By Loren Collingwood (University of New Mexico), Gabriel Martinez (University of New Mexico), and Kassra A.R. Oskooii (University of Delaware)
In 1982, Tucson, Arizona, birthed the sanctuary movement, with a minister of Southside Presbyterian declaring his church a sanctuary for immigrant refugees fleeing civil conflict in El Salvador and Guatemala. However, in 2019, despite being a broadly progressive city with a 2 to 1 advantage in registered Democrats, Tucson voted down a ballot initiative (Proposition 205) that would have made the city a sanctuary. While no single definition exists, sanctuary cities have two common elements: 1) an ordinance that forbids local law enforcement from inquiring into residents’ immigration status and, 2) limits on local law enforcement’s cooperation with Immigration and Customs Enforcement (ICE).
In our paper, “Undermining Sanctuary? When Local and National Partisan Cues Diverge,” we show that Tucson Democrats split their vote on the sanctuary initiative ultimately leading to its downfall. To understand this outcome, we advance a partisan mixed-cues theoretical framework in which Democratic elites communicated a split cue environment — some supported the initiative while others opposed it, whereas Republicans communicated universal opposition to the initiative. For instance, while the local Pima County Democratic Party backed the initiative, outgoing Democratic mayor, Jonathan Rothschild, penned a “vote no”’ op-ed in the Arizona Daily Star newspaper just three weeks before the 2019 vote, where he asserted: “If passed, Prop. 205 would harm our community in ways that have nothing to do with immigration. And, while intended to protect immigrants, it may actually make their situation worse.”
To capture the elite partisan dynamics, we conducted a media text analysis on news coverage about the initiative in both English and Spanish. We show that textual references to “Democrats” shows elites mixed in their support for the initiative, whereas references to “Republicans” show elites uniformly opposed.
Next, we gathered precinct-level election results from the Tucson City Clerk’s office, which includes the number of registered voters and the results for Prop. 205. Second, for each precinct, we generated a measure of latent partisanship by averaging the percent vote share in the 2018 election for the two top of the ticket Democrats. Finally, we conducted an ecological inference regression to assess whether Republicans are in uniform opposition to Prop. 205 and Democrats mixed.
The results reveal that 49% of Democrats supported Prop. 205 and 51% opposed it. In contrast, 99% of Republicans opposed the proposition. As a sensitivity analysis, we also analyzed the mayoral vote, which occurred contemporaneously as Prop. 205. An estimated 88 percent of Democrats backed Romero (the Democratic candidate), while an estimated 92 percent of Republicans voted for Ackerley (the independent candidate). Overall, these findings provide support for a mixed partisan elite cues explanation for the ballot initiative outcome.
To provide additional support for our findings, we conducted two real-world comparisons. Tucson is not alone in holding a public vote on sanctuary status. Two other recent ballot initiatives occurred in Humboldt County, CA, and the township of Greenfield, MA. We replicate our research design in these locations. As it turns out, Democratic elites in these two locations publicly voiced support for the sanctuary initiatives. The Humboldt County ecological inference regression results show partisan polarization as predicted by a partisan elite cues model: 82.5% of Democrats supported the pro-sanctuary initiative, with just 17.5% in opposition. As with Tucson, Humboldt Republicans opposed the measure. Similarly, the Greenfield results show 96% of Democrats backing the initiative, with just 1% in opposition. However, 77.5% of Republicans opposed the initiative, while 22% supported it.
Overall, all three analyses support the claim that local political contexts can still exert significant influence on citizens voting behavior and that in-party cues appear to matter more than previously assumed.
Finally, to take a deeper look at whether in-party cues could influence how Democrats respond to sanctuary city initiatives, we created a framing experiment. We randomly exposed self-identified Democrats to one of three conditions. Condition one is a control group with an informational statement on sanctuary cities. Condition two is the same as condition one but adds on language that Democratic mayors and council members are “divided” on supporting sanctuary city policies. Condition three includes the same language as the control but primes respondents that Democratic mayors and city council members are “united”. Results show that in the elites “divided” treatment group respondents are about 20-25 points less supportive of a sanctuary ballot initiative than are respondents in the control condition. Taken together, these results provide additional support that in-party cues may exert significant influence on whether Democrats display support toward sanctuary city policies.
These findings lead us to two important conclusions about partisan voting behavior in local elections. First, it suggests that local elites still have influence over their constituents even in the face of conflicting national elite cues. This means that scholars and political advocates should consider paying more attention to the ways in which political dynamics may unfold in local contexts. Another implication of our study is that local politics can still exert influence on voting behavior. However, the influence of local politics will be limited to ballot initiatives where a clear co-partisan divide between local and national elites is present. This sort of divergence is less likely to emerge on more crystallized issues. Democrats in states that are mostly controlled by Republicans may be more careful or conservative in their approach to sanctuary city policy. In the face of a hostile national and state government, they may not see much upside to officially declaring their city a sanctuary for undocumented immigrants.
Loren Collingwood is an associate professor in the Department of Political Science at University of New Mexico. He is the author of “Sanctuary Cities: The Politics of Refuge” (2019) and “Campaigning in a Racially Diversifying America: When and How Cross-Racial Electoral Mobilization Works” (2019) both with Oxford University Press.
Gabriel Martinez is a Ph.D. student in the Department of Political Science at University of New Mexico. His research interests include Latino Politics and Immigration Politics.
Kassra A.R. Oskooii is an associate professor in the Department of Political Science and International Relations at the University of Delaware. His research focuses on the interplay between the contextual and psychological determinants of political opinions and behaviors of high and low status group members.