THEORIES OF URBAN CONSTRAINT AND INNOVATION

Traditional theories of urban politics emphasize cities’ constraints rather than their innovation. Cities are subject to both state and federal regulations that constrain their policy-making autonomy. Moreover, theories of fiscal federalism argue that the smaller geographic scope of cities makes them more sensitive to mobility and, consequently, less able to redistribute. If cities raise taxes to redistribute services, higher-income individuals and businesses may move elsewhere (Peterson Reference Peterson1981; Tiebout Reference Tiebout1956). In part as a result, urban officials often ally with local businesses in regime-governance arrangements (Stone Reference Stone1989).

In recent years, however, large cities have often defied these expectations, producing urban policy innovation including redistributive efforts (Einstein and Glick Reference Einstein and Glick2018). Amid federal political polarization, some commentators even suggest that cities should rule the world given their pragmatic, problem-solving approach to policy (Barber Reference Barber2013). Efforts to explain the increasing innovation and influence of cities have led some scholars to argue that cities today are situated within a multilevel governance structure. Multilevel governance envisions a vertical dimension, in which intergovernmental influence is multidirectional, and a horizontal dimension, in which city decision making involves the private and nonprofit sectors as well as other jurisdictions in metropolitan areas (Kubler and Pagano Reference Kubler, Pagano, John, Mossberger and Clarke2012). For a situation to resemble multilevel governance, it must exhibit both this vertical dimension and a horizontal dimension that challenges “state/society boundaries” (Caponio and Jones-Correa Reference Caponio and Jones-Correa2018). I argue that the vertical dimension deserves particular attention regarding municipal responses to immigrants. Top-down federal regulatory constraints demonstrably shape local innovation through federal policy feedbacks. In turn, however, the vertical dimension likely will not be unidirectional—the local policies that emerge may shape future state and federal policy making.

FEDERAL POLICY FEEDBACK AND URBAN RESPONSES TO IMMIGRANTS

Few would dispute that local governments are constrained by state and federal policy mandates in a top-down intergovernmental relationship. Yet, federal policies also produce feedbacks that shape subsequent local innovation. Specifically, federal policies generate resources, incentives, and interpretive tools that shape local officials’ behavior and beliefs (Pierson Reference Pierson1993). Theories of policy feedback often examine how policies shape the subsequent behavior of interest groups (Skocpol Reference Skocpol1992) or mass publics (Campbell Reference Campbell2005; Mettler Reference Mettler2002). Instead, I focus on the effects of federal policies on local officials. Federal policies have both an instrumental effect—telling local officials what to do—and a symbolic effect—telling them what to believe (Bloemraad Reference Bloemraad2006; Marschall, Rigby, and Jenkins Reference Marschall, Rigby and Jenkins2011). In shaping local officials’ normative understanding of their role, federal policies powerfully shape and reinforce the trajectory of local political agendas.

Urban responses to immigrants provide a clear example of federal policy feedback. Federal policies constrain local officials to serve immigrants in several key ways. Schools must educate immigrants, including providing instruction for English-language learners, regardless of legal status (Lau v. Nichols 1974; Plyler v. Doe 1982). Where a substantial language-minority population is present, Title VI of the Civil Rights Act requires that local agencies provide language access. These policies provide an instrumental push requiring local officials to develop the resources and capacity to serve immigrants. At the same time, they frame immigrants as clients of local government services, shaping officials’ interpretations of foreign-born residents. In this way, top-down federal policies that require local services to immigrants generate resources and interpretive tools that encourage subsequent innovation to serve immigrants, setting cities on a trajectory toward local welcoming policies. Whereas federal policies, on balance, encourage local officials to serve immigrants, the increasing devolution of federal enforcement to the local police also can frame immigrants as lawbreakers and discourage local welcoming responses (Williamson Reference Williamson2018). Theories of bureaucratic incorporation illustrate how accommodating federal policies can amplify bureaucrats’ professional service ethos, while restrictive policies can heighten their regulatory impulses (Marrow Reference Marrow2011).

If, indeed, federal policies produce feedbacks that shape cities’ responses to immigrants, we expect that cities more exposed to welcoming federal policies would produce more welcoming local responses, whereas cities more exposed to restrictive federal policies would produce more restrictive local responses. Variation across cities in exposure to refugee resettlement and county-level detention of unauthorized immigrants allows us to test this hypothesis. These policies offer a particularly compelling test because cities do not control their exposure to refugees or county-level detainees. Therefore, preexisting preferences surrounding immigration are unlikely to drive the results. Although these federal policies are aimed at select subsets of immigrants, I demonstrate that differing exposure shapes responses not only to refugees and immigrant detainees but also to immigrants more broadly.

Refugee resettlement voluntary agencies determine the destinations of newly arrived refugees based on assessments of local capacity to absorb the newcomers. Local officials may be consulted in the process, but they do not approve resettlement and they rarely request it (Singer and Wilson Reference Singer and Wilson2006). Upon arrival, however, local officials must provide services to refugees beyond what is required for non-refugee immigrant newcomers. In addition to serving refugees in schools and public health clinics, and through language access, refugees are immediately eligible for locally administered federal social programs (Bernstein and DuBois Reference Bernstein and DuBois2018). In contrast, since the 1996 welfare reform, most other immigrants have been excluded from these programs for their first five years in the country, unless states opt to cover the expenses (Gelatt and Fix Reference Gelatt, Fix and Fix2007). Even beyond locally administered federal programs, local officials are more likely to encounter refugees, as voluntary agency caseworkers and co-sponsorship volunteers advocate for the newcomers. Qualitative studies demonstrate that federal refugee resettlement policies shape local officials’ responses by framing refugees as worthy clients of services (Horton Reference Horton2004). I further argue that exposure to refugee resettlement generates the capacity and inclination to serve not only refugees, but also spills over to immigrants more generally, spurring innovative local welcoming responses (Williamson Reference Williamson2018).

Likewise, with respect to immigrant detention, city-level officials do not control whether their resident county enters an intergovernmental service agreement with the Department of Homeland Security to house immigrant detainees. Nonetheless, when county facilities are involved with immigration enforcement, town and city officials—particularly law enforcement officials—are exposed to policies that frame immigrants as lawbreakers rather than as clients or community members. Often, town- and city-level arrestees are housed in county-level facilities, even in large cities. When local police book arrestees into county jails that house immigrant detainees, they interact with facilities and personnel who are involved in the day-to-day work of immigration enforcement (Armenta Reference Armenta2012). In this way, greater exposure to federal immigration enforcement at the county level is associated with diminished local initiative to implement welcoming policies.

To substantiate these claims, I drew on analysis of a survey of local government officials’ responses to immigrants, coupled with government administrative data. Full results are presented in Williamson (Reference Williamson2019) and summarized here. Evidence on local responses to immigrants is from the 814 towns and cities responding to the 2016 Municipal Responses to Immigrants Survey.^{Footnote 1} In each town, surveys were sent to the mayor, a randomly selected city councilor, the police chief, and the city manager or a comparable lead appointed official. The survey asked officials about local policies and informal practices in responding to immigrants, as well as their views on the local government role in this sphere.^{Footnote 2} The dependent variable of interest was an index of welcoming responses that captures the proportion of welcoming practices that a town implemented among those it was asked about. The index incorporates 19 local immigrant-serving practices, all of which surpass mere compliance and represent independent local efforts to welcome newcomers. Specifically, the index includes practices such as appointing immigrants to local boards, funding immigrant organizations, and partnering with nonprofit organizations to provide services to immigrants.

To measure whether exposure to refugee resettlement promotes local welcoming responses, I drew on the Worldwide Refugee Admissions Processing System administered by the State Department. I constructed a measure that sums refugee arrivals in each city between 2002 and 2016, then calculates the number per 1,000 local residents. To measure whether exposure to county-level immigrant detention dampens local welcoming responses, I drew on Immigration and Customs Enforcement (ICE) data to construct a measure of the county-level immigrant detention population in 2015 per 1,000 local residents.^{Footnote 3} To ensure that these variables are driving local welcoming responses rather than conflated factors, I controlled for a broad range of variables identified as predictors of local responses in past literature, including cities’ capacity, foreign-born presence, experience of ethnic threat, and partisanship (Hopkins Reference Hopkins2010; Provine et al. Reference Provine, Marie, Varsanyi, Lewis and Decker2016; Ramakrishnan and Wong Reference Ramakrishnan, Wong and Varsanyi2010; Steil and Vasi Reference Steil and Vasi2014; Williamson Reference Williamson2018).^{Footnote 4}

Even when controlling for these variables in multivariate analyses, both proxies of exposure to federal policies remain statistically significant predictors of local welcoming responses. Greater exposure to refugee resettlement is associated with an increased proportion of welcoming practices, whereas greater exposure to immigrant detention is associated with a decreased proportion. Specifically, holding other variables at their means, moving from a city with no refugee arrivals per 1,000 residents to a city with 15 refugees per 1,000 residents (roughly a standard deviation above the mean) is associated with an increase in accommodation of three percentage points. Conversely, when holding other variables at their means, moving from a city with no county-level detainees per 1,000 residents to a city with one detainee per 1,000 residents is associated with a decline in accommodation of three percentage points. Because the accommodating index consists of 16 practices on average, a change of three percentage points is equivalent to an increase or decrease of half of an accommodating practice. The magnitude is not massive, but it is remarkable that these relatively small changes in refugee or detainee presence—two factors outside of local officials’ control and not indications of their preferences—result in significant changes in local accommodation, even when holding constant local partisanship and a broad range of other factors.

In summary, when local officials are more exposed to federal requirements to serve refugees as worthy local clients, cities produce more independent local welcoming responses. When local officials instead are exposed to immigrant detention, they devote less energy to welcoming immigrants. These predictions echo the earlier findings on the bureaucratic incorporation of immigrants, which argued that state and federal policies toward immigrants shape service-oriented or regulatory responses from bureaucrats (Marrow Reference Marrow2011). In applying theories of policy feedback, I emphasize the processes through which federal policies produce both local capacity to serve immigrants and a normative understanding of immigrants’ role as community members. Moreover, the fact that varying exposure to these policies affects town-wide implementation of welcoming practices suggests that federal policies shape not only micro-level bureaucratic behavior but also the broader political agenda among municipal appointed and elected officials.

Although I argue that the top-down, vertical influence of the federal government is central in shaping local responses, it is worth considering the horizontal dimension. Refugee resettlement involves both the top-down dimension of determining which refugees may settle where and the horizontal dimension in which voluntary agencies (federal contractor nonprofit organizations) carry out many resettlement tasks. Likewise, immigrant detention involves vertical cooperation in which the federal government relies on county governments to provide detention beds it would not otherwise have through intergovernmental service agreements. However, it also involves a horizontal aspect, as evidenced by municipal partnerships with private detention facilities or participation in “pass-through contracts,” through which private contractors provide detention facilities with federal funding (Reiter Reference Reiter2018). Although these horizontal dimensions no doubt exist, this study argues that the vertical dimension is particularly influential when it comes to urban responses to immigrants. Regardless of whether voluntary agencies or private detention contractors influence local officials, it is federal policies that drive these institutions’ local presence. Local officials have little power over whether refugees are settled locally or whether their resident county partners with ICE to detain immigrants; nonetheless, exposure to these federal policies shapes their understanding of immigrants and their policy making. Moreover, even when controlling for elements of the horizontal dimension—such as the number of immigrant NGOs per 1,000 residents and the dominance of local agribusiness interests (i.e., percentage of agricultural employment)—the influence of exposure to these two federal policies remains consistently significant.

URBAN INNOVATION AND THE FUTURE OF FEDERAL REFORM

These results make clear that federal policy feedbacks channel urban policy innovations from the top down. When these data were collected from February to May 2016, local officials already were experiencing both federal requirements to welcome immigrants and federal efforts to involve local law enforcement in immigration enforcement. Since then, the Trump administration has promoted more restrictive policies and rhetoric, as well as greater political polarization with respect to immigration. Nevertheless, federal requirements to serve immigrants in schools and through language access remain, as does the legacy of these policies in shaping local innovations to welcome immigrants. Indeed, in the face of the Trump administration’s attempts to curb urban efforts to welcome unauthorized immigrant residents, cities have fought back.

These innovative urban policies to welcome immigrants can, in turn, feedback to shape the future of federal reform efforts from the bottom up. Gulasekaram and Ramakrishnan (Reference Gulasekaram and Karthick Ramakrishnan2015) argued that immigration policies established on the state and local level anchor national partisan preferences, thereby constraining options for future federal reform. Immigration federalism, they wrote, represents a “multi-jurisdictional turf battle to instantiate competing visions of desired national immigration policy” (Gulasekaram and Ramakrishnan Reference Gulasekaram and Karthick Ramakrishnan2015, 118). Innovative urban welcoming policies are particularly likely to generate policy feedbacks that shape the future of comprehensive reform. These local policies redefine membership in ways that (1) shape elite and mass understandings of immigrants, and (2) potentially stimulate participation among immigrants in the interest of expansionist reform.

For a policy feedback to occur, the implementation of a particular policy must represent a critical juncture that generates resources, incentives, and interpretive tools that reproduce political incentives in a path-dependent process (Rast Reference Rast2012). Mettler and Soss (Reference Mettler and Soss2004) have further specified that policy feedbacks can powerfully define community membership, delineate salient groups, and build those groups’ capacity for participation. Cities’ recent efforts to provide municipal identification to unauthorized immigrants, or to declare a sanctuary zone from immigration enforcement, define immigrants—even those who lack national citizenship status—as municipal citizens (Maas Reference Maas, Shachar, Bauböck, Bloemraad and Vink2017). In these cities, millions of residents are receiving a symbolic message that unauthorized immigrants are worthy of membership, shaping the views of mass publics and associated elites. At the same time, these innovative urban welcoming policies shape the beliefs of the immigrants themselves, creating the potential to further activate immigrants as a political constituency and to generate participation in favor of expansionist reform.

Urban policy making is distinct given the tertiary position of cities in the intergovernmental system and the resulting federal and state policy constraints they face. Some scholars have argued that cities are increasingly “acting at eye level with upper-level authorities” in multilevel governance arrangements (Kubler and Pagano Reference Kubler, Pagano, John, Mossberger and Clarke2012, 124). Other scholars herald the rise of a new “progressive federalism,” evident in the policy innovations emerging across large, liberal cities, as well as the influence these policies have on the national debate (Gerken Reference Gerken2012). The example of local government responses to immigrants substantiates the claim that ideas and influence can flow upward from the local level. Yet, the evidence presented here also demonstrates that differing exposure to federal policies is associated with corresponding variation in local practices. Even in an age of increasing urban influence, federal policies remain demonstrably powerful not only in what they ask local officials to do but also in what they encourage them to believe, thereby shaping the agenda for subsequent, independent policy innovations.