Cities, Culture Culture & Society
May 18th, 2021 11 Minute Read Issue Brief by Daniel J. Hopkins

Declining Turnout in Big-City Elections: A Growing Problem for Democratic Accountability


The killing of George Floyd in the custody of Minneapolis police officers, along with the disturbing video of the incident, set off a wave of protests across the United States. On June 6, 2020, for example, an estimated 50,000–80,000 people demonstrated in Philadelphia.[1] They sought reforms to policing, and one of their targets was the administration of the city’s mayor, James Kenney.

Such large crowds and public interest in municipal policy were striking, given that seven months earlier, in November 2019, Kenney had been reelected in a contest for which only 27% (286,904) of the city’s registered voters turned out.[2] It’s not that voter turnout in Philadelphia is always so low: some 749,000 Philadelphia residents cast ballots in the November 2020 presidential election, more than 2.5 times the number who had voted in the previous year’s municipal election.[3] Despite the mayor’s authority over the police department and despite what had already been a growing and deep public passion about policing and criminal justice, Philadelphia residents were far more engaged with the 2020 presidential race than with the 2019 mayoral race.

Philadelphia’s experience is by no means unique. In Chicago, the total turnout in the 2020 presidential election was 2.2 times higher than in its 2019 mayoral election. In Kansas City, Missouri, the same ratio was 2.0. In Boston, the ratio between turnout in the 2016 presidential election and the 2017 mayoral election was 2.5. Researchers and commentators have written extensively about declining turnout in presidential elections, but turnout in big-city municipal elections is substantially lower and more troubling.[4]

Such gaps—and such low levels of public engagement in municipal elections—are especially striking, given that America’s federalist institutions give substantial authority to states and localities. Of every tax dollar in the U.S., states and localities spend 48 cents.[5] Consider Philadelphia: its 2020 budget was $4.8 billion, meaning that it spent over $3,000 per resident. Its policy responsibilities include not only policing and criminal justice but also K–12 education, land use, development, sanitation, and infrastructure—many of the keys to building a safe, prosperous, equitable city. Yet 2019 wasn’t an anomaly—turnout in Philadelphia’s recent mayoral elections was even lower in 2011 and 2015.

Turnout gaps are also important because of what they mean for the inequality of political representation. As research by J. Eric Oliver, Sarah Anzia, William A. Fischel, and Bernard Fraga has demonstrated, low-turnout elections are likely to favor certain groups, including organized interests, homeowners, higher-income voters, and white voters.[6] Such turnout differentials, refracted through what are often highly segregated urban neighborhoods, are readily apparent to politicians and can have a profound influence on policymakers’ priorities.[7] When cities face fiscal distress, as has been the case for many localities during the pandemic, city officials are unlikely to prioritize communities with low levels of turnout. Turnout differentials may also contribute to policies that limit urban development, pushing more development to lower-density areas associated with higher carbon dioxide emissions.[8]

Voter turnout in America’s big cities hasn’t always been this low. In 1989, 1.9 million New Yorkers voted for mayor in the election that pitted David Dinkins against Rudolph Giuliani—and made Dinkins the city’s first black mayor. That represented 89% of the city’s turnout in the presidential election in the previous year. Six years earlier, in 1983, the 1.3 million voters who participated in Chicago’s mayoral election actually outpaced the number of voters who cast ballots in the subsequent year’s presidential race. Also in 1983, more than 700,000 Philadelphia residents voted in the mayoral race, which was 93% of the ballots cast in the following year’s presidential election.

But there’s an obvious explanation for those turnout surges: each of those elections pitted a black candidate against a white candidate and represented the first election of a black mayor in the city’s history. Whether it was Harold Washington in Chicago or Wilson Goode in Philadelphia, such elections were typically highly racialized, divisive contests.[9] The question is how to increase turnout in big-city elections without fostering racial divisions.


Why Big-City Turnout Is Down

Several explanations for the recent declines in big-city turnout do not emphasize race. Another feature of those 1980s’ big-city mayoral races is that they were highly competitive. In New York, Dinkins won with 51% of the vote, and in Chicago, Washington’s victory was with 52%. But as political polarization has increased nationally, fewer voters are willing to break with their political party, even in down-ballot races, with the consequence that cities with partisan primaries often have lopsided, uncompetitive general elections.[10] In New York’s 2017 mayoral election, Bill DeBlasio won 66% of the vote despite polls showing that only 50% of voters gave his performance an “A” or a “B.”[11] And in Philadelphia’s 2019 mayoral election, the incumbent mayor won 80% of general-election votes. Partisan primaries, coupled with the concentration of Democratic voters in the nation’s big cities, have led many general elections to be uncompetitive and have thus contributed to declining turnout levels.[12]

Yet partisan primaries are far from the only explanation. Chicago and Los Angeles are among the big cities that use nonpartisan mayoral elections, yet Chicago’s 2019 mayoral turnout of about 557,000 was less than half its turnout in 2020’s presidential election. Similarly, the 2013 Los Angeles mayoral turnout was just over one-third of its presidential turnout the previous year.

It borders on tautology to say that turnout is lower because voter interest is lower; that statement simply leads us to ask, why is voter interest lower? Low voter engagement with city politics is something of a paradox since cities control a variety of policy levers that can have major impacts on citizens’ day-to-day lives—and also since cities’ smaller sizes relative to states means that blocs of voters are more likely to influence election outcomes. But one key part of the explanation is the transforming media environment.

Markets for local television stations and print newspapers are typically centered in larger cities. Since these media outlets define their audience geographically, they have strong incentives to cover news of particular interest to people living in and around those cities. That includes not only local sports and weather but also politics. In surveys I’ve conducted, respondents are much more likely to report getting local news from print newspapers and local television.[13]

However, these two staples of information about local government are rapidly losing their audiences. From 1990 to 2012, according to Pew Research Center surveys, the fraction of Americans who regularly read print newspapers or watch local television dropped markedly and has continued to decline. This has led to a hollowing out of local news coverage—some outlets close, and others cut back on their newsrooms.[14] Other local television stations have been replaced by national, partisan networks such as Sinclair.[15] In a panel survey, I tracked self-reported media usage by the same respondents and found a four-percentage-point decline in newspaper usage between 2008 and 2012 alone.[16] When I speak to local audiences about these issues, I joke that as someone in my early forties, I am the youngest print subscriber to the Philadelphia Inquirer I know—and no one has ever contradicted me.

Meanwhile, cable television and online news have captured increasing attention—and the news coverage through both types of media is highly national in orientation.[17] Fox News and MSNBC have incentives to build national audiences by focusing on national issues; they cover state and local politics only when it fits with national stories and frames. Even the Internet, which in theory has the potential to lower production costs for local news, has instead centralized political attention while siphoning the advertising revenues that local outlets relied on.[18] As voters get more of their news from cable television and online, they are getting their information from outlets that have a national or even an international focus. Accordingly, it is no surprise that Americans’ knowledge about and engagement with local politics are declining.

Such trends can have clear impacts on voter choice as well as voter turnout. For example, incumbent mayors perform worse when running for reelection if local unemployment is rising.[19] However, that relationship only holds true in cities that dominate their media market—and thus where local unemployment gets sufficient news coverage. For democratic accountability to work in big cities, local media is a necessary ingredient.

A Menu of Reforms

Declining engagement with big-city politics has several causes, and policymakers should consider a wide range of reforms, including:

  • Holding elections concurrently with state and federal elections, which has been shown to increase municipal turnout; and
  • Ending partisan primaries, which could make city elections more competitive and decouple national partisan loyalties from local voting decisions.[20]

With respect to holding municipal elections concurrently with other elections, the benefits are sizable and straightforward. American democracy stands out for the number of different elections that we ask voters to participate in—and the frequency of American elections can dampen and distort voter turnout, advantaging organized interests. But cities can increase turnout, often dramatically, by holding their mayoral elections alongside statewide or federal elections.[21] Baltimore is an example: turnout for its 2011 mayoral election was 13% of registered voters. But turnout grew to 60% in 2016, when the city held its mayoral election at the same time as the presidential election.[22] Similarly, Austin, Texas, saw big turnout gains in 2014 when it made its city elections contemporaneous with statewide ballots.[23] A recent study found that voter turnout grew by almost three times in 54 California cities that moved to on-cycle elections in the wake of a 2015 state law.[24]

Due to polarization and nationalization, many large cities’ electorates are now overwhelmingly Democratic in national politics. When these cities use party primaries to nominate candidates who then become heavy favorites in general elections,[25] they can reduce general-election turnout—and can also reduce the political voice of those not registered with the dominant local party, usually the Democrats. In Philadelphia’s 2019 mayoral election, for example, just shy of 250,000 registered voters were outside the Democratic Party and thus unable to cast ballots in its primary, despite the fact that the Democratic nominee is typically the overwhelming favorite to win the general election.[26] In this way, voters’ national loyalties can constrain their ability to cast meaningful local ballots.[27] Here, a promising reform is to open primaries to all candidates, with the top two vote-getters going on to contest the general election. Christopher Warshaw notes in a report on possible reforms to Baltimore’s elections: “Although a top-two primary would probably result in two Democratic mayoral candidates advancing to the general election, both candidates would be incentivized to compete for the support of the largest number of voters in the general election, making the winner more representative of Baltimore voters’ preferences.”[28]

More broadly, researchers and policymakers would do well to consider the ways in which contemporary federalism—with its complex, overlapping jurisdictions—is jeopardizing political accountability in local and state elections. In recent decades, policy areas that have traditionally fallen to state and local governments have increasingly experienced federal intervention.

Such trends are evident in policy arenas from health care and transportation to education and criminal justice. While any single such policy intervention may be impeccably justified on policy grounds, the collective impact is likely to weaken the connection between mayors and other local officials and the policy areas within their influence. To some extent, nationalized voting patterns are an understandable response to the blurring of lines separating federal, state, and local authority.[29]

Problems of this magnitude cannot be solved by any single reform, and certainly not by nudge-style interventions alone. Nonetheless, it is noteworthy that in research I have conducted with the City of Philadelphia and the Philadelphia City Commissioners, even the receipt of postcards communicating about voting and elections has had meaningful effects on voter turnout.[30] For example, in one trial in 2019, we sent randomly selected Philadelphia registrants to receive zero, two, or four postcards, and increased turnout by 1.4 percentage points in the 2019 general election for those who received four postcards.

In part, we hypothesize that the interactive nature of the postcards may have helped foster a sense of reciprocity with the recipients. Government officials and accessible, voter-friendly election administration can have meaningful impacts on voter turnout, too. Nonetheless, these are small changes in political systems that could frequently benefit from more fundamental changes.

Still, it is critical to consider the larger media environment. Here, federal investments and regulations that encourage high-quality, nonpartisan local news coverage are an important tool.[31] For example, in 2019, the IRS approved a waiver allowing the Salt Lake Tribune to become a nonprofit, three years after the Philadelphia Inquirer was purchased by a nonprofit.[32] The Federal Communications Commission could also reconsider its 2017 decision to end the requirement that broadcasters have a physical station within the areas in which they broadcast, or else consider other policies that may have the same effect of promoting locally oriented coverage.[33] As the U.S. considers the regulation of social media, it should consider ways of incentivizing the production and sharing of high-quality local news as well.


See endnotes in PDF


Are you interested in supporting the Manhattan Institute’s public-interest research and journalism? As a 501(c)(3) nonprofit, donations in support of MI and its scholars’ work are fully tax-deductible as provided by law (EIN #13-2912529).