runontheat-largetickethavebeensuccessfulbut again you can tell by the numbers unfortunately theydon’talwaysgleanthemostamountofvotes. Councilor Sanders’s response hints at the presence of raciallypolarizedvotinginIndianapolis.Otherinterview- ees expressed similar sentiments. When asked why black candidates had been unsuccessful in winning at- large seats, Councilman Isabel ascribed the outcome to polarization: 146 566 JESSICA TROUNSTINE AND MELODY E. VALDINI FIGURE 1 EstimatedBenefitofDistrictsComparedtoAt-LargeElectionsbyGroup Population Size, 1986–2001 -0.02 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 Difference in ProbabilityTiny Population Moderate Population Majority Population -0.01 0.00 0.01 0.02 0.03 0.04 0.05 0.06 Difference in ProportionAfrican American Latino Probability of Having Any Black or Latino Councilors Predicted Proportion of Black and Latino Councilors Source:U.S. Census Bureau 1990 & 2000; International City/County Manager’s Association (ICMA) surveys of 1986, 1992, 1996, and 2001. IthinkwhiteshaveareluctancetovoteforAfrican Americans.We’reinthesouth,Ithinkupnorthor back where you are out west, there may be some differences.ButIthinkhereinthesouthwhitesre- allyhaven’tcometotherealitythatAfricanAmer- icans can represent them well. Similarly, Councilor Burgess suggested that “subtle racism”kept African Americans from winning citywide elections. A number of our interviewees also insinuated thatsomegroupsformedmorecohesivevotingblocsthan others.CouncilorRodriguezfromDenverhighlightedthe benefits of districts for Latino candidates because of the strategy of “single-shot[ting]”where voters pool votes in a multicandidate, at-large race for a single candidate. She told us that the African American community used thisapproachsuccessfullytoelectrepresentativeswhereas Latinos tended to divide their votes among a slate of can- didates and so “every time a Hispanic candidate would run, they would be defeated.”According to these coun- cil members, racially polarized voting continues to be a significant factor in city council elections, and different groupsareaffectedindifferentwaysbythesetypesofvote patterns. Conclusions One final possibility in explaining the benefit of districts for female and minority council membership is the at- traction of running in a district versus citywide election. Districts might aid racial and ethnic minorities because more traditionally underrepresented candidates choose to run in district races. If the organizing, fundraising, 147 THE CONTEXT MATTERS 567 and campaign costs are substantially different between districtsandcitywideraces,thismaywellbeafactor.Fur- thermore,itremainstobeexplainedwhysomecitieselect more women than others: it does not appear to make a substantial difference whether women run in district or at-largeelectoralsystems.Nonetheless,webelievethatwe have taken a step forward in explaining the effect of elec- toral systems on underrepresented groups. By taking advantage of the institutional variation across cities in the United States, we have gained a more nuanced understanding of the representation of women and racial and ethnic minorities. Single-member district systems can increase diversity only when underrepre- sentedgroupsarehighlyconcentratedandcomposemod- erate portions of the population. These factors are most important in an arena where polarized voting predomi- nates and where groups leverage their population size to achieve descriptive representation. In addition, the effect of the electoral system is not constant across all people of color, nor is it constant across both genders; race and gender interact to produce different results. Our findings demonstrate the need for caution when making declara- tions of the benefit or detriment of institutional settings; while the electoral rules certainly have an effect, the con- text in which they are employed is also crucial to gain a complete understanding. References Alozie, Nicholas, and Lynne Manganaro. 1993.“Women’s CouncilRepresentation:MeasurementImplicationsforPub- lic Policy.”Political Research Quarterly 46(2): 383–98. Arrington, Theodore, and Thomas Gill Watts. 1991.“The Elec- tion of Blacks to School Boards in North Carolina.”The Western Political Science Quarterly 44(4): 1099–1105. Barreto, Matt A. 2004.“The Role of Latino Candidates in Mo- bilizing Latino Voters: Revisiting Latino Vote Choice.”Pre- paredforLatinoPolitics:TheStateof theDiscipline.College Station: Texas A&M University. Bobo, Lawrence. 1988.“Group Conflict, Prejudice, and the Paradox of Contemporary Racial Attitudes.”In Eliminat- ing Racism: Profiles in Controversy, ed. Phyllis A. Katz and Dalmas A. Taylor. New York: Plenum Press, 85–116. Brace, Kimball, Bernard Grofman, Lisa Handley, and Richard Niemi.1988.“MinorityVotingEquality:The65PercentRule in Theory and Practice.”Law and Policy 10(1): 43–62. Bridges, Amy. 1997.Morning Glories: Municipal Reform in the Southwest. Princeton, NJ: Princeton University Press. Brockington,David,ToddDonovan,ShaunBowler,andRobert Brischetto. 1998.“Minority Representation under Cumula- tiveandLimitedVoting.”TheJournalofPolitics 60(4):1108– 25. Brown, Clyde. 1994.“Judgments about the Capabilities of City Councilors and Support for Female Representation on City Council.”Social Science Journal 31(4): 355–73. Brown, Clyde, Neil Heighberger, and Peter Shocket. 1993. “Gender-Based Differences in Perceptions of Male and Fe- male City Council Candidates.”Women & Politics 13(1): 1– 17. Bullock, Charles, and Susan MacManus. 1990.“Structural Fea- turesof MunicipalitiesandtheIncidenceof HispanicCoun- cilmembers.”Social Science Quarterly 71(4): 665–81. Bullock,CharlesS.,III,andSusanA.MacManus.1991.“Munic- ipalElectoralStructureandtheElectionof Councilwomen.” Journal of Politics 53(1): 75–89. Bullock,CharlesS.,III,andSusanA.MacManus.1993.“Testing Assumptions of the Totality-of-the-Circumstances Test: An Analysis of the Impact of Structures on Black Descriptive Representation.”AmericanPoliticsQuarterly 21(3):290–306. Cole, Leonard. 1974.“Electing Blacks to Municipal Office: Structural and Social Determinants.”Urban Affairs Quar- terly 10(1): 17–39. Conover, Pamela. 1988.“The Role of Social Groups in Political Thinking.”British Journal of Political Science 18(1): 51–96. Darcy, Robert, and Charles Hadley. 1988.“Black Women in Politics—The Puzzle of Success.”Social Science Quarterly 69(3): 629–45. Darcy, Robert, Charles D. Hadley, and Jason F. Kirksey. 1993. “ElectionSystemsandtheRepresentationofBlackWomenin AmericanStateLegislatures.”WomenandPolitics 13:73–89. Darcy, Robert, and Sarah Schramm. 1977.“When Women Run Against Men.”Public Opinion Quarterly 41(1): 1–12. Darcy, Robert, Susan Welch, and Janet Clark. 1985.“Women Candidates in Single-Member and Multimember Districts: American State Legislative Races.”Social Science Quarterly 66(4): 945–53. Davidson, Chandler, and Bernard Grofman, eds. 1994.Quiet Revolution in the South: The Impact of the Voting Rights Act, 1965–1990. Princeton, NJ: Princeton University Press. Davidson, Chandler, and George Korbel. 1981.“At-Large Elections and Minority-Group Representation: A Re- Examination of Historical and Contemporary Evidence.” The Journal of Politics 43(4): 982–1005. DeSipio, Louis. 1996.Counting on the Latino Vote: Latinos as a New Electorate. Charlottesville: University of Virginia Press. Dolan, K. 2004.“The Impact of Candidate Sex on Evaluations of Candidates for the US House of Representatives.”Social Science Quarterly 85(1): 206–17. Engstrom, Richard, and Michael McDonald. 1982.“The Un- derrepresentation of Blacks on City Councils: Comparing the Structural and Socioeconomic Explanations for South/ Nonsouth Differences.”Journal of Politics 44(4): 1088–99. Garcia Bedolla, Lisa, Katherine Tate, and Janelle Wong. 2005. “IndelibleEffects:TheImpactofWomenofColorintheU.S. Congress.”In Women and Elective Office: Past, Present, and Future, 2nd ed., ed. Sue Thomas and Clyde Wilcox. Oxford: Oxford University Press, 152–75. Gay, Claudine, and Katherine Tate. 1998.“Double Bound: The ImpactofGenderandRaceonthePoliticsofBlackWomen.” Political Psychology 19(1): 169–84. Githens, Marianne, and Jewell Prestage, eds. 1977.A Portrait of Marginality:ThePoliticalBehaviorofAmericanWomen.New York: McKay. Grofman, Bernard, and Lisa Handley. 1989.“Black Represen- tation: Making Sense of Electoral Geography at Different 148 568 JESSICA TROUNSTINE AND MELODY E. VALDINI Levels of Government.”Legislative Studies Quarterly 14(2): 265–79. Guinier, Lani. 1992.“Voting Rights and Democratic Theory— WhereDoWeGoFromHere?”In Controversies in Minority Voting: The Voting Rights Act in Perspective, ed. B. Grofman and C. Davidson. Washington, DC: The Brookings Institu- tion, 283–91. Gurin,Patricia.1985.“Women’sGenderConsciousness.”Public Opinion Quarterly 49(2): 143–63. Hajnal, Zoltan, and Jessica Trounstine. 2005.“Where Turnout Matters: The Consequences of Uneven Turnout in City Pol- itics.”Journal of Politics 67(2): 515–35. Heilig, Peggy, and Robert Mundt. 1983.“Changes in Repre- sentational Equity: The Effect of Adopting Districts.”Social Science Quarterly 64(2): 393–97. Herrick, Rebekah, and Susan Welch 1992.“The Impact of At- Large Elections on the Representation of Black and White Women.”National Political Science Review 3: 62–77. Huddy, Leonie, and Nayda Terkildsen. 1993.“The Conse- quences of Gender Stereotypes for Women Candidates at Different Levels and Types of Office.”Political Research Quarterly 46(3): 503–25. Hutchings, Vincent, and Nicholas Valentino. 2005.“The Cen- tralityof RaceinAmericanPolitics.”AnnualReviewofPolit- ical Science 7: 383–408. Jones-Correa,Michael.1998.BetweenTwoNations:ThePolitical Predicament of Latinos in New York City. Ithaca, NY: Cornell University Press. Karnig, Albert. 1979.“Black Resources and City Council Rep- resentation.”Journal of Politics 41(1): 134–49. Karnig,Albert,andSusanWelch.1979.“SexandtheEthnicDif- ferences in Municipal Representation.”Social Science Quar- terly 60(3): 465–81. Karnig,Albert,andSusanWelch.1982.“ElectoralStructureand BlackRepresentationonCityCouncils.”SocialScienceQuar- terly 63(1): 99–114. Karnig,AlbertK.,andOliverWalter.1976.“Electionof Women to City Councils.”Social Science Quarterly 56(4): 605– 13. Kaufmann, Karen. 1994.The Urban Vote: Group Conflict and MayoralVotingBehaviorinAmericanCities.AnnArbor:Uni- versity of Michigan Press. Koch, J. 2000.“Do Citizens Apply Gender Stereotypes to In- fer Candidates’Ideological Orientations?”Journal of Politics 62(2): 414–29. Leal,David,ValerieMartinez-Ebers,andKennethJ.Meier.2004. “The Politics of Latino Education: The Biases of At-Large Elections.”Journal of Politics 66(4): 1224–44. Lien, Pei-Te. 1998.“Does the Gender Gap in Political Attitudes and Behavior Vary across Racial Groups?”Political Research Quarterly 51(4): 869–94. Massey,DouglasS.,andNancyDenton.1988.“TheDimensions of Residential Concentration.”Social Forces 67(2): 281–315. Matland,Richard,andDonleyStudlar.1996.“TheContagionof Women Candidates in Single-Member District and Propor- tional Representation Electoral Systems: Canada and Nor- way.”Journal of Politics 58(3): 707–33. Matland,RichardE.1995.“HowtheElectoralSystemStructure HasHelpedWomenClosetheRepresentationGap.”InClos- ing the Gap, Women in Nordic Politics, ed. Lauri Karvonen and Per Selle. London: Dartmouth, 281–312. Matland, Richard E., and Deborah Dwight Brown. 1992.“Dis- trict Magnitude’s Effect on Female Representation in U.S. State Legislatures.”Legislative Studies Quarterly 17(4): 469– 92. McClain,Paula,NiambiCarter,andMichaelBrady.2005.“Gen- derandBlackPresidentialPolitics:FromChisholmtoMose- ley Braun.”Journal of Women Politics & Policy 27(1–2): 51– 68. McDermott, M. 1997.“Voting Cues in Low-Information Elec- tions: Candidate Gender as a Social Information Variable in ContemporaryUnitedStatesElections.”AmericanJournalof Political Science 41(1): 270–83. Mladenka, Kenneth R. 1989. Barriers to Hispanic Employment Success in 1200 Cities.Social Science Quarterly 70(2): 391– 407. Montoya, Lisa J., Carol Hardy-Fanta, and Sonia Garcia. 2000. “Latina Politics: Gender, Participation, and Leadership.”PS: Political Science and Politics 33(3): 555–61. Mundt, Robert, and Peggy Heilig. 1982.“District Representa- tion: Demands and Effects in the Urban South.”Journal of Politics 44(4): 1035–48. Norrander,Barbara.2003.“TheIntrapartyGenderGap:Differ- ences between Male and Female Voters in 1980–2000 Pres- idential Primaries.”PS: Political Science and Politics 36(2): 181–86. Norris, Pippa. 1985.Politics and Sexual Equality.Boulder,CO: Reinner. Pachon, Harry. 1999.“California Latino Politics and the 1996 Elections: From Potential to Reality.”In Awash in the Main- stream, ed. Rodolfo O. de la Garza and Louis DeSipio. Boul- der, CO: Westview Press, 167–90. Philpot, Tasha, and Hanes Walton, Jr. 2007.“One of Our Own:BlackFemaleCandidatesandtheVotersWhoSupport Them.”American Journal of Political Science 51(1): 49–62. Polinard, Jerry, Robert Wrinkle, and Tomas Longoria. 1991. “The Impact of District Elections on the Mexican Ameri- can Community: The Electoral Perspective.”Social Science Quarterly 72(3): 608–14. Popkin,Samuel.1991.TheReasoningVoter.Chicago:University of Chicago Press. Rice,BradleyRobert.1977.ProgressiveCities.Austin:University of Texas Press. Rule,Wilma.1992.“MultimemberLegislativeDistricts:Minor- ity and Anglo Women’s and Men’s Recruitment Opportu- nity.”In United States Electoral Systems: Their Impact on Women and Minorities, ed. Wilma Rule and Joseph F. Zim- merman. New York: Greenwood Press, 57–72. Rule, Wilma. 1994.“Parliaments of, by, and for the People: Ex- cept for Women?”In Electoral Systems in Comparative Per- spective: Their Impact on Women and Minorities, ed. Wilma Rule and Joseph F. Zimmerman. Westport, CT: Greenwood, 15–30. Sanbonmatsu, Kira. 2002.“Gender Stereotypes and Vote Choice.”American Journal of Political Science 46(1): 20–34. Sass, Tim R. 2000.“The Determinants of Hispanic Representa- tion in Municipal Government.”Southern Economic Journal 66(3): 609–30. 149 THE CONTEXT MATTERS 569 Schwindt-Bayer,LeslieA.,andWilliamMishler.2005.“AnInte- gratedModelofWomen’sRepresentation.”JournalofPolitics 67(May): 407–28. Taebel,Delbert.1978.“MinorityRepresentationonCityCoun- cils: Impact of Structure on Blacks and Hispanics.”Social Science Quarterly 59(1): 142–52. Tam, Wendy K. 1995.“Asians—A Monolithic Voting Bloc?” Political Behavior 17: 223–49. Tate, Katherine. 2003.Black Faces in the Mirror: African Ameri- cansandTheirRepresentativesintheU.S.Congress.Princeton, NJ: Princeton University Press. Thornburg v. Gingles, 478 US 30 (1986). Valdini, Melody Ellis. 2006.“Electoral Institutions and Infor- mation Shortcuts: The Effect of Decisive Intraparty Com- petition on the Behavior of Voters and Party Elites.”PhD dissertation. University of California, San Diego. Vedlitz, Arnold, and Charles Johnson. 1982.“Community Racial Concentration, Electoral Structure, and Minority Representation.”Social Science Quarterly 63(4): 729–36. Welch, Susan. 1990. The Impact of At-Large Elections on the Representation of Blacks and Hispanics.Journal of Politics 52: 1050–76. Welch, Susan, and Albert Karnig. 1979.“Correlates of Female Office Holding in City Politics.”Journal of Politics 41(2): 478–91. Welch, Susan, and Lee Sigelman. 1992.“A Gender Gap among Hispanics? A Comparison with Blacks and Anglos.”The Western Political Quarterly 45(1): 181–99. Welch, Susan, and Donley Studlar. 1990.“Multimember Dis- tricts and the Representation of Women: Evidence from BritainandtheUnitedStates.”JournalofPolitics 52(2):391– 412. 150 At-Large versus Ward Elections: Implications for Public Infrastructure Author(s): Douglas R. Dalenberg and Kevin T. Duffy-Deno Source: Public Choice , 1991, Vol. 70, No. 3 (1991), pp. 335-342 Published by: Springer Stable URL: https://www.jstor.org/stable/30025474 JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at https://about.jstor.org/terms Springer is collaborating with JSTOR to digitize, preserve and extend access to Public Choice This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 23:12:48 UTC All use subject to https://about.jstor.org/terms 151 Public Choice 70: 335-342, 1991. @ 1991 Kluwer Academic Publishers. Printed in the Netherlands. At-large versus ward elections: Implications for public infrastructure* DOUGLAS R. DALENBERG Department of Economics and Finance, John Carroll University, University Heights, OH 44118 KEVIN T. DUFFY-DENO Department of Economics, Weber State University, Ogden, UT 84408 Submitted 1 August 1989; accepted 26 March 1990 1. Introduction Since Gordon Tullock's (1959) seminal piece on logrolling, economists and po- litical scientists have attempted to find empirical support for the argument that vote trading leads to larger municipal expenditure levels. In order for logrolling to occur there must be clear potential gains from trade, negotiations must be feasible and less costly than the value of the trade, and enforcement must be possible (Buchanan and Tullock, 1962; Mueller, 1979). The city council en- vironment appears to satisfy these requirements. Apart from Zax (1985), how- ever, who examined municipal employment and compensation, no empirical evidence of the effect of logrolling on municipal expenditures levels has been uncovered. Different institutional structures may provide different incentives to trade votes. In particular, ward election systems may provide a greater incentive to logroll than at-large systems (Bradford, 1911; Goetz, 1977). Ward councilors rely on neighborhood loyalties for their re-election. Thus, they are interested in providing public services whose benefits are geographically concentrated but whose costs are spread city-wide. At large councilors, whose constituency is city-wide, are more concerned with city-wide benefits. This difference in loyal- ties suggests that ward councilors are more apt to engage in vote trading be- havior than are at-large councilors. One reason for the lack of empirical evidence of this behavior is that researchers have not examined the appropriate measure of municipal output. The local public good whose provision is most likely to be affected by logrol- * An earlier version of this paper was presented at the Public Choice Society Meetings in March 1989. The authors wish to acknowledge the valuable comments of Stephen Mehay and David Weimer. Partial funding was received from the Southeastern Massachusetts Research Committee. This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 23:12:48 UTC All use subject to https://about.jstor.org/terms 152 336 ling behavior is public infrastructure. The benefits of public capital projects (e.g., street and sidewalk repair, libraries, police and fire stations, sewer and water supply) are primarily concentrated in individual communities while the costs are borne by the entire city. Moreover, since capital projects are highly visible and durable, councilors can maintain a high profile in their community by working behind the scenes to procure such projects. Thus, as Goetz (1977) argues, evidence of logrolling under the ward election system may be found by examining a municipality's public investment decisions. Evidence of logrolling under the ward system can be detected by examining a cumulative measure of local decisions. Unfortunately both operating budgets and capital budgets fluctuate dramatically from year to year and are not strictly cumulative processes. However, public infrastructure which is the accumula- tion of year to year capital investments should capture the cumulative effect of such vote trading behavior. The stock of public infrastructure contains a record of political decisions and deals over the history of the city. The study provides such evidence by examining the effect of city council representation on municipal capital stocks. A cross-section of cities over time shows that cities with ward representation do have higher per capita stocks when other demand and supply factors are held constant. This increase in capi- tal stocks results from a capital bias rather than a proportional increase in both public capital and public labor. 2. Empirical model We are not aware of any previous study that has examined the effect of local institutions on public capital stocks.1 Thus, we hypothesize that per capita public capital (PCSTOCK) can be explained by time (TDUMs), the region of the country in which the city is located (RDUMs), various demographic and city characteristic variables, prices, and the form of city council representation. Demographic factors which may influence PCSTOCK include city density (DEN), whether or not city population is growing (POPCHG), per capita in- come (INCOME), median age (AGE), median education (EDUC), and percent of population that is white (WHITE). We also include the number of manufac- turing firms per person (MANUF) in the SMSA to account for the differing requirements and tastes of manufacturing sectors. SMSA population relative to city population (SPOP) is included in order to control for the importance of the city to the SMSA.2 The percentage of the city's housing stock built be- fore 1950 (HSG50) is used to proxy city age since older cities had a headstart in accumulating public capital. We also include intergovernmental aid per per- son (AID) and the number of municipalities per person (MUNI) within the SMSA (to capture local competition for residents and businesses). Price effects are captured through two variables. Cities often raise signifi- This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 23:12:48 UTC All use subject to https://about.jstor.org/terms 153 337 cant portions of their budgets through property taxes. As the percentage of residents who own their homes rises, more residents may perceive the actual tax burden, and demand lower tax rates and lower expenditures. We include the percentage of owner occupied housing (OWN) to account for this price ef- fect on capital outlays. We also include the ratio of the price of public capital to the price of public labor (INPUT P) to account for the cost of public capital to municipal governments.3 Finally, WARD is a dummy variable indicating whether or not city coun- cilors are elected by wards (= 1) or at-large (= 0). By estimating the model in log form the coefficients become elasticity estimates and heteroscedasticity problems are reduced (Theil, 1971). The resulting capital equation is: In PCSTOCK = a + aiE TDUMs + bjE RDUMs + ckE In Xk + di WARD + e, (1) where X is a vector of explanatory variables described above and e is the error term. If there is a greater degree of logrolling on capital projects in ward cities, then the cumulative effect of this process should cause di to be significantly positive. 3. Data Equation (1) is estimated using data for 30 cities for the years 1960 through 1981. A list of cities, regional definitions, and relevant descriptive data appears in Table 1. The sample of cities was selected randomly subject to two constraints. The sample was chosen such that there would be the same number of ward cities as at-large cities.4 The sample was also constrained to cities that remained large over time so that a public investment data series could be constructed back to 1910.5 This was necessary in order to apply the perpetual inventory technique to the public investment series to generate capital stock estimates for the years 1960 through 1981.6 Data for the public investment series are found in the City Finances series published by the U.S. Census Bureau. Public investment was constrained to functions common to all cities such as sewer construction, road construction, hospital construction, water supply work, etc. All nominal values are convert- ed into real values using 1972 as the base year.7 4. Empirical results The per capita public capital stock estimation results appear in column I of Ta- ble 2.8 The adjusted R-squared statistic of 0.74 indicates that the model fits This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 23:12:48 UTC All use subject to https://about.jstor.org/terms 154 338 Table 1. Cities, regions, council form, and public capital 1981 council Public capital City State form per resident New England Region (7 cities) Boston MA AL $4556 Bridgeport CT W' 2425 Hartford CT AL 4375 New Haven CT W 4519 Providence RI W 4415 Springfield MA AL 3608 Worcester MA AL 3528 Middle Atlantic Region (3 cities) Erie PA AL 2388 Pittsburgh PA AL 4452 Utica NY W 4217 Southern Region (4 cities) Baltimore MD W' 5880 Birmingham AL AL 2832 Louisville KY AL 4730 Oklahoma City OK W' 2909 Midwest 1 Region (8 cities) Chicago IL W 3506 Cleveland OH W 5709 Columbus OH AL 2352 Dayton OH AL 3594 Flint MI W 3752 Grand Rapids MI W' 2991 Toledo OH AL 2413 Youngstown OH W 3324 Midwest 2 Region (4 cities) Milwaukee WI W 3887 Minneapolis MN W 3499 St. Louis MO W' 4484 Wichita KS AL 4169 Western Region (4 cities) Los Angeles CA W 4113 Portland OR AL 2952 San Diego CA AL 1649 Spokane WA AL 1893 W = ward elections; AL = at-large elections; W' = one member elected at-large, treated as a ward election. Public capital per resident is 1981 real stock per resident measured in 1972 dollars. the data well. Moreover, most of the explanatory variables are statistically sig- nificant and have the expected signs. For example, cities that are relatively This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 23:12:48 UTC All use subject to https://about.jstor.org/terms 155 339 Table 2. Regression results In PCSTOCK In STOCK/EMPLOYMENT Variable Coefficient Std. error Coefficient Std. error CONSTANT 6.427*** 0.547 12.924*** 1.408 NE - 0.004 0.005 - 0.088*** 0.016 MIDATL 0.005 0.005 - 0.137*** 0.021 SOUTH 0.028*** 0.005 -0.007 0.016 MIDWEST 2 0.012*** 0.004 -0.020 0.014 WEST - 0.015*** 0.005 0.038** 0.016 20 TIME DUMMIES 20 TIME DUMMIES In DEN - 0.048*** 0.017 - 0.185*** 0.059 POPCHG - 0.015*** 0.003 - 0.003 0.011 In SPOP 0.013 0.015 -0.074 0.049 In INCOME 0.032 0.310 -0.029 0.099 In AGE 0.060 0.071 - 0.209 0.202 In EDUC -0.017 0.112 - 0.914*** 0.267 In WHITE - 0.123** 0.052 0.115 0.144 In MANUF 0.008 0.013 -0.112*** 0.031 In HSG50 0.492*** 0.035 -0.103 0.094 In AID 0.031*** 0.005 - 0.126*** 0.014 In MUNI 0.032*** 0.010 0.215*** 0.025 In OWN - 0.343*** 0.048 -0.491*** 0.102 In INPUT P -0.007 0.011 - 0.643*** 0.035 WARD 0.010*** 0.003 0.019** 0.008 Observations 630 630 F 46.49*** 33.35*** R-squared 0.75 0.69 Adjusted R-squared 0.74 0.67 Autocorrelation corrections were made for each city using the OLS residuals. Midwest 1 and 1961 are the omitted dummy variables. * indicates significant at the 10% significance level. ** indicates significant at the 5% significance level. *** indicates significant at the 1% significance level. older, receive greater intergovernmental aid, and face a higher degree of com- petition from neighboring cities are characterized by larger per capita public capital stocks. As hypothesized, cities which elect their council members through ward elec- tions rather than through at-large elections have significantly higher public capital stocks at the 1 percent significance level. Ward elections are associated with a 1.0 percent increase in per capita public infrastructure over time and across cities, ceteris paribus. The 1.0 percent increase in per capita stocks can be applied to those cities which elect their representatives at large to estimate the potential gain in capital This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 23:12:48 UTC All use subject to https://about.jstor.org/terms 156 340 Table 3. Municipal capital stock additions for at-large cities if they had used a ward system Additions to 1981 municipal capital measured in: City 1972 dollars 1987 dollars Birmingham $ 8,070,000 $20,346,000 Boston 24,832,000 62,606,000 Columbus 13,547,000 34,154,000 Dayton 7,404,000 18,667,000 Erie 2,818,000 7,105,000 Hartford 6,038,000 15,223,000 Louisville 13,717,000 34,583,000 Pittsburgh 18,609,000 46,917,000 Portland 10,626,000 26,790,000 San Diego 14,874,000 37,500,000 Spokane 3,218,000 8,113,000 Springfield 5,340,000 13,463,000 Toledo 8,541,000 21,543,000 Wichita 11,798,000 29,745,000 Worchester 5,681,000 14,323,000 Average 10,341,000 26,072,000 stocks that could have resulted if the city had originally instituted a ward sys- tem and had experienced the associated capital bias. These results are presented in Table 3 which shows increases in 1981 capital stocks ranging from $2.8 mil- lion in Erie to $24.8 million in Boston measured in 1972 dollars. In the case of Boston this inflates to $62.6 million dollars in 1987 dollars which is enough money to pay for a significant capital project.9 While the results of the estimation of equation (1) indicate that cities with ward election systems do have more public capital when other factors are held constant, the results do not distinguish between two possible cases. Cities with ward election systems could be spending proportionately more on both public capital and public labor, or they could be spending proportionately more on public capital, generating a capital bias. The visibility of many public capital projects may make them more attractive for logrolling than increases in public employment. In order to investigate this hypothesis we re-estimated equation (1) replacing PCSTOCK with the ratio of public capital stock to public employ- ment. The empirical results are shown in column 2 of Table 2. The positive and statistically significant coefficient on WARD indicates that such a capital bias does exist. Cities with ward election systems are associated with a 1.9 percent higher public capital - public labor ratio. This capital bias parallels the Averch-Johnson effect found in the regulation literature (Averch and Johnson, 1962) and is the direct result of the differing incentives under ward and at-large systems. This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 23:12:48 UTC All use subject to https://about.jstor.org/terms 157 341 5. Summary of findings and implications While city council members elected at large wish to form city-wide special in- terests groups to help keep them in office, councilors elected through ward elec- tions rely on neighborhood loyalties for their re-election. These differences lead to different incentives for councilors. In particular, ward councilors are biased toward large capital projects which are visible and have concentrated neighborhood benefits. A cumulative measure of public decisions, public capi- tal stock, captures this bias. In our sample of 30 large, U.S. cities during the 1960 to 1981 period, we find that cities which elect their councilors by wards are characterized by a 1 percent larger stock of public infrastructure. In terms of cities' 1981 capital stock, this translates into, on average, $26 million dollars (measured in 1987 dollars). Government officials, politicians, and economists are recognizing the vital role public capital plays in a local economy. Providing an adequate level of public capital is crucial to the economic development and health of cities. The findings of this study suggest that the structure of city government may play an important role in cities' attempts to deal with their deteriorating public in- frastructure. Notes 1. McEachern (1978) examined the effect of referendum voting rules on community debt patterns. Holcombe and Zardkoohi (1980) analyzed the impact of differences in state legislatures on capi- tal outlays. 2. While the cities of Cleveland and Columbus Ohio have similar population sizes, Cleveland serves as the center of a much larger SMSA and therefore needs more public capital. 3. The price of public capital is proxied by the rate of return on Moody's general obligation bonds for the relevant city divided by the expected inflation rate. The expected inflation rate is as- sumed to equal the previous year's inflation rate. 4. Five of the cities (Bridgeport, Baltimore, Oklahoma City, Grand Rapids and St. Louis) elect one council member at large and the remainder by wards. These cities are counted in the sample as ward cities, giving 15 at-large and 15 ward cities. 5. 48 percent of U.S. cities with a population greater than 100,000 residents used an at-large system, 16 percent used a ward system, and 36 percent used a combination system in 1980 (Sanders, 1983). 6. See Faucett (1977) and Usher (1980) for a discussion of the perpetual inventory technique. 7. A complete data appendix describing variable definitions and data sources can be obtained from the authors upon request. 8. Since we employ a pooled cross-sectional and time-series sample, heteroskedasticity and au- tocorrelation may be a problem. We avoid the problem of heteroskedasticity by analyzing per capita values and using a logarithmic form. In order to correct for autocorrelation, we estimated the equation for the full sample and used the residuals to estimate separate autocorrelation parameters for each city, via a Cochrane-Orcutt estimation procedure (Kmenta, 1986). We then used these parameters to quasi-difference the data, which required the 1960 observations be dropped. The transformed sample for 1961 through 1981 was then reestimated. 9. For example, this is more than the amount of money needed for the proposed Rock and Roll Hall of Fame to be built in Cleveland. This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 23:12:48 UTC All use subject to https://about.jstor.org/terms 158 342 References Averch, H. and Johnson, L. (1962). Behavior of the firm under regulatory constraint. American Economic Review 52 (December): 1052-1069. Bradford, E. (1911). Commission government in American cities. New York: Macmillan Press. Buchanan, J. and Tullock, G. (1962). The calculus of consent. Ann Arbor: University of Michigan Press. Faucett, J.(1977). Capital stock estimates for input-output industries: Methods and data. BLS Bulletin 2034 Washington, DC: U.S. Department of Labor. Goetz, C. (1977). Some influences of the form of political representation on majority rule fiscal choices. In P.B. Downing (Ed.), Local service pricing policies and their effect on urban spatial structure, 217-229. Victoria: University of British Columbia Press. Holcombe, R. and Zardkoohi, A. (1980). Public investment in a democracy. Southern Economic Journal 47 (July): 210-217. Kmenta, J. (1986). Elements of econometrics. New York: McMillan. McEachern, W. (1978). Collective decision rules and local debt choice: A test of the median-voter hypothesis. National Tax Journal 31 (June): 129-135. Mueller, D. (1979). Public choice. Cambridge: Cambridge University Press. Reid, G. (1987). Logrolling and the median voter: Pork barrel at the local level? Unpublished manuscript. Sanders, H. (1983). The government of American cities: Continuity and change in structure. In International City Management Association, Municipal yearbook 1983, 179-184. Washington, DC: ICMA. Theil, H. (1971). Principles of econometrics. New York: Wiley. Tullock, G. (1959). Problems of majority voting. Journal of Political Economy 67 (December): 571-579. Usher, D. (1980). Introduction. In D. Usher (Ed.), The measurement of capital, 1-21. Chicago: University of Chicago Press. Zax, J. (1985). Economic effects of municipal government institutions. National Bureau of Eco- nomic Research working paper No. 1657. This content downloaded from 153.90.18.165 on Thu, 12 Mar 2026 23:12:48 UTC All use subject to https://about.jstor.org/terms 159 1 INFORMATIONAL BRIEF Hybrid City Council Election Systems Francisco Jasso,PhD,Pat Brown Institute –California State University,Los Angeles Timothy Krebs,PhD,University of New Mexico Core Question This brief explores different types of election systems (district,at-large,mixed),the prevalence of hybrid/mixed city councils and their accompanying pros and cons,effects of multimember districts,and ends with an exploration of nested regional seats. Summary of Topline Findings The limited amount of research on hybrid city councils suggests a hybrid council made up of mostly district seats will provide as much representation as we may expect from a purely district-based system and may have the added benefit of providing greater gender diversity via its inclusion of at-large seats.Most current hybrid city councils are majority district-based arrangements and as such the force of arguments against purely at-large systems may be mitigated.With precedent in the California state constitution,the nesting of state Senate, Assembly,and Board of Equalization seats may provide a guide for the Los Angeles city council in designing a nested city council system. Background &Information Across the US,and particularly in California,many cities have transitioned away from at-large elections toward district-based elections on the basis of greater representation and influence for racial/ethnic groups and neighborhoods.California cities’transition from at-large to district-based elections began in the 1920s and dramatically accelerated after the implementation of the 2002 CA Voting Rights Act.To this day,many cities continue to transition and adopt district-based systems.While many cities retain at-large electoral systems,several of the largest cities and many smaller cities,are structured through a hybrid of at-large and district-based seats. Cities began to abandon at-large systems in favor of district-based or hybrid systems mainly since the 1965 Voting Rights Act (MacManus 1999).At the present,several American cities elect their city council through a mixture of at-large and district-based seats,see Table 1. 160 2 Table 1.US City Councils using Hybrid of District and At-Large Elections. Population (Approximate)City District Seats At-Large Seats 2.3 million Houston,TX 11 5 2 million Metro Council of Nashville & Davidson County 35 5 1.57 million Philadelphia,PA 10 7 955k Jacksonville,FL 14 5 907k Columbus,OH Currently 7 at-large seats,but will transition to 9 at-large general election seats to be first voted on by-district in a primary election* 880k Charlotte,NC 7 4 734k Seattle,WA 7 2 713k Washington,D.C.8 5 711k Denver,CO 11 2 655k Boston,MA 9 4 641k Portland,OR Currently 4 at-large seats,but in 2024 will expand to 3 ranked-choice seats per each of 4 districts 633k Detroit,MI 7 2 508k Kansas City,MO 6 6 496k Atlanta,GA 13 3 470k Raleigh,NC 5 2 434k Oakland,CA 7 1 300k Greensboro,NC 5 3 377k New Orleans,LA 5 2 285k Durham,NC 3 3 210k Rochester,NY 4 5 112k Peoria,IL 5 5 53k Wheaton,IL 4 2 Note:Table created by Francisco Jasso,PhD.Cities that describe their structure as hybrid in reference to having their mayor elected at-large are excluded. *This election system is also referred to as a “from district”system. Where are Hybrid Election Systems Situated in the Literature? There is very little literature on the prevalence or dynamics of hybrid city councils despite many cities across the nation using this hybrid structure.Research on this seems to be limited to early research from the 1970s-1990s on the emergence of hybrid city council systems post-1965 Voting Rights Act and greater representational outcomes relative to at-large election systems (see MacManus 1990).Research focusing on southern cities in the 1970s found cities adopting hybrid systems produced increases in representation and minority influence (MacManus 1999).In general,hybrid systems were found to yield greater Black &Latino representation than at-large systems (MacManus 1978),certainly if the Black population was smaller than 20%of the city population (Welch 1990). 161 3 Apart from a handful of school district studies (Leal,Martinez-Ebers,&Meier 2004;Meier & Rutherford 2014),most research thereafter focuses on the representational benefits from transitioning away from at-large election systems toward district-based systems.In general,the literature finds district-based elections to increase racial/ethnic representation,particularly if the group in question constitutes a large enough share of the city population (Collingwood &Long 2021;Marschall,Ruhil,&Shah 2010;Trounstine &Valdini 2008).Although,modified at-large elections have also been found to yield representational gains akin to district-based elections (Brockington,Donovan,Bowler,&Brischetto 1998). Analysis of new jurisdictions transitioning to district-based elections post-2002 California Voting Rights Act finds increases in Latino representation (Abott &Magazinnik 2020;Hertz 2023; Levitt &Johnson 2016).Nonetheless,while district-based elections increase racial/ethnic representation,at-large systems increase gender representation (Trounstine &Valdini 2008;See Huang,Murray,&Hofer (2018)for a review on the racial and gender trade-offs between at-large and district-based elections). The dominant critique in the literature against purely at-large systems is racial/ethnic vote dilution and underrepresentation (see Davidson &Korbel,1981),while the dominant critique against purely district-based elections is the fear and incentive of logrolling and parochialism – privileging their district to the detriment of the city’s general interests (Banfield &Wilson 1963). Burnett &Kogan (2014)challenge this critique leveraging roll call votes from the LA City Council and argue that council members rather practice a norm of conditional deference where they defer to council members on legislation affecting their districts but are very willing to oppose bills from individual members that negatively affect the whole city.This may be the only relevant study that focuses on the LA City Council. There may be other representational costs to the addition of at-large councilors.Again,with the caveat that most research compares pure district systems to pure at-large ones,at-large city councilors spend less time helping constituents,and are focused more on citywide issues and business concerns compared to district councilors (Welch and Bledsoe 1988;see also Polinard, Wrinkle,Longoria,and Binder 1994).Communities of color also may feel less efficacious and voter engagement may suffer as a consequence of at-large seats (Hajnal and Trounstine 2005; Lindgren 2007;Welch and Bledsoe 1988,110).At-large seats may also produce councilors of considerably higher socioeconomic status (Welch and Bledsoe 1988),and cause overall levels of campaign spending to increase (Lindgren 2007). Potential Questions That May Arise A proposal for a hybrid city council for LA can raise a few questions,namely: 1.Would a hybrid system stand Constitutional scrutiny? 2.How would a hybrid system affect current representation? 3.How would a hybrid city council change behavior between city council members? There is very little research on hybrid city councils to offer confident answers to these questions. However,a 2016 National League of Cities document on at-large,district-based,and hybrid 162 4 election systems suggests hybrid systems would be more likely to stand Constitutional scrutiny provided that most seats were district-based.The document does not provide justification for this expectation but given that the literature’s critiques of at-large systems refer to purely at-large systems,their application to majority district-based hybrid systems should be mitigated.A recent report on San Francisco’s governance structure additionally explored mixed systems concluding that “While this system has been less studied,the available evidence is that mixed systems provide governing boards with a more balanced policy perspective that accounts for both neighborhood and citywide needs.Mixed systems are not immune from suit under the CVRA, but are likely to be less vulnerable than purely at-large systems.”1 As Table 1 shows,most current hybrid councils are majority district-based arrangements and as such the force of arguments against purely at-large systems may be weaker. A bit of research on hybrid city councils suggests a hybrid council made up of mostly district seats will provide as much representation as we may expect from a purely district-based system and may have the added benefit of providing greater gender diversity via its inclusion of at-large seats.As such,LA communities worried that an expanding City Council would mean loss of their current representation may be more willing to support an expanded council with at-large seats they could additionally compete for.This of course calls for an LA-specific consideration given the uniqueness of LA and its small council relative to its population. At-large councilors will provide a citywide focus to policy problems,which may prove beneficial to local policymaking (Welch and Bledsoe 1988;Polinard,Wrinkle,Longoria,and Binder 1994).At-large seats would align the constituencies of the mayor and the at-large members of the city council,which may result in greater policy action and success.On the other hand,however,with their constituencies aligned,at-large members may confront powerful incentives to challenge an incumbent mayor for reelection thus reducing the incentive to cooperate on policy.Interviews with key informants in Philadelphia and Seattle suggest little for this proposition,however (see Krebs_mixed_ystem_interview summaries).Within the council, at-large seats reduce district-based turf battles,and engender less policymaking conflict (Welch and Bledsoe 1988).For less geographically concentrated communities of color,namely Latinos, at-large seats may produce opportunities to gain seats (Trounstine and Valdini 2008). On the 3rd question raised,it seems that the only article we can draw from is Burnett &Kogan (2014).We cannot say how council member behavior would change from the addition of at-large seats,but we can expect that a proposal for a hybrid LA City Council with majority district-based seats would not negatively affect the norm of conditional deference and general concern for the city’s interests that Burnett &Kogan find. Effects of Multi-Member State Legislative Districts At-large city councilors can be thought of as representing a multi-member district (MMDs),the boundaries of which are the city as whole.But at-large elections,whether they are part of a mixed electoral structure or a stand-alone system of representation,are not present just at the city level.In fact,the use of MMDs was quite common at the state level until the 1960s when court 1 See Heidorn,N.,Miller,K.P.,Nadon,B.,&by TogetherSF,C.(2023).Re-Assessing San Francisco’s Government Design. 163 5 decisions about their legality under the federal Voting Rights Act motivated state legislatures to adopt single-member districts (SMDs).Today,only nine states use multi-member districts to select the representatives of at least one of their legislative chambers,and only four –Arizona, New Jersey,South Dakota,and Washington—use this format to elect members of both chambers.2 The literature on state politics sheds light on a number of things relevant to the current proposal for a mixed system in LA.For example,scholars have examined whether the incumbent electoral advantage was greater in MMDs relative to SMDs.Cox and Morgenstern (1995)examined 16 years’worth of election data in 40 states and found that while the incumbent advantage grew under both systems,the rate of growth in MMDs was slower than that of SMDs,suggesting a greater incumbency advantage in SMDs.Hirano and Snyder (2009),however,showed that incumbent advantages were substantial in multi-member districts relative to SMDs and that the source of the advantage was officeholder benefits (e.g.,committee assignments,leadership positions)and incumbent quality.They also found that the effect of district office holder benefits is greater in more professionalized legislatures,which characterizes the LA city council. Research on the electoral effects of MMDs on the election of women and minorities to state legislatures suggests that women candidates are advantaged (Darcy,Welch,and Clark 1985; King 2002),while minority candidates,especially African American,candidates are not (Bullock and Gaddie 1993;Gerber,Morton,and Rietz 1998).The latter view may be changing,however. In a recent study of state legislative elections in Maryland,Herrnson,Rouse,and Taylor (2020) demonstrate that African American candidates may not be hindered by MMDs in patterns of candidate emergence,or primary and general election vote share.They attribute this break with conventional wisdom to several factors,including a reduction in racial gerrymandering and racial intimidation in more recent election cycles;deracialized campaign strategies on the part of African American candidates;“single shot”voting whereby African American voters select only African American candidates in multi-member contests;and greater willingness on the part of non-black voters to support black candidates.Not surprisingly,partisan polarization also plays a role in reducing the negative effect of candidates’race;since most African American candidates are Democrats,they benefit from increases in polarization among party voters. In addition to effects associated with descriptive representation,research on MMDs has focused on substantive representation.Larimer (2005)found that higher concentrations of MMDs in upper state legislative chambers negatively affected the generosity of welfare programs,while the concentration of MMDs in lower chambers had no effect on generosity,controlling for minority and female descriptive representation in state legislatures.More generally,evidence suggests that representation in MMDs is more ideologically diverse than in SMDs (Adams 1996; Bertelli and Richardson 2008;Hale 2019). Another topic is collaboration,or the willingness of representatives from MMDs to cooperate on shared policy and district objectives.According to research by Snyder and Ueda (2007)counties receive more in state intergovernmental transfers when represented by people elected in MMDs compared to SMDs.This is so because members cooperate to deliver district improvements,thus 2 See https://ballotpedia.org/State_legislative_chambers_that_use_multi-member_districts#cite_note-26,accessed October 22,2023. 164 6 solving the collective action problems expected in these representation schemes.Research by Kirkland (2012)also shows that legislators representing MMDs collaborate effectively. Scholars also have examined the effects MMDs on civic engagement.According to Herrnson, Taylor,and Curry (2015),MMDs may experience higher levels of voter roll-off or the tendency of voters to submit incomplete ballots.Candidates’campaign donor bases are also more diverse in MMDs than SMD elections (Curry,Herrnson,and Taylor 2013). Nested Regional Seats -An L.A.Solution to an L.A.Problem? One difference between district and at-large seats is the geographic constituency that casts votes. At-large seats encompass the entire city jurisdiction while districts encompass a very limited geographic area.We can imagine however the possibility of seats that compromise between the geographic scope of district and at-large seats:regional seats.Regional seats would be larger than individual districts,yet smaller than the city-wide jurisdiction.It is important to note that regional seats are best understood in the presence of districts.In the absence of districts,the concept of regional seats is indistinguishable from districts as they serve the same purpose in relation to at-large seats. There does not appear to be any research on regional seats as the bulk of the literature regards at-large systems,district systems,and the transition from the former to the latter.An additional, large area of research that is not explored in this brief regards alternative voting formulas such as ranked choice or proportional ranked choice formulas.Table 1 above shows two alternatives to the plurality-winner,single member districts system where at-large elections occur but initiate from districts (Columbus,OH)or have multi member ranked choice district elections (Portland, OR).San Francisco,Berkeley,Oakland,and San Leandro use a single-member ranked choice voting system and Albany,CA uses a multi-member ranked choice version.3 Fifty-one US jurisdictions including 2 states and 3 counties use ranked choice voting.4 A proposal for Los Angeles to have a mixed system of districts and regional (instead of at-large) seats would be a unique configuration not found in other cities of comparable size and likely not practiced in any smaller city either.This would amount to an LA solution to an LA problem: regional seats superimposed over districts.In the case of LA,regional seats may aim to fulfill the purpose of at-large seats in terms of policy focus but would ensure diversity of geographic representation by preventing geographic areas of the city from having a near monopoly on these seats. The California State Constitution,however,does provide precedent for the use of nesting.5 Proposition 11,voter-approved in 2008,added language to the California Constitution in Article 21 ranking state redistricting criteria in order of priority.After the redistricting requirements of 5 Other than Article 21 of the California State Constitution,Special Masters appointed by the state Supreme Court in the 1970s and 1990s also nested their districts (see Kogan &McGhee,2012). 4 See FairVote,“Where is Ranked Choice Voting Used?” https://fairvote.org/our-reforms/ranked-choice-voting-information/,accessed November 22,2023. 3 See Heidorn,N.,Miller,K.P.,Nadon,B.,&by TogetherSF,C.(2023).Re-Assessing San Francisco’s Government Design. 165 7 equal population,compliance with the VRA,district compactness,etc.,it includes the nesting of state Assembly,Senate,and Board of Equalization districts: “To the extent practicable,and where this does not conflict with the criteria above,each Senate district shall be comprised of two whole,complete,and adjacent Assembly districts,and each Board of Equalization district shall be comprised of 10 whole, complete,and adjacent Senate districts.”6 For the state 2010 redistricting cycle,perfect nesting was not required nor achieved given compliance with higher ranked criteria.Tension can exist between the different ranked criteria and therefore the pursuit of one higher ranked criterion will limit the extent to which a lower ranked criterion is met,hence “to the extent practicable.”7 In the case of Los Angeles,there may be a lower chance of tension given the smaller geography of Los Angeles relative to the entire state. Nonetheless,tension among the criteria was less pronounced than otherwise expected in the state 2010 cycle.8 The maps drawn in the 2010 redistricting cycle achieved 3 nested Senate districts in the high 60s percent range,15 in the 70 percent range,10 in the 80 percent range,9 in the 90 percent range,and 3 100%nested seats.9 The maps drawn in the 2020 cycle achieved 1 district nested at 52.3%,3 in the high-60s percent range,16 in the 70s range,7 in the 80s range,12 in the 90s range,and 1 100%nested district.10 If modeled on the practice of California state redistricting,the California Constitution establishes the legal framework for drawing districts in a nested manner. 10 See California Citizens Redistricting Commission.(2021).Report on Final Maps,page 49. 9 See State of California Citizens Redistricting Commission Final Report on 2011 Redistricting (2011): https://wedrawthelines.ca.gov/wp-content/uploads/sites/64/2011/08/crc_20110815_2final_report.pdf,page 25. 8 See Kogan &McGhee (2012),pages 14-16. 7 See Barabas &Jerit (2004)and Cain &MacDonald (2007). 6 https://law.justia.com/constitution/california/article-xxi/section-2/ 166 8 References Abott,C.,&Magazinnik,A.(2020).At‐Large Elections and Minority Representation in Local Government. American Journal of Political Science, 64(3),717-733. Adams,Greg D.1996.“Legislative Effects of Single-Member vs.Multi-Member Districts.” American Journal of Political Science 40(1),129–44. Banfield,E.C.,&Wilson,J.Q.(1963).City Politics.Cambridge:Harvard Univ.Press. Barabas,J.,&Jerit,J.(2004).Redistricting principles and racial representation.State Politics & Policy Quarterly,4(4),415-435. Bertelli,Anthony,and Lilliard E.Richardson Jr.2008.“Ideological Extremism,Branding,and Electoral Systems:Multimember versus Single Member Systems.”Public Choice 137 (1–2),347–68. Brockington,D.,Donovan,T.,Bowler,S.,&Brischetto,R.(1998).Minority representation under cumulative and limited voting. The Journal of politics, 60(4),1108-1125. Bullock,Charles S.III,and Ronald Keith Gaddie.1993.‘‘Changing from Multimember to Single-Member Districts:Partisan,Racial and Gender Consequences.’’State and Local Government Review 25:155–163. Burnett,C.M.,&Kogan,V.(2014).Local logrolling?Assessing the impact of legislative districting in Los Angeles. Urban Affairs Review, 50(5),648-671. Cain,B.E.,Mac Donald,K.,Boyle,N.,Palazzo,W.,Russian,N.,&Bonovich,N.(2007).The Implications of Nesting in California Redistricting.Institute of Governmental Studies, University of California at Berkeley. California Citizens Redistricting Commission.(2021).Report on Final Maps. https://wedrawthelines.ca.gov/wp-content/uploads/sites/64/2023/01/Final-Maps-Report-w ith-Appendices-12.26.21-230-PM-1.pdf Collingwood,L.,&Long,S.(2021).Can states promote minority representation?Assessing the effects of the California Voting Rights Act. Urban Affairs Review, 57(3),731-762. Cox,Gary W.and Scott Morgenstern.1995.“The Incumbency Advantage in Multimember Districts:Evidence from the U.S.States.”Legislative Studies Quarterly 20(3),329-349. Curry,James M.,Paul S.Herrnson and Jeffrey A.Taylor.2013.“The Impact of District Magnitude on Campaign Fundraising.”Legislative Studies Quarterly 38:517-543. 167 9 Darcy,R.,Susan Welch,and Janet Clark.1985.“Women Candidates in Single-and Multi-Member Districts:American State Legislative Races.”Social Science Quarterly 66 (4):945–53. Davidson,C.,&Korbel,G.(1981).At-large elections and minority-group representation:A re-examination of historical and contemporary evidence.The Journal of Politics,43(4), 982-1005. FairVote.(Accessed Nov.22,2023).“Where is Ranked Choice Voting Used?” https://fairvote.org/our-reforms/ranked-choice-voting-information/ Gerber,Elizabeth R.,Rebecca Morton,and Thomas A.Rietz.1998.‘‘Minority Representation in Multimember Districts.’’American Political Science Review 92:127–144. Hajnal,Zoltan,and Jessica Trounstine.2005.Where Turnout Matters:The Consequences of Uneven Turnout in City Politics.The Journal of Politics,67(2):515-35. Hale,Isaac.2020.“Candidate Extremism and Electoral Design in U.S.State Legislative Elections.”Social Science Quarterly,101(2),861-877. Heidorn,N.,Miller,K.P.,Nadon,B.,&by TogetherSF,C.(2023).Re-Assessing San Francisco’s Government Design. Herrnson,Paul S.,Jeffrey A.Taylor,and James M.Curry.(2015).‘‘The Impact of District Magnitude on Voter Drop-Off and Roll-Off in American Elections.’’Legislative Studies Quarterly 40:627–650. Herrnson,Paul S.,Stella M.Rouse,and Jeffrey A.Taylor.(2020).“The Impact of Electoral Arrangements on Minority Representation:District Magnitude and the Election of African American State Legislators.”Election Law Journal 19(1),64-77. Hertz,L.Z.(2023).Does a Switch to By-District Elections Reduce Racial Turnout Disparities in Local Elections?The Impact of the California Voting Rights Act.Accepted at Election Law Journal. Hirano,S.,&Snyder,Jr,J.M.(2009).Using multimember district elections to estimate the sources of the incumbency advantage.American Journal of Political Science,53(2), 292-306. Huang,C.,Murray,R.,&Hofer,S.(2018).The Trade-Offs between At-Large and Single-Member Districts. 168 10 King,James D.(2002).“Single-Member Districts and the Representation of Women in American State Legislatures:The Effects of Electoral System Change.”State Politics and Policy Quarterly 2(2):161–75. Kirkland,Justin H.2012.“Multimember Districts’Effect on Collaboration between U.S.State Legislators.”Legislative Studies Quarterly 37(3),329-353 Kogan,V.,&McGhee,E.(2012).Redistricting California:An evaluation of the citizens commission final plans.California Journal of Politics and Policy,4(1). Larimer,Christopher W.2005.“The Impact of Multimember State Legislative Districts on Welfare Policy.”State Politics &Policy Quarterly 5(3):265-282. Leal,D.L.,Martinez-Ebers,V.,&Meier,K.J.(2004).The politics of Latino education:The biases of at-large elections. The Journal of Politics, 66(4),1224-1244. Levitt,J.,&Johnson,D.(2016,November 3).Quiet revolution in California local government gains momentum. Rose institute of State and Local Government. https://roseinstitute.org/wp-content/uploads/2016/11/CVRA-White-Paper.pdf Lindgren,Eric A.2007.Comparing San Francisco’s At-Large and District Supervisor Elections’ Average Spending and Participation Rates.California Politics &Policy,June:38-46. MacManus,S.A.(1978).City council election procedures and minority representation:are they related? Social Science Quarterly, 59(1),153-161. MacManus,S.A.(1999).Mixed Election Systems:The Newest Reform Structure. Local Government Election Practices. Marschall,M.J.,Ruhil,A.V.,&Shah,P.R.(2010).The new racial calculus:Electoral institutions and black representation in local legislatures. American Journal of Political Science, 54(1),107-124. Meier,K.J.,&Rutherford,A.(2014).Partisanship,structure,and representation:The puzzle of African American education politics. American Political Science Review, 108(2), 265-280. National League of Cities.(2016,December 2014).Cities 101 –At-Large and District Elections. https://www.cityofws.org/DocumentCenter/View/13991/Cities-101---District-v-At-Large- Elections---National-League-of-Cities-pdf Polinard,J.L.,Robert D.Wrinkle,Tomas Longoria,and Norman E.Binder 1994.Electoral Structure and Urban Policy:The Impact on Mexican-American Communities.New York: Routledge. 169 11 Snyder,Jr.James M.and Michiko Ueda.2007.“Do Multimember Districts Lead to Free-Riding?”Legislative Studies Quarterly 32(4):649-679. State of California Citizens Redistricting Commission Final Report on 2011 Redistricting (2011). https://wedrawthelines.ca.gov/wp-content/uploads/sites/64/2011/08/crc_20110815_2final _report.pdf Trounstine,J.,&Valdini,M.E.(2008).The context matters:The effects of single‐member versus at‐large districts on city council diversity. American Journal of Political Science, 52(3),554-569. Welch,S.(1990).The impact of at-large elections on the representation of Blacks and Hispanics. The Journal of Politics, 52(4),1050-1076. Welch,Susan,and Timothy Bledsoe.1988.Urban Reform and Its Consequences:A Study in Representation.Chicago:University of Chicago Press. 170 Memorandum REPORT TO:Study Commission FROM:Caeleb Heinen, Recording Secretary Mike Maas, Ex Officio SUBJECT:Future Education Discussion and Needs for April 2 and April 15 Meetings MEETING DATE:March 25, 2026 AGENDA ITEM TYPE:Citizen Advisory Board/Commission RECOMMENDATION:Education process and needs for April 2 and April 15 for the following topics around community representation and involvement: wards/districts neighborhood associations advisory boards STRATEGIC PLAN:1.1 Outreach: Continue to strengthen and innovate in how we deliver information to the community and our partners. BACKGROUND:This is an opportunity for Study Commissioners to discuss what further materials or education they will need on the aforementioned topic. UNRESOLVED ISSUES:None identified ALTERNATIVES:As per the Study Commission FISCAL EFFECTS:None identified Report compiled on: March 12, 2026 171