Rethinking Sanctuary: The Origins of Non-Cooperation Policies in Social Welfare Agencies

Abstract

Too often, scholarship on immigration conflates sanctuary ordinances with the non-cooperation policies, often embedded in these ordinances, which limit cooperation between local officials and federal immigration authorities. In this article, I disentangle the two by tracing the rise of non-cooperation policies in health and welfare agencies since the New Deal. Doing so challenges assumptions about the origins, targets, consequences, and significance of early sanctuary policies. It reveals that non-cooperation was federal policy between 1935 and the early 1970s, when local, state, and federal officials began to experiment with cooperation. When the consequences of such practices became clear, welfare and health officials were forced to reaffirm non-cooperation just before the sanctuary movement burst onto the scene. This research clarifies why scholars see early sanctuary ordinances as largely symbolic: because many local, state, and federal officials had largely abandoned cooperation in practice. It also challenges the widespread assumption that non-cooperation fundamentally represents local resistance to federal power. Instead, I demonstrate the key role played by the federal government in the rise of non-cooperation in health and welfare agencies. Lastly, this research reaffirms the significance of the fragmented nature of federal institutions for promoting immigrant rights.

INTRODUCTION

In the wake of the 2016 presidential election, hundreds of communities across the United States adopted or reaffirmed sanctuary policies (Graber and Avila 2019). While critics denounced these efforts (Villazor and Gulasekaram 2018), a debate ensued among scholars, advocates, and pundits alike about what exactly “sanctuary” entailed (Villazor 2008; Bauder 2017). “Sanctuary” can encompass a variety of actions taken by private or public actors (Villazor 2008, 137; Villazor and Gulasekaram 2018; Ayers 2021), including declarations of support or solidarity (McMillam 1987); sheltering immigrants in churches (Coutin 1993); providing attorneys to non-citizens in removal proceedings; and “policies and practices designed to make state and local services available to immigrants” (Kagan 2018, 393–94; Motomura 2018).

In practice, sanctuary city policies nearly always include some pledge of limited or non-cooperation between state and local officials and federal immigration authorities. Among other things, this may include limits on information sharing or limiting compliance with immigration detainer requests. Local officials may also refrain from collecting data about legal status to limit the possibility of cooperation with immigration authorities. Non-cooperation has thus become virtually synonymous with sanctuary (Varsanyi 2010, 3n4; Villazor 2010; American Immigration Council 2015; Bauder 2017, 176; Colbern 2017, 204; A. Garcia 2018, 186; Lasch et al 2018, 576n20; Collingwood and Gonzalez O’Brien 2019; Kramer 2020; Su 2020; de Graauw 2021).

The conflation of sanctuary and non-cooperation may obscure more than it reveals. Many scholars trace the origins of sanctuary policies (and, therefore, non-cooperation policies and practices) to the sanctuary movement in the mid-1980s. Most of the ordinances passed at that time were inspired by the mobilization of religious groups offering church-based sanctuary to Central American migrants fleeing civil war (Merina 1985; Colbern 2017). But the non-cooperation policies that were often embedded in these sanctuary ordinances have a much longer history. In this article, I detail the longer history behind non-cooperation policies in welfare and health agencies. Doing so reveals that—in line with broader concerns over privacy and confidentiality—non-cooperation was federal policy between 1935 and the early 1970s when local, state, and federal officials began several experiments in cooperation between welfare, health, and immigration agencies. When the consequences of such practices became clear, local, state, and federal officials were forced to reconsider these experiments in cooperation. By the time cities began to adopt sanctuary ordinances in 1983, many local, state, and federal officials had already reaffirmed policies and practices that greatly limited cooperation between welfare, health, and immigration officials.

Disentangling sanctuary from non-cooperation helps challenge several assumptions about the origins, targets, consequences, and significance of early sanctuary policies. It reveals that concerns about undocumented immigrants, not simply Central Americans fleeing civil war, motivated the re-emergence of non-cooperation policies and practices in the 1970s and 1980s, especially in cities with large numbers of undocumented immigrants. It also helps us better understand why many scholars deemed sanctuary ordinances to be largely symbolic: because many local, state, and federal officials had already abandoned cooperation a few years earlier.

This research also challenges the widespread assumption that non-cooperation policies necessarily represent local resistance to federal power. Instead, I demonstrate the key role played by the federal government in the rise of non-cooperation policies. What was novel about sanctuary ordinances in the 1980s was not non-cooperation but, rather, the public declaration of non-cooperation as citywide policy and, in many, though not all, cities, the public expression of solidarity for the plight of Central American and other migrants. While the expressive function of sanctuary represented local resistance to federal power, the policy of non-cooperation in health and welfare agencies did not.

This article is divided into four parts. First, I review the multidisciplinary literature on sanctuary and make the case for disentangling sanctuary from non-cooperation. Second, I outline the origins and development of non-cooperation policies in federal welfare law since the New Deal. Third, I briefly describe the rise of experiments in cooperation between welfare, health, and immigration officials in the 1970s. Using two short case studies, I demonstrate the problems with cooperation and the backlash that ensued, leading to the reemergence of non-cooperation policies shortly before the advent of the sanctuary movement. Lastly, I reassess the significance of the early sanctuary ordinances in light of this longer history, and I detail the implications for the broader literature.

FROM SANCTUARY TO NON-COOPERATION

Sanctuary

With few exceptions, immigration scholarship—in sociology, law, political science, and geography—shares a common set of assumptions about the origins, targets, consequences, and significance of early sanctuary city policies. This literature suggests that sanctuary ordinances (or resolutions) first emerged in the United States in the mid-1980s (Bau 1994; Wells 2004; Pham 2006; Coleman 2007; Ridgley 2008; Villazor 2008; Hing 2012; Mancina 2016; Bauder 2017; Lai and Lasch 2017; A. Garcia 2018; Colbern, Amoroso-Pohl, and Gutiérrez 2019). Some scholars note precursors to sanctuary ordinances in earlier periods, including Medieval Europe (Bauder 2017) and ancient Egyptian, Greek, or Roman cultures (Bauder 2017; Ayers 2021) or highlight its biblical roots (Crittenden 1988; Ridgley 2012; Bauder 2017; Ayers 2021) or similarities to the mobilization against fugitive slave laws (Crittenden 1988; Colbern 2017; Colbern, Amoroso-Pohl, and Gutiérrez 2019; Ayers 2021). Others argue that sanctuary evolved from resisters and conscientious objectors to the war in Vietnam in the early 1970s (Bau 1994, 51; Ridgley 2012; Bauder 2017; Ayers 2021).

The actors who pushed for the adoption of sanctuary ordinances were religious groups (Bau 1994; Wells 2004; Pham 2006; Coleman 2007; Ridgley 2008; Villazor 2008; Hing 2012; A. Garcia 2018), though the coalition grew to include “university campuses, legal, human rights and civil liberties groups” (Ridgley 2008, 66) as well as “organized civil society actors working for social justice in El Salvador” (Perla and Coutin 2012, 73). The target group—those to be protected by sanctuary—are identified as Central American undocumented migrants fleeing US-backed civil wars who were, according to activists, inappropriately denied asylum (Bau 1994; Wells 2004; Pham 2006; Coleman 2007; Ridgley 2008; Villazor 2008; Hing 2012; Mancina 2016; Bauder 2017; A. Garcia 2018). While the church-based sanctuary movement began in 1982 (Coutin 1993), the first city to pass a sanctuary resolution (in Madison, Wisconsin) did so in 1983. By 1987, twenty-nine cities and four states had adopted a sanctuary policy (Colbern 2017, 178–79). The rise of public sanctuary declarations, Allan Colbern argues, reflects a shift in strategy among sanctuary movement leaders after repeated efforts to change federal law failed as well as the well-publicized federal prosecution of sanctuary movement leaders in Arizona (2017, 185).

Sanctuary city ordinances varied in their content (McMillam 1987). Some established their city as a “sanctuary,” a “refuge,” or a “City of Peace” or commended local groups for providing services and assistance to “refugee” communities. Some declarations supported pending federal legislation or recommended changes in federal policy toward Guatemalan and Salvadoran refugees. Beyond that, many sanctuary resolutions included concrete policies prohibiting city employees from cooperating with immigration officials for enforcement purposes or “inquiring or disseminating information about a person’s immigration status unless it was affirmatively required by federal or state statute, regulation, or court decision” (Ridgley 2012, 225; Merina 1985; Bau 1994; Mancina 2012).

Because early sanctuary city ordinances often included non-cooperation policies, immigration scholars often suggest that non-cooperation policies emerged from the sanctuary city movement (Varsanyi 2010, 3n4; Mancina 2019, 253) and only later evolved beyond the purpose of protecting Central Americans to protect the greater undocumented community (Ridgley 2012, 225; A. Garcia 2018, 190; Collingwood and Gonzalez O’Brien 2019, 4) or to apply as “more general protections for all immigrants” (Villazor 2008, 143; see also Bau 1994, 54). According to Huyen Pham (2006, 1385), by 1996, “the practical impact of the sanctuary movement had diminished because its intended beneficiaries, Guatemalans and Salvadorans, became eligible for special refugee consideration.” Harald Bauder (2017, 176) argues that the New Sanctuary Movement, which started in 2007, represents the moment when the movement “shifted focus from newly-arrived refugees to illegalized migrants.”

Scholars also typically portray these sanctuary ordinances as symbolic (Gulasekaram and Villazor 2009; Bauder 2017) because they were nonbinding (Hing 2012); they “did not directly alter the day to day operations of the city” (Ridgley 2008, 68); they were “unlikely to yield any substantive change to current legal doctrine” (A. Garcia 2018, 189); they “merely reiterate rights already protected with the state police power or other constitutional limitations” (A. Garcia 2018, 206); or because they could not prevent the federal government from deporting people in their jurisdictions (Villazor 2008, 150). ¹ Peter Mancina (2012, 212) says San Francisco’s ordinance was “largely moral and symbolic in its purpose…. It created no institutional oversight body, no chain of command, no procedures for serving refugees, and no guidelines for when or how disciplinary action should be administered to non-compliant city employees.” Miriam Wells (2004, 1319) argues that these resolutions provided few specifics as to what noncooperation meant and few sanctions for its breaching. As such, the federal government “tended to ignore these first sanctuary resolutions,” notes Jennifer Ridgley (2008, 68), since the Immigration and Naturalization Service (INS) “rarely depended on city resources or service providers to assist with the enforcement of immigration law, and federal legal opinions and judicial decisions … supported only a very limited role for local police.”

The fact that such ordinances were often symbolic does not mean they were not significant (Edelman [1964] 1974). Symbolic law can empower activists and boost organizing or affect the ideological conditions of a movement (Calavita 1983). Thus, even though sanctuary laws were “not impactful on substantive legal doctrine,” Alyssa Garcia (2018, 208) writes, they carried “significant normative influence simply because they are laws.” Such laws “served as a reference point for governmental legislators.” They also “provided activists and politicians a point of focus for political work” (Mancina 2012, 212) or became a “mechanism for organizing and public education” (Ridgley 2008, 67). They also helped to build trust with local immigrant communities, with positive consequences for public safety (A. Garcia 2018, 207), including higher reporting of violent crime (Martínez-Schuldt and Martínez 2021).

Lastly, immigration scholars tend to argue that sanctuary ordinances represented local resistance to federal power (Colbern 2017; Villazor and Gulasekaram 2018; Colbern, Amoroso-Pohl, and Gutiérrez 2019). While religious organizations pressed for these policies, sanctuary ordinances, we are told, “are strictly the products of local and state governments” (A. Garcia 2018, 187). As such, the concerns that sanctuary—and, therefore, non-cooperation—raises are ones related to federalism (Bauder 2017; Lai and Lasch 2017), the separation of powers, and the police powers of states. Doctrinal debates about the old or new sanctuary movement, therefore, consider whether state and local government can refuse to enforce federal law or shield migrants from the force of the federal government (Bau 1994) or “whether the Tenth Amendment prohibits the federal government from conscripting state and local governments into doing their work for them” (Villazor and Gulasekaram 2018, 553–54; Pham 2006).

Non-Cooperation

In this article, I call for greater conceptual distinction between sanctuary ordinances and the non-cooperation policies often imbedded in these ordinances. I shift the focus away from the origins of sanctuary ordinances to look instead at the origins of non-cooperation policies and practices in public welfare and health agencies. Non-cooperation in this context refers to policies prohibiting information sharing with immigration officials. Focusing on non-cooperation challenges several assumptions in the sanctuary literature. It reveals that the modern origins of non-cooperation in welfare and health institutions date to the 1970s, not the 1980s. The targets (or beneficiaries) of non-cooperation were not Central American asylum seekers but, rather, all undocumented immigrants. The key actors behind this push for non-cooperation in the 1970s were a coalition of immigrant rights activists, medical professionals, legal aid lawyers, as well as local, state, and federal officials. Some of these victories were dependent on laws established by the federal government during the Great Depression and therefore represented a return to a practice of non-cooperation that had been in place nationally for decades.

To be sure, some scholars acknowledge that sanctuary ordinances in the 1980s were built on earlier non-cooperation policies like Special Order 40 in Los Angeles, passed in 1979, which limited cooperation between the Los Angeles Police Department (LAPD) and the INS (National Immigration Law Center 2008; Villazor 2008, 142n59; Sullivan 2009; Mitnik and Halpern-Finnerty 2010; Hing 2012, 253–54). ² Yet allusions to Special Order 40 in the sanctuary literature are usually passing references (Villazor 2008, 142n59; Sullivan 2009; Mitnik and Halpern-Finnerty 2010, 54), even as some declare it “the country’s first sanctuary policy” (Sullivan 2009, 571). Loren Collingwood and Benjamin Gonzalez O’Brien (2019, 18) devote the most attention to Special Order 40—a single page of their book—acknowledging that the “legal precedent for the sanctuary city was thus set well before the beginning of the Central American refugee crisis.” However, they argue that there was no “significant spread” of non-cooperation “policies like Special Order 40” before the advent of the sanctuary movement, ignoring non-cooperation policies in other agencies. ³ Indeed, there has been virtually no attention to the history of non-cooperation in welfare and health agencies, even though they were far more common than policies limiting cooperation between police and immigration officials. ⁴

Specifically, I demonstrate that confidentiality provisions adopted by federal welfare officials and encoded in the Social Security Act largely prohibited cooperation between welfare and immigration officials from the mid-1930s through the early 1970s. ⁵ During the early 1970s, however, the federal government adopted policies that, for the first time, barred undocumented immigrants from Social Security numbers and virtually all federal welfare programs. The federal government also weakened the confidentiality provisions that prohibited cooperation between welfare and immigration officials. Thereafter, various hospitals and welfare agencies across the country began reporting immigrants to the INS for verification or enforcement purposes. Fairly quickly, local, state, and federal officials discovered that cooperation caused problems. It endangered immigrant health as well as broader public health. It often violated the rights of US citizens, existing federal or state laws as well as the ethics of medical professionals. It also resulted in the shifting of costs to lower levels of government and onto hospitals. Local and state officials, therefore, largely abandoned these experiments in cooperation—at times because of the demands of the US Department of Health, Education and Welfare (HEW)—all shortly before the advent of the sanctuary movement. Indeed, in 1977, Texas was forced to abandon its policy of reporting to immigration officials all undocumented immigrants identified by Aid to Families with Dependent Children (AFDC), Food Stamp, or Medicaid staff, a policy that had been in effect since 1973.

My argument draws on research for an ongoing study on the rise of legal status restrictions in federal welfare policy since the New Deal. That study includes in-depth case studies of California, New York, and Texas, allowing me to examine the origins, implementation, and consequences of welfare policy choices. I rely on a wide range of sources, including federal, state, and local government archives, reports, statutes, and hearings; a database of early sanctuary ordinances; ⁶ the personal papers of immigrant advocates; newspaper articles; and court cases. Examining this history of non-cooperation policies in public welfare and health agencies helps us better understand why 1980s sanctuary ordinances were deemed largely symbolic—because many health and welfare officials had already stopped cooperating with the INS before the sanctuary movement emerged. In addition, it illustrates the key role played by the federal government in the rise of non-cooperation policies, challenging the idea that non-cooperation always represents local resistance to federal power. It also demonstrates the significance of the fragmented nature of federal institutions for facilitating the promotion of immigrant rights (Wells 2004; de Graauw 2021). Relatedly, it helps explain why federal officials did little to resist these sanctuary ordinances when they first arose: because different federal agencies disagreed about the value and propriety of cooperation. When federal officials did resist, they resisted not the policy of non-cooperation but, rather, the public declaration of sanctuary/non-cooperation, fearing that such proclamations would encourage undocumented migration and settlement.

THE RISE OF NON-COOPERATION

Non-cooperation policies were embedded in federal welfare law during the Great Depression. The Social Security Act of 1935 established the modern American welfare state. ⁷ In addition to establishing means-tested programs such as Aid to Dependent Children, it also created social insurance programs, including Social Security. There were no federal citizenship or legal status restrictions for any of these programs. Even undocumented immigrants were eligible for benefits on the same basis as citizens, though states could bar non-citizens from means-tested programs if they desired (Fox 2016).

Social Security

By design, confidentiality provisions limited cooperation between Social Security and immigration officials. The passage of the Social Security Act “set in motion a huge effort to build the infrastructure needed to support a program affecting tens of millions of individuals” (Puckett 2009, 55). Employers had to deduct payroll taxes from the wages of workers, and federal officials had to craft a method to track contributions. To do so, they devised a numbering scheme assigning all eligible workers Social Security numbers. The Social Security Board’s (SSB) Informational Service embarked on a campaign to encourage workers to complete an “application-for-account number” and return it to the post office (59–60, 73). From the outset, many people—native and foreign born alike—were concerned about how the information collected would be used, and the SSB was keen to allay their concerns (Crank 1985; Igo 2018).

Social Security officials made two key decisions that limited the possibility of cooperation with immigration officials. First, the SSB did not collect information that could be used to track down non-citizens. Worried that immigrants were hesitant to apply for Social Security numbers, the Foreign Language Information Service (FLIS) put out a press release in 1937 to clarify the rules under which the SSB was operating. They assured immigrants that “[t]he application forms … do not ask whether the applicant is a citizen or an alien, or, if an alien, when or in what manner he or she entered the United States” (qtd. in Fox 2012, 261). Second, the SSB promised that whatever information they did collect would not be disclosed to anyone. In 1936, “the day before applications for social security numbers … were distributed, the Board issued a press release stating: … ‘The information required of every worker on this form would be regarded as confidential.’ … Only … [g]overnment employees having official responsibility in connection with the social security files will have access to this information” (Crank 1985, 8).

Perhaps to ensure that non-citizens would feel more comfortable filing for Social Security numbers, the FLIS encouraged immigrants to file under the name “by which the applicant is known on the payroll of his employer. For record identification purposes, it does not matter whether this is the alien’s true name, or a name assumed since arrival in America” (qtd. in Fox 2012, 261–62). Being able to file under their “American” name was critical because it decreased the chances of the INS using the information in the Social Security files if it was ever granted permission. There were instances in which the INS requested such permission. In 1937, a post office in Minneapolis, asking on behalf of an immigration inspector, requested “certain information from the files furnished the Social Security Board” (262). Because the SSB regarded the information they collected to be confidential, they had previously rejected any requests for such information. Frances Perkins, who as secretary of labor also oversaw the INS, did not protest the withholding of such information, noting that “it is not the intention of the Department of Labor to make any applications for information which are contrary to this well established [confidentiality] policy” (262). Immigration field officers were thereafter instructed to “refrain from requesting information from the Social Security Board for immigration, naturalization or other purposes” (262).

The SSB inscribed this principle of confidentiality in its very first regulation, the “Disclosure of Official Records and Information,” published in 1937. Two years later, “congress incorporated … the provisions of Regulation No. 1 into section 1106 of the Social Security Act,” strengthening it by including “a criminal penalty for any unauthorized disclosure of information in social security files” (Crank 1985, 8–9). Though Congress initially ratified the confidentiality provisions developed by federal welfare agencies, it attempted to increase cooperation between Social Security and immigration officials in the early 1950s (US Senate Committee on the Judiciary 1950, 330, 333). A bill introduced in the US Congress (1952, 70) provided that Social Security officials notify the Attorney General “whenever any alien is issued a social security number and social security card.” The measure also provided “that the Administrator shall also furnish such information as may be requested by the Attorney General regarding aliens employed in the United States.” To do so, Social Security officials would have to ask applicants about their citizenship status.

The Federal Security Agency (FSA), which oversaw the SSB, resisted collecting information about the citizenship or legal status of Social Security applicants and broader efforts to promote cooperation between the SSB and the INS (US Congress 1952, 69–75). The FSA worried that anything beyond very limited cooperation for the purpose of national defense and security would undermine trust in the Social Security system. In addition, the FSA doubted “the potential value” of Social Security records in locating undocumented immigrants, especially considering the “burden and expense of the contemplated procedure” (71). In response to these concerns, Congress adopted new language, included in the 1952 Immigration and Nationality Act, which authorized the release of information about non-citizens who applied for Social Security numbers to the Attorney General, but only upon the Attorney General’s request for information on a specific individual. ⁸

In the immediate aftermath, cooperation between Social Security and immigration officials remained limited in large part because Social Security officials continued to refrain from asking questions about applicants’ citizenship or legal status. It was only in 1972, when the federal government first barred undocumented immigrants from obtaining Social Security numbers, that any widespread cooperation with the INS became possible. Even then, cooperation remained limited, especially after the passage of the Privacy Act of 1974. ⁹

Means-Tested Assistance

Social Security was a federal program but Old Age Assistance, Aid to the Blind, and Aid to Dependent Children were administered (and partially financed) by state and local communities. In line with their efforts to modernize state and local welfare provision, federal welfare officials tried to “persuade state and local welfare departments to keep welfare records confidential” (Tani 2016, 161). Confidentiality provisions were in keeping with the principles “basic to professional social welfare practice” (Social Work Journal 1952, 88, 93) in order to “protect beneficiaries against humiliation and exploitation” (US Congress 1939a, 1939b). ¹⁰ Specifically, federal welfare officials objected to the common poor-law practice of sharing information to shame or discipline welfare recipients. Some local officials published the names of relief recipients in the newspaper. Others supplied such names to “taverns … to prevent them from spending the money on drink” (Tani 2016, 161). Federal officials wanted communities to abandon these practices.

Federal welfare officials were also concerned that state and local agencies would share information with other government officials. The SSB objected to politicians’ use of “recipient lists to solicit votes” (Tani 2016, 161). Federal welfare officials also worried that state and local welfare agencies might share information with immigration authorities, a practice common in the Southwest in the early years of the Great Depression (Fox 2012, 149–53). After Franklin Delano Roosevelt was elected president, therefore, federal welfare officials often refused when communities asked that the names of relief recipients be shared with immigration officials. An INS district director asked a federal welfare official in Arizona to share “the names and addresses of alien members of the Communist Party” so that the INS could deport them. ¹¹ The welfare official contacted Harry Hopkins, Roosevelt’s top relief official, asking whether cooperation was advisable. Hopkins told her to “have nothing to do with this request.” ¹²

There was nothing in federal law to prevent the sharing of information about welfare recipients with the public, politicians, or other government agencies. The SSB, therefore, recommended that Congress amend the Social Security Act to condition “states’ grants on their agreement to safeguard client information” (Tani 2016, 161). Congress agreed, and the 1939 Amendments to the Social Security Act contained provisions that required that state plans for means-tested assistance must “provide safeguards which restrict the use or disclosure of information concerning applicants and recipients to purposes directly connected with the administration” of categorical assistance programs (Social Service Review 1952, 229). ¹³ I found no evidence that concerns about cooperation between state and local welfare and federal immigration officials specifically drove the SSB to press for confidentiality provisions (229n2). Nevertheless, until the early 1970s, federal and state officials interpreted this confidentiality provision to prevent cooperation between state and local welfare agencies and federal immigration officials (US General Accounting Office 1973, 41–46).

EXPERIMENTING WITH COOPERATION

Starting in the early 1970s—amid a broader backlash against welfare and undocumented immigration—local, state, and federal welfare officials began to experiment with cooperation (Fox 2016, 2019). California moved first, barring undocumented immigrants from access to welfare in 1971, part of a larger bill to reform the state welfare system. Thereafter, non-citizens who applied for AFDC had to fill out a status verification form that was forwarded to the INS (Fox 2019). Shortly thereafter, the federal government moved into the fray. The 1972 Amendments to the Social Security Act denied undocumented immigrants’ access to Supplemental Security Income (SSI) and Social Security numbers. ¹⁴ Within four years, the federal government had also barred undocumented immigrants from AFDC, Medicaid, Food Stamps, and Unemployment Insurance (Fox 2016).

In 1972, the federal government also loosened up its confidentiality provisions. This move appeared to open the door for cooperation across agencies, providing that “information under welfare programs could be disclosed to law enforcement officials” (US General Accounting Office 1973, 42, 44–46). In theory, the relaxation of the federal confidentiality provisions applied to Social Security and SSI. ¹⁵ Through a legislative oversight, it did not initially apply to AFDC and Medicaid. Federal officials nevertheless concluded that state welfare officials were permitted to share information with the “INS about applicants and recipients for the purpose of determining the legal status of these persons.” ¹⁶ State and local welfare agencies were thus expressly permitted to follow California’s lead in conducting status verification to determine whether an individual was undocumented and therefore ineligible for aid (Dwight 1973).

In 1974, Congress amended the Social Security Act to require “that a State plan permit disclosure of information on recipients or applicants to public officials who need it in their official duties.” ¹⁷ This measure applied to AFDC, and it was inserted to ensure that the 1939 confidentiality provisions could “not be used to prevent” law enforcement or other public officials “from obtaining information required in connection with his official duties such as obtaining support payments or prosecuting fraud or other criminal or civil violations” (US Senate Committee on Finance 1974, 49). It is not clear whether HEW actively encouraged states to cooperate with the INS, but it did approve state welfare plans that required cooperation with the INS, either for verification or for immigration enforcement purposes.

In 1974, the US Department of Agriculture (USDA) issued regulations limiting food stamps to citizens and permanently residing aliens (Feltner 1974, 25996; Yeutter 1974, 3643; Congressional Research Service 1975, 26). The regulation said nothing about cooperation (Feltner 1974, 25996), but the new food stamp certification manual did, requiring verification of legal status for “questionable” cases (US Department of Agriculture, Food and Nutrition Service 1974, 31). The manual also contained a reporting requirement for immigration enforcement purposes. Under the new directive, eligibility workers were told that “[i]f in the application process, it becomes known to the State agency that an alien has entered or remained in the United States illegally … such alien shall be promptly brought to the attention of the INS … for appropriate action” (US Department of Agriculture, Food and Nutrition Service 1974, 34; see also US Congress 1974, 4, 145–48). It is unclear why the USDA included this reporting requirement in their handbook. HEW officials were puzzled at the move, noting that “this provision is not found in any public regulation.” ¹⁸ Despite the shaky legal foundations, the regulation was enforced.

During the 1970s, welfare officials across the country began to cooperate with the INS either to verify the legal status of applicants or to report undocumented immigrants to the INS for enforcement purposes. ¹⁹ The Food Stamp policies applied nationally, while each state developed its own policies for jointly financed and administered welfare policies. Texas and New York opted for a direct reporting requirement, while California required cooperation for status verification only (Farber 1974). ²⁰ The primary purpose of status verification in California was not to facilitate deportation but, rather, to ensure that applicants were eligible for services. Nevertheless, the distinction was often meaningless since the INS could use the information for enforcement purposes (Public Hearing on Undocumented Persons 1978). Various counties in California also adopted direct reporting policies of their own. ²¹ We know less about cooperation outside California, New York, and Texas, but these three states alone were home to roughly 70 percent of undocumented immigrants in the United States at the time (Passel 1986).

Immigrant rights’ lawyer Peter Schey and a colleague explained that by linking welfare and immigration agencies—whether through status verification or direct reporting—welfare agencies in the 1970s were being transformed into “the law-enforcement arm of the migra.” They charged that “aliens in ‘questionable cases’ have been locked in rooms while trying to apply for public assistance while INS was called to come and pick them up.” As a result, a “high percentage of aliens being deported by INS … are people who were forced to ‘cooperate’ with INS as a condition of receiving welfare.” ²² Schey (1977, 7) grew especially concerned about a California regulation that required applicants to attend an INS interview, which had turned “the welfare system into the most effective law enforcement mechanism available to the INS in the entire Southwest.” He claimed that “[a]pplicants, naïve enough to attend ‘interviews’ with INS as a pre-requisite to receiving benefits, have disappeared and lost contact with their families.” The new policies “have led to the motto: ‘Every Welfare Worker an Immigration Law-Enforcement Officer’. … Many thousands of potentially qualified applicants simply avoid contact with the welfare system out of fear for the ‘INS interview’” (7).

It is not clear how many people were deported due to such cooperation. There is reason to believe that cooperation induced a chilling effect, as immigrants who did not know their immigration status or those with US-born children avoided seeking health care or applying for benefits for which they remained eligible. We do have some sense of the scale of the verification program in southern California. Between 1975 and 1977, the district director of immigration in Los Angeles processed more than fifteen thousand status verification forms. Of these, 27 percent of applicants were in the country legally. The rest were either in the country illegally (30 percent) or failed to appear for an interview (43 percent). INS officials presumed these no-shows were unlawfully present as well (Public Hearing on Undocumented Persons 1978, 97–98). It is possible that as many as 11,400 people might have been expelled because of the status verification program in and around Los Angeles in this two-year period alone.

Fairly quickly, however, local, state, and federal officials discovered that cooperation between welfare, health, and immigration officials had negative consequences. ²³ At the urging of immigrant activists, medical professionals, legal aid lawyers, and, eventually, local, state, and federal welfare officials, they largely abandoned these experiments in cooperation. A few examples drawn from California and Texas help to illustrate the larger trend.

County Officials Reevaluate Cooperation

In the summer of 1977, Orange County officials adopted a policy requiring that indigent patients apply for Medi-Cal, the state’s Medicaid program (Public Hearing on Undocumented Persons 1978, 90). By state regulation, “information on suspected undocumented immigrants” who applied for AFDC or Medi-Cal was “sent to the INS for verification of immigration status.” The county further decided that those who refused to sign a Medi-Cal application “would be considered ‘uncooperative’”; University of California Irvine (UCI) Medical Center, which contracted with the county to provide care for indigent patients, would be required to forward their names and addresses to the INS as well. ²⁴

Orange County officials adopted the policy in part to cut costs. Per state policy, if immigrants applied for Medi-Cal the state would share the cost of medical care provided by the county while the individual’s legal status was being verified, a process that could take three to six months. If immigrants refused to apply for Medi-Cal, the entire cost of care was borne by the hospital and/or county.

Some UCI medical staff were incensed at the new cooperation policy. The College of Medicine dean reported feeling “disgust, shame, and anger” at being pressured to “implement what are essentially immigration service functions.” ²⁵ “Recent county maneuvers to qualify uncertified workers for Medi-Cal have mocked the human rights” of immigrants, he averred. “People have rights whether they are citizens or not. Uncertified citizens have rights, particularly in a country which clearly has chosen to tolerate if not encourage their presence.” He saw the new policy as a “serious breach in the doctor-patient relationship,” ²⁶ and he wrote that “[t]o provide these names to the INS … is a profound violation of basic human rights.” ²⁷ Medical students also resisted, adding: “This policy violates our long-held premise that medical care is a right for all, not simply a privilege for those who can afford it.” ²⁸

Medical personnel highlighted the potential costs and public health implications of these policies. Medical students predicted that undocumented immigrants would “soon avoid seeking aid from our facilities except in the most critical cases; thus pre-natal care, well-baby care, and early acute care (when treatment is best, easiest, and cheapest) would cease for such patients.” They told their Dean that they “deplore this highly unethical practice, and request your assistance in rectifying the situation before too many individuals are hurt, and before our reputation with the community becomes too damaged.” ²⁹ The dean noted the risks to other residents of Orange County—including the risks of contracting tuberculosis and other communicable diseases—but added that he regarded such consequences

as the least of our problems and one which we perhaps richly deserve in the face of our callous indifference to the real needs of a group of people who correctly perceive that they are welcome in Orange County as long as they don’t get sick or wish to have children. To attack the illegal alien by direct or indirect denial of needed medical care is a wretched act of human injustice. To pass the cost to the state through a mechanism which violates a basic human right of the sick to a reasonable degree of privacy is no less a travesty. ³⁰

The Orange County Board of Supervisors appointed a task force made up of “representatives of business, labor, churches, politics and academic life” to study the question. Their report was blunt. Under the new policy, “the seeking of medical care from public agencies became synonymous with the threat of deportation.” Very quickly, “word spread” “within the tightly-knit community … about the danger of visits to health centers. Many people seemed to have heard of someone who had been deported when they had sought help or of someone who had been summoned by the Immigration authorities after seeking treatment at a health center.” ³¹ Though the local public health department and a few community clinics “considered the county policy unwise or inapplicable to them, and consequently ignored the directive,” the Task Force on Medical Care for Illegal Aliens believed that the new policy made immigrants reluctant to seek health care from any Orange County agency, whether the agency cooperated with the INS or not. From the immigrants’ perspective, they explained, “[t]he only ‘safe’ course of action was to avoid any health service whenever possible.” The task force worried “that the fear generated by the directive” made undocumented parents reluctant to let their American children utilize “these services as well as all others—a state of affairs which, if uncorrected, will be paid for dearly by the Orange County taxpayers.” ³² As predicted, the county documented a significant increase in communicable diseases. “Only a year and a half after the county” adopted the new policies, “medical problems multiplied: extrapulmonary tuberculosis increased 57%; salmonellosis increased 47%; infectious hepatitis increased 14%; rubella increased 53%; syphilis increased 153%” (Flores 1979).

Community advocates demanded that the county and medical center abandon their reporting practices. “For an emotion-packed 2 ½ hours” in March 1978, “Orange County supervisors listened to a parade of speakers tell them that the county’s illegal aliens were being frightened away from needed medical care at the UCI Medical Center because they believed that seeking treatment would lead to their deportation.” The advocates were successful: “When the hearing ended, the board unanimously agreed that county medical services and enforcement of federal immigration laws ‘should be separate and independent of one another.’ The vote provoked loud and sustained applause from the overflow audience, nearly all of whom were there as advocates for the aliens” (Turner 1978, A1).

Even the county acknowledged problems with their policy. The county director of the Medical Services Administration, who initially favored the policy in part because it would shift some county costs to the state, said that “even though … we have worked very hard to encourage aliens to sign the WR-6” form required to verify their status, “we are only getting about 10% of those individuals to apply for Medi-Cal.” The policy was thus not “working out as well as we had hoped.” That was because “people are extremely fearful of signing the WR-6, because it is sent to INS.” To increase state reimbursement, the county now wanted “either legislative or procedural changes which would keep the WR-6 from mandatorily going to INS.” In other words, they wanted to end the state status verification system: “We feel that we would have a much higher incidence of undocumented persons applying for Medi-Cal in those instances, which would then mean the county taxpayer doesn’t pick up the cost” (Public Hearing on Undocumented Persons 1978, 91–93). Orange County, the birthplace of modern conservativism (McGirr 2015), had abandoned its cooperation policy and asked the state to do the same.

A Federal Agency and State’s Reporting Policies Are Rescinded

Unlike California, which, at least under state policy, limited cooperation to status verification, Texas welfare officials were required to report undocumented immigrants to immigration officials for enforcement purposes. And unlike in California where aid was granted pending verification, Texas applicants had to provide proof of legal status before receiving aid. ³³ Thus, Texas had no need to communicate with the INS to verify eligibility for welfare. Nevertheless, Texas adopted AFDC regulations in 1973 providing that, “[w]hen an individual is denied assistance because the facts indicate he was not lawfully admitted to this country,” the INS “must be notified.” The INS referral was even required when an applicant “failed to complete the application process” or when undocumented parents applied for benefits for US citizen children. ³⁴ It is not clear how many individuals were reported to immigration officials, but one INS official in El Paso said that, at his office, “we get hundreds of leads from the Texas department of public welfare about illegals who are trying to get on welfare” (US News and World Report 1977).

The policy endangered the well-being of undocumented immigrants as well as US-born children in mixed-status families. In the border region, officials estimated that 10–15 percent of the caseload was composed of mixed-status families. ³⁵ One boy affected by this policy was a US citizen whose parents were undocumented. The parents had not applied for aid because they knew that they would “be reported to the INS and deported” if they did. The boy had a “bronchial condition which a doctor” had advised would “lead to his death by suffocation if not properly treated.” His mother had “taken him to several doctors, but none” would “treat the child because the family” was not eligible for Medicaid, nor did they have the cash to pay the medical fees. The child also lacked “adequate, healthy housing or nutrition.” Under federal law, the boy was eligible for AFDC, Medicaid, and Food Stamps. However, his parents could not apply on his behalf “for fear” that they would be deported. Without assistance, “this child, a U.S. citizen, and many others” like him might “not survive.” ³⁶

To remedy this situation, two legal aid lawyers—Fred McLeroy and Clyde Farrell—filed two lawsuits against the Texas Department of Public Welfare. ³⁷ One case involved Martha Rodriguez, a US citizen whose food stamps were “cut off when she refused to give information concerning undocumented aliens who were allegedly residing in her house” (A. Smith 1979, 110). Another case, a class action, involved undocumented parents and their US-born children. The lead plaintiff in Doe v. Texas Department of Public Welfare had been living in the United States without authorization “for a number of years.” ³⁸ She had two American-born children whose AFDC and food stamps were “terminated” after a Hidalgo County welfare worker “advised her that she would be reported to the Immigration Service and deported if she renewed her application for the children’s benefits.” Without assistance, her family was forced to rely on her meager $60 monthly income. ³⁹ “The only way the children can exercise their rights to AFDC or food-stamp benefits when they are eligible,” Farrell explained, “is through an application by their caretaker. Texas officials have put them in a cruel ‘Catch-22’ by using this application process to cause deportation of the caretaker.” ⁴⁰ The reporting policy not only violated federal law, McLeroy argued, it also “works against one particular class of Texas citizens, Mexican-Americans.” “Many laws enforcing restrictions on immigrants from Mexico,” he explained, “have the effect of terrorizing Mexican-Americans who are legally in Texas. Only Mexican-Americans must display their papers in border areas. Only Mexican-Americans are questioned and intimidated about citizenship at check points or by roving patrols. Only Mexican-Americans will be discouraged from applying for food stamps and welfare.” ⁴¹

The lawyers claimed that Texas policy violated both federal and Texas statutes, which still guaranteed some measure of confidentiality. ⁴² They explained that the confidentiality provisions in the Social Security Act were there to ensure individuals need not fear applying for benefits: “Without the assurance that the information provided will not be used against the applicant, many deserving persons will be afraid to apply for welfare.” ⁴³ “Shortly after the Rodriguez case was filed,” the Texas Department of Public Welfare “moved to join HEW and USDA” as co-defendants in the case. “HEW responded to the Motion for Joinder with the observation that” Texas’s reporting policy was illegal. ⁴⁴ Though amendments to the Social Security Act in 1972 and 1974 appeared to allow cooperation between welfare and immigration officials, and HEW had approved Texas’s state plan, which required reporting undocumented immigrants to the INS, some federal welfare officials were skeptical that these amendments allowed such cooperation. The acting administrator for the Social and Rehabilitation Service acknowledged as early as 1975 that the 1974 amendments “permits States to inform INS of alien status when so requested. … But … welfare agencies should not be expected to carry out responsibilities of INS. Hence, it may not be proper for public assistance agencies to initiate individual care reports to INS.” ⁴⁵ In other words, welfare agencies should cooperate with law enforcement when asked about a specific individual, but they should not engage in broad reporting practices.

By 1977, the leadership at HEW had changed as well. After Jimmy Carter was elected president in 1976, he appointed Joseph Califano to head HEW. Under this new leadership and pressed by legal aid lawyers to reevaluate their position, HEW now said that they believed that Texas’s reporting mandate violated the Social Security Act “and was therefore improperly approved.” ⁴⁶ HEW was now planning to file an amicus brief in the Doe case, taking the position that Texas’s reporting policy was illegal. ⁴⁷ The USDA, however, disagreed with HEW and sided with the defendants, “seeking to uphold the Texas policy.” ⁴⁸ HEW came to believe that this reporting requirement conflicted with other federal regulations. Specifically, one federal regulation required that “[e]ach State agency shall restrict the use of disclosure of information obtained from applicant households to persons directly connected with the administration or enforcement of the provisions of the Food Stamp Act.” Apparently, “their General Counsel’s office did not, … clear the conflicting Handbook provision” and was in fact unaware of the regulation at all.” ⁴⁹

Farrell and McLeroy wrote to the White House in 1977 for assistance in convincing the USDA to follow HEW’s lead. ⁵⁰ “We have the sad spectacle of two federal administrative departments litigating against each other over the application of statutes and regulations whose meaning is plain,” Farrell wrote: “Counting the two Legal Aid Associations now representing plaintiffs in these cases, we have the legal resources of five government agencies involved in litigating in two federal district courts. Given the critical importance of the states’ unlawful policy to the most impoverished of our clients, every Legal Aid Association in this state should be litigating this question.” ⁵¹ McLeroy warned the White House that “a group of lawyers in California has been inquiring into the status of the Rodriguez case and appears to be considering a lawsuit along its lines.” ⁵² The USDA could avoid further litigation if they changed their policy.

The lawyers’ efforts were successful. As a result of the consent decree in Rodriguez v. Texas Department of Public Welfare, the USDA eliminated its reporting requirement from the Food Stamp handbook (Foreman 1981). ⁵³ In turn, both the USDA and HEW told Texas officials to abolish their reporting policy. ⁵⁴ By November 1977, the Texas Department of Public Welfare had complied; “illegal aliens identified by AFDC, Food Stamp, or Medicaid staff” would no longer “be referred to INS.” ⁵⁵ In addition to amending the food stamp handbook, the USDA also clarified its position on reporting in its amendments to the Food Stamp Act in 1977. ⁵⁶ The new federal law specified “that a State agency shall not use or disclose information obtained from applicant households to persons not directly connected with the administration or enforcement of the Food Stamp Act or its regulations.” In addition, it reiterated that “[c]urrent regulations prohibit contact with INS by certification personnel without written consent by the household” (Foreman 1981, 4645).

While direct reporting of food stamp applicants was no longer permitted, state welfare agencies were still permitted “to contact INS for verification of alien status if other verification is not available, or if the reported status appears questionable.” ⁵⁷ And the INS was still free to use the verification process for enforcement purposes. While Texas had no need for status verification (since they required proof of status before benefit distribution), status verification was still mandated in California. At a public hearing in Los Angeles in 1977, immigrant advocates praised the change in federal food stamp policy but pressed for more safeguards to protect undocumented immigrants and their children, urging “that the state legislature mandate that no social or state agency shall act as a reporting arm of I.N.S.” (Public Hearing on Undocumented Persons 1978, 125). While the state continued to engage in status verification, advocates were telling immigrants by 1980 that status verification “doesn’t usually lead to deportation,” though they added that it “could.” ⁵⁸

In 1981, after two years of impassioned debate, the Los Angeles County Board of Supervisors (LACBS) adopted a proposal requiring that indigent individuals apply for Medi-Cal as a condition of nonemergency medical treatment at any county health facility. Per state policy, information about non-citizens who applied for Medi-Cal would be forwarded to the INS for status verification. County officials stressed that undocumented immigrants would continue to receive emergency medical assistance as well as free treatment for communicable diseases, regardless of their willingness to apply for Medi-Cal. ⁵⁹ Despite seemingly strong public support for the measure, it was never implemented. Legal advocates, county health workers, labor unions, religious groups, and other community organizations organized the County Health Alliance, and they convinced a judge to issue a temporary restraining order (Oliver 1981a, 1981b). In April 1985, the LACBS “ratified without discussion or dissent the settlement” of the County Health Alliance lawsuit. Under the terms of the settlement, the county agreed “not to try to require undocumented aliens to complete status verification forms as a condition of being treated at hospitals or clinics.” The county would continue to provide (emergency and) nonemergency care “without inquiring into citizenship status” and without contact with the INS. ⁶⁰

RETHINKING SANCTUARY

By the time the sanctuary movement began to gain strength and the first cities adopted sanctuary ordinances, some communities, welfare agencies, and hospitals had already begun to abandon efforts to cooperate with the INS for enforcement purposes without any prodding from the sanctuary movement. Rather, out of concern for undocumented immigrants, their children, and the broader community—not simply Central American migrants—immigrant rights activists, medical professionals, and legal aid attorneys pushed local and federal welfare officials to adopt limited or non-cooperation policies. Federal welfare officials not only abandoned their own cooperation policies, but, in some cases, they also pressured local communities to abandon theirs. Indeed, by the early 1980s, the Department of Health and Human Services (formerly HEW) was discouraging hospitals from reporting undocumented immigrants to immigration officials for enforcement purposes. ⁶¹ They clarified that they considered undocumented immigrants to be “‘persons’ protected” by the “community service obligation” under the Hospital Survey and Construction Act, commonly referred to as the Hill-Burton Act, to provide charity care. ⁶² Federal officials also pressured a hospital in Yuma, Arizona, to stop reporting patients to the INS. The action came after Community Legal Services filed a complaint with the Department of Health and Human Services Office of Civil Rights. In a 1982 settlement, the hospital “agreed to stop asking potential patients for proof of their legal immigration status, to stop reporting persons who did not provide satisfactory documentation to INS and the Border Patrol, and to remove signs the hospital had posted stating that it cooperated with the Border Patrol” (Drake 1986, 508; see also Anonymous 1983).

Even when the federal government continued to encourage limited cooperation, states often chose the narrowest interpretation available. The US Congress (1980, 134–38; see also Foreman 1981) approved amendments to the Food Stamp Act in 1980 that mandated reporting of undocumented immigrants to the INS in certain circumstances. ⁶³ At the urging of immigrant rights and anti-poverty groups, officials in Texas and California decided that, while local “workers must report to the INS any illegal alien household members” to comply with federal law, the states would choose to define “illegal aliens” very narrowly as “aliens who have received final orders of deportation” (Johnston 1983, 2081–82; McKinsey 1984). Other states, activists suggested, did the same (Wheeler and Leventhal 1986, 916).

While the majority of sanctuary ordinances adopted in the 1980s were no doubt inspired by the sanctuary movement and the plight of Central American migrants (Colbern 2017), in practice, the limited or non-cooperation policies encoded in many of these ordinances applied to all undocumented immigrants whatever their origins. Chicago’s 1985 executive order held that “[n]o agent or agency shall request information about or otherwise investigate or assist in the investigation of the citizenship or residency status of any person unless such inquiry or investigation is required by statute, ordinance, federal regulation or court decision.” In addition, it stipulated that “[n]o agent or agency shall disseminate information regarding the citizenship or residency status of any person unless required to do so by legal process.” ⁶⁴ Indeed, Chicago’s executive order made no mention of “sanctuary” or Central Americans at all.

Even when these ordinances declared their cities “sanctuaries” and specifically named Central Americans as the targets of these policies—as most of these resolutions did—local officials were likely open to including non-cooperation provisions as part of these ordinances because they had recently witnessed the problems with, and debates about, cooperation or had recently adopted or reaffirmed their commitment to non-cooperation. ⁶⁵ Los Angeles’s resolution was adopted six months after the county ratified the County Health Alliance settlement that limited cooperation between non-emergency health providers and immigration officials and six years after the city issued Special Order 40 that limited cooperation between the LAPD and immigration officials. ⁶⁶ Los Angeles city councilman Michael Woo affirmed that the city’s sanctuary resolution was built in part from these earlier efforts. The resolution, he maintained, “reaffirms a Police Department policy not to arrest or detain undocumented immigrants merely for being in this country illegally” (see also Kramer 2020). ⁶⁷

Indeed, New York’s “sanctuary” ordinance is perhaps best understood as a non-cooperation order. In October 1985, Mayor Edward Koch issued a memo to all city agency heads telling them that they should not cooperate with immigration officials. Koch explained that it was already the “current policy of most city agencies not to report aliens to immigration authorities unless the alien has given signed permission for a status check or the alien appears to be engaged in some kind of criminal behavior.” The purpose of the 1985 memo was “to reaffirm this as city-wide policy,” stressing that this was “not a new but a more comprehensive city policy on undocumented city aliens.” ⁶⁸ While some local officials “said the memo was a response to public hearings that involved discussions over the plight of refugees and questions about the sanctuary movement,” a spokesman for Koch said that the memo “was not tied to the sanctuary issue” (qtd. in Merina 1985). Unlike Los Angeles’s sanctuary ordinance, New York’s memo made no mention of Central Americans or sanctuary. Rather, it framed the issue more broadly: “New York City is home … to somewhere between 400,000 and 750,000 undocumented aliens.” ⁶⁹ The mayor stressed that undocumented immigrants’ presence “is not a New York City crime” (qtd. in Los Angeles Times 1985). And “he said he did not want to turn every city clerk into a supercop preoccupied with tracking down illegal immigrants” (qtd. in Schmalz 1985).

Like other cities during the 1970s, New York had had its own experiments in cooperation. In 1974, New York state passed legislation stipulating that applicants or recipients of AFDC, Medicaid, and general assistance who were “unlawfully residing in the United States” must be referred to the INS or nearest consulate so they might “take appropriate action or furnish assistance.” ⁷⁰ State welfare officials subsequently stipulated that “the immigration service should be notified immediately of any alien who is unable to verify his legal residency or who presents documents of ‘questionable validity’” (Farber 1974). Around the same time, Bellevue, a large public hospital in Manhattan, adopted a policy of encouraging all indigent patients to apply for Medicaid. By state law, “supplying the necessary documentation” to apply for Medicaid “could result in deportation” of non-citizens without legal status (Flores 1979, 38–39). As Elizabeth Bogen (1987, 175), director of New York City’s Office of Immigrant Affairs, explained, “[a]pplicants were not always informed that this INS check would be made.” As a result, Bellevue “developed the reputation, especially in the Chinese community, as the agency turning names in to the INS.” Hospital administrators later noticed that “many persons stopped seeking care at the hospital” because they viewed “Bellevue as threatening.” Eventually, the hospital reconsidered its policy and began to discourage undocumented immigrants from applying for Medicaid (Flores 1979, 2; Dallek 1980, 407–14). Over time, the city concluded that they should seek funds from the federal government to cover the cost of services to undocumented immigrants rather than try to cut costs on the backs of immigrants (Buder 1986). ⁷¹ In 1979, Koch told a reporter that the city had “no choice but to render services” to all immigrants “on a humanitarian basis.” “New York provides services for whoever needs them,” he said. “We have no way of knowing if we’re helping illegal aliens or not” (qtd. in Blum 1979).

When Koch issued his non-cooperation order in 1985, he explained that, “[f]or the most part, these aliens are self-supporting, law-abiding residents. The greatest problem they pose to the city is their tendency to underuse services to which they are entitled and on which their well-being and the city’s well-being depend.” Specifically, the mayor worried that “victims of crime, consumer fraud or workplace safety violations may decide not to report their victimization for fear” that they would be reported immigration authorities. “Persons who need medical care may decide not to seek it, some families may keep their children out of school, and adults may fail to avail themselves of ESL classes. … It is to the disadvantage of all who live in the City if some of its residents are uneducated, inadequately protected from crime, or untreated for illness.” He continued: “Undocumented aliens should not be discouraged from making use of those city services to which they are entitled; on the contrary, for the public weal, they should be encouraged to do so.” ⁷² Few cities went as far as New York in encouraging undocumented immigrants to access city services. ⁷³ And federal laws still barred undocumented immigrants from most federal programs. At the same time, by the mid-1980s, the experiments in cooperation that were so popular in the 1970s had largely faded from view.

An attorney for the National Center for Immigrant Rights explained that by 1986 even the status verification procedure was generally no longer “being used as an enforcement mechanism by INS officials” (Wheeler and Leventhal 1986, 916, 923). The separation of status verification from immigration enforcement was facilitated by a new federal law, which expanded the practice of verification nationwide. With the passage of the Immigration Reform and Control Act (IRCA) in 1986, the federal government mandated that welfare agencies verify the status of applicants for federal benefits through the Systematic Alien Verification for Entitlements (SAVE) program (US General Accounting Office 1989). ⁷⁴ SAVE was a computerized database operated by the INS that allowed state and local agencies to verify eligibility in a matter of seconds rather than months. Critics charged that due to the INS’s shoddy record-keeping system, eligible individuals were denied aid. ⁷⁵ The IRCA, however, specified that SAVE was to be “used only to verify eligibility for benefits and cannot be used to initiate deportation or removal proceedings (with exceptions for criminal violations)” (Broder and Blazer 2011). ⁷⁶ Thus, there was “an explicit federal policy of not using information obtained by health and social service officials to enforce civil immigration law” (P. Smith 1995, 160).

CONCLUSION

By the mid-1980s, when dozens of cities began to issue sanctuary ordinances, many government officials had already concluded that cooperation between welfare, health, and immigration officials was unwise, ineffective, counterproductive, and, in some cases, against federal law. While some scholars suggest that cities now insert limited cooperation policies within broader policies that guarantee confidentiality to shield themselves from legal challenges against sanctuary laws (Mitnik and Halpern-Finnerty 2010, 54–55), I have demonstrated that non-cooperation in welfare agencies emerged from those broad confidentiality policies in the first place. Specifically, confidentiality provisions adopted by federal welfare officials and encoded in the Social Security Act largely prevented cooperation between welfare and immigration officials from the mid-1930s through the early 1970s.

While the adoption of sanctuary city ordinances in the 1980s were overwhelmingly inspired by desires to protect Central American migrants, perhaps not all of them were (for example, New York) (Colbern 2017). Five cities and one state made no mention of Central Americans in their resolutions (Chicago, New York, Rochester, San Jose, San Diego, and Massachusetts), while others also mentioned Haitians (Cambridge, Somerville, Oakland, Detroit), Brazilians (Somerville), South Africans (Oakland), and the Irish (Somerville). More to the point, the form that many of these sanctuary ordinances took (that is, limited or non-cooperation policies) were policy tools that local officials had already deployed (or were imposed upon them), often inspired by broader concerns about undocumented immigrants and the negative consequences of cooperation for them and for the broader community. ⁷⁷ As such, I challenge the contention that it was only over time that “city sanctuary became more explicitly connected to debates about information sharing and cooperation between various institutions of government” and that cities started to defend “their policies through appeals to the necessities of municipal governance, and the health and safety of residents” (Ridgley 2012, 228; see also Colbern 2017, 193). In cities where undocumented immigrants were numerous such as New York and Los Angeles, these debates about cooperation predated the sanctuary movement by a decade. Non-cooperation policies were also incorporated into state policies in places such as Texas where no sanctuary ordinances passed. What is more, at least ten of the non-cooperation policies imbedded in sanctuary ordinances were written to apply broadly to all undocumented immigrants (Los Angeles, Sacramento, Burlington, Madison, Chicago, Takoma Park, New York, Massachusetts, Seattle, and Rochester), even in cases where these ordinances were clearly inspired by concerns about Central Americans.

This article demonstrates that non-cooperation policies emerged not out of a desire to protect Central Americans who were improperly denied asylum. Rather, the origins of non-cooperation involved concerns about undocumented migrants (many of whom were Mexican) and their family members, a broader critique about privacy and confidentiality that concerned all social service applicants and about the propriety of institutions devoted to service and well-being taking on the functions of law enforcement agencies. Indeed, by conceptually and empirically untangling sanctuary ordinances and non-cooperation policies, one can see that limitations on cooperation between welfare, health, and immigration officials are not simply the products of sanctuary ordinances or local communities. Rather, limits on cooperation can also be found in federal welfare law (US Congress 1939a), federal privacy law (the Privacy Act in 1974, the Family Educational Rights and Privacy Act, and the Health Insurance Portability and Accountability Act), ⁷⁸ and federal immigration law (the IRCA in 1986). As such, non-cooperation policies need not always represent local resistance to federal power. In some circumstances, such policies may even represent federal resistance to local power.

To be sure, some non-cooperation policies do represent local resistance to federal power, perhaps especially in the criminal legal system (Lasch et al 2018). What is more, the public declaration of sanctuary or non-cooperation can also represent local resistance to federal power (Kramer 2020). Local, state, and federal officials recognized the difference between the content and the more symbolic elements of sanctuary. Harold Ezell, a western regional INS commissioner, asserted “that the declaration of Los Angeles as a sanctuary, regardless of the resolution’s purpose or content, would be received as an open invitation by the rest of the world to seek a better way of life in Los Angeles as a safe haven from the INS.” ⁷⁹ Eventually, critics pressured the Los Angeles city council to rescind the resolution and pass a compromise instead, “omitting the word ‘sanctuary’ but retaining most of the declaration, including the stipulation that law enforcement personnel not report undocumented immigrants to the INS” (Colbern 2017, 188; see also Kramer 2020). Immigration officials also objected to New York’s executive order publicly mandating non-cooperation, which did little to change actual policies and practices on the ground.

In line with scholarship on the expressive (Sunstein 1996) and symbolic (Edelman [1964] 1974; Calavita 1983) functions of law and the interpretive consequences of social policies (Campbell 2012), this broader history reaffirms the symbolic significance of declaring oneself a “sanctuary city” or of declaring a policy of non-cooperation, above and beyond the policy and practice of non-cooperation. ⁸⁰ At the same time, the policy and practice of non-cooperation had important material consequences. Non-cooperation allowed undocumented parents to apply for benefits for their US-citizen children and allowed undocumented individuals to make use of vital health services.

This article also reaffirms the significance of the fragmented nature of federal institutions for facilitating the promotion of immigrant rights (Wells 2004). While federal immigration officials often called for greater cooperation from welfare and health agencies, federal welfare officials were often more hesitant about taking on the functions of law enforcement agencies (Fox 2019). This was a point that President Gerald Ford’s Domestic Council Committee on Illegal Aliens noted in their preliminary analysis of failed efforts to increase cooperation between the Social Security Administration and the INS. Headed by the Attorney General, and composed of various relevant cabinet members, they concluded that “[f]uture policymakers should be extremely wary of drawing agencies whose central purpose is not enforcement into arrangements which require them to act in ways contrary to their historical nature. The SSA has always seen itself as a service agency, and thus has been slow to adjust to” the new efforts to increase cooperation between SSA and INS. ⁸¹

It was not until 1996, with the passage of the Personal Responsibility and Work Opportunity Reconciliation Act and the Illegal Immigration Reform and Immigrant Responsibility Act, that the federal government passed laws to limit the spread of non-cooperation policies (Pham 2006, 1384). ⁸² These laws also required greater cooperation between immigration authorities and welfare or local law enforcement officials (Committee on Ways and Means 1998, Appendix J, 1408). While cooperation between police and immigration authorities increased significantly after the 9/11 attacks with the rise of 287(g) agreements and later the Secure Communities program (Armenta 2017), cooperation between hospitals, welfare agencies, and immigration authorities has remained limited. The reporting requirement for welfare agencies is “quite narrow in scope” (Broder, Moussavian, and Blazer 2015, 9). And federal officials have issued guidance to states recommending that they limit the information collected from non-applicant family members. Cooperation with hospitals is even more circumscribed: “[T]here is no federal reporting requirement in health programs” (10). In 2013, the Department of Homeland Security issued a memo “confirming that information submitted by applicants or family members seeking Medicaid, CHIP, or health care coverage under the Affordable Care Act would not be used for civil immigration enforcement purposes” (10). Immigration authorities have also issued memos indicating that “immigration enforcement actions are to be avoided at sensitive locations, including at hospitals and other health care facilities” (National Immigration Law Center 2017, 2).

These sorts of differences in expectations of, and provisions for, cooperation in immigration enforcement across institutional spheres (including welfare, health, and law enforcement agencies) suggests that scholars should not only distinguish between non-cooperation policies and sanctuary ordinances, as I have done here, but also among both non-cooperation and even sanctuary policies as well. Scholars should examine the origins of non-cooperation policies for local law enforcement agencies. More work is also needed on the varied politics that inspired the adoption of early sanctuary ordinances as well as the relationship between sanctuary activists and the broader immigrant rights community. When we do, we might further reconsider our assumptions of the origins, targets, consequences, and significance of early “sanctuary” policies.