The City That Won’t Agree

Ann Arbor—population 125,700, home to the University of Michigan, the kind of place where yard signs outnumber parking spots—records an 18.19% dissent rate across 7,100+ agenda items, which is triple the next-closest city in our dataset and roughly three hundred times Chicago’s. That rate has held above 12% every single year we measured, through recessions, a pandemic, and complete turnovers of the council itself.

To understand how strange this is, consider the politics. Ann Arbor has not elected a Republican to any office since 2003. All eleven council seats belong to Democrats. The city is one of only two in Michigan—along with Ypsilanti—that still holds partisan municipal elections, which means the August Democratic primary effectively decides every race on a turnout of 25–35%. And yet this single-party city produces more internal disagreement than councils in places that are actually split between the two national parties.

The fights are entirely intramural: urbanists vs. preservationists, climate hawks vs. neighborhood advocates, the mayor’s organized coalition vs. a rotating cast of independents whose only common ground is skepticism of the mayor’s organized coalition. Locals call it “The People’s Republic,” a nickname that dates to the 1960s, when Students for a Democratic Society was founded on campus and the University hosted the nation’s first Vietnam War teach-in. The Human Rights Party won council seats in the 1970s. This city has been arguing with itself for sixty years, and the data says it does so at an intensity no other American city comes close to matching.

Taylor was describing the 2018–2020 council. Seven members the local analyst Sam Firke labeled “Protectors” held the majority. They wanted roads fixed, sewers maintained, police staffed, and neighborhoods left alone. The mayor’s four-member minority, the “Strivers,” pushed for density, transit, and climate spending. Debates ran past 3 a.m. The Protector majority fired City Administrator Howard Lazarus in February 2020 without cause—a 7–4 vote that provoked a public backlash and a $223,600 severance payout.

Then the map flipped. In August 2020, Taylor’s endorsed slate swept all five contested ward seats. Lisa Disch, Linh Song, Travis Radina, Jen Eyer, and Erica Briggs replaced Protector-aligned members or claimed open seats. By 2022, the progressive faction held the entire council. The contention barely paused. Dissent ran 20.3% in 2019, 19.5% in 2020, 22.0% in 2021, 19.8% in 2022. Only in 2023, with the last Protector-aligned members gone, did it dip to 12.3%—though a smaller sample of votes that year clouds the signal.

So was the fighting structural—hard-wired into a city where everyone reads the staff reports and nobody defers to anyone? Or was it factional, driven by a cohort of dissenters who eventually aged out? The seventeen-year baseline above 12% points to culture. The 2023 dip points to personnel. Ann Arbor probably breeds dissent by temperament, and certain councils crank the volume up.

Jeff Hayner represented Ward 1 from 2018 to 2022. He voted no on 46.9% of everything that came before him—nearly half. He and Lisa Disch, who holds that same Ward 1 seat today, agreed on 32.4% of the 293 votes they shared. On a typical American city council, any two members agree above 95%. Hayner and Disch, both Democrats, occupying the same seat four years apart, were worse than a coin flip.

Hayner was the council’s most visible contrarian: an environmental advocate who championed the Gelman Plume cleanup—a 1,4-dioxane contamination that has been spreading beneath Ann Arbor since the 1960s—and who led the decriminalization of entheogenic plants. He sponsored PILOT resolutions targeting the University of Michigan’s enormous tax-exempt footprint. In April 2021, he posted a passage from Hunter S. Thompson’s “Fear and Loathing in Las Vegas” on Facebook containing a homophobic slur. Hayner argued it was literary commentary, not endorsement. Council voted 8–2 to strip him of all committee appointments. A recall effort won approval from the Washtenaw County Election Commission but fell short of the 2,264 signatures required.

Julie Grand, who has held Ward 3 since 2014, sits at the other pole. She and Linh Song vote together 97.2% of the time—a locked bloc. She and Hayner agreed just 41% across 719 shared votes. She and Anne Bannister, 52%. The pattern repeats across every Protector–Striver pairing in the dataset. The heatmap below makes the divide physical: a warm cluster of seven members in the upper-left, a tight progressive bloc of four in the lower-right, and a cold diagonal fault line between them.

Not everyone accepts the two-faction story. Dave Askins, former editor of the Ann Arbor Chronicle, applied multidimensional scaling to 100 roll-call votes in 2019 and found only 14 that split strictly along factional lines. The Protectors showed far more internal variation than the Strivers, who voted as a near-perfect bloc. Elizabeth Nelson, a councilmember then loosely grouped with the Protectors, wrote that there was really “only one faction”—the mayor’s organized coalition—while the supposed opposition was seven independent voices who happened to converge occasionally.

The voting data gives both camps ammunition. Among the Protectors, agreement rates ranged widely, from 62.8% between Hayner and Lumm to 78.5% between Eaton and Bannister—a loose alliance that held together on big votes but fractured on the details. Among the Strivers, agreement ran above 82% across every pairing: Taylor, Grand, Disch, and Song moved as a unit. That asymmetry goes a long way toward explaining why the Strivers eventually won. They didn’t need to agree on everything—just enough to show up with the same vote when it counted.

On the evening of November 9, 2014, Ann Arbor police officers responded to a domestic disturbance on West Summit Street. They shot and killed Aura Rosser, a 40-year-old Black woman, inside her own home. The county prosecutor declined to press charges. The killing became a wound that did not close, and when George Floyd’s murder nationalized the policing debate six years later, Ann Arbor did not need to be convinced there was a problem. It had been arguing about the problem since 2014.

Council earmarked $3.5 million in American Rescue Plan Act funds for an unarmed crisis response pilot: a 16-person team with its own emergency number, its own dispatchers, and de-escalation specialists drawn from people with direct experience of incarceration or mental health crisis. Donnell Wyche, senior pastor of Vineyard Church, led the Coalition for Re-envisioning Our Safety in developing the model. Estimated cost: $3 million a year for a two-year pilot.

City Administrator Milton Dohoney found the only submitted bid non-responsive to the RFP requirements, citing the cost of building separate dispatch infrastructure. Whether that was a legitimate procurement finding or a bureaucratic kill shot depends on who you ask. In July 2024, facing the federal ARPA spending deadline, council reallocated the $3.5 million to a Barton Dam embankment project and park improvements. The program born from Aura Rosser’s killing never launched.

The ARPA fight was one front. In 2017, the city restructured water rates, shifting costs from multifamily apartments to single-family homeowners. Jane Lumm—the council’s lone independent, a former Republican in a city that treats the label like a communicable disease—became the champion of furious ratepayers. A lawsuit, Hahn v. City of Ann Arbor, alleged the city had overcharged “in the tens of millions” by exceeding the cost of service. A judge denied class certification. Council split 6–5 to delay rate increases during COVID, then pushed through a combined 14% increase over six months in 2021.

In November 2024, Ann Arbor voters considered two ballot proposals that might have defused some of the fighting: nonpartisan elections and the state’s first local campaign finance reform. Mayor Taylor opposed both, calling himself a “proud Democrat.” Critics of the measures argued they were designed to hand a back door to candidates who had lost in Democratic primaries. Both proposals failed. Ann Arbor voted, in effect, to keep the system that produces its contention.

30 items in the Ann Arbor dataset passed or failed by a single vote—6–5 on an eleven-member council, or 5–5 ties that killed the motion outright. These were not procedural throwaways. They include annual budgets, neighborhood rezonings, the Gelman contamination settlement, and whether to end the city’s COVID local emergency—each one decided by a single person’s presence or willingness to break from their usual allies.

The penultimate entry is the one that makes Ann Arbor, Ann Arbor. In September 2024, council split 5–5 on a leaf blower ban. The motion failed. A city that divides evenly on leaf blowers is a city that will divide on anything.

If you are a cynic about local government, the 93.83% unanimity rate confirms everything you already believe: the meetings are a rubber stamp, public testimony is theater, and the real decisions get made long before anyone sits down at the dais. Ann Arbor is the counterargument. Its council members actually read the 400-page budget packet, show up with amendments, vote no when they disagree, and occasionally lose a 5–5 fight over leaf blowers at 11 p.m. on a Monday night.

The cost of that kind of deliberation is not abstract. Meetings that run past 3 a.m., a fired city administrator and his $223,600 severance payout, and a $3.5 million crisis response program that consumed years of political energy before being reallocated to a dam embankment—those are real invoices, paid by real taxpayers, for a council that refuses to let the professionals run things quietly.

But if you believe the meetings should be what they claim to be—open deliberation by elected representatives who disagree honestly and settle those disagreements where the public can watch them—then Ann Arbor starts to look less like an outlier and more like the only city in this dataset that is actually doing the job. Across 142 cities and 8.1 million votes, it stands alone, which raises an uncomfortable possibility: maybe the problem isn’t Ann Arbor.