Making noncitizens ’ rights real: Evidence from immigration scam complaints

Noncitizens seeking to make sense of US immigration systems encounter a labyrinth of information and deception. This paper is the first national study of scams targeting noncitizens seeking immigration legal services. I construct a county-year database ( N = 3135 over a four-year time period, 2011 – 2014) across secondary data sources to analyze the correlates of immigration scam complaints submitted to the Federal Trade Commission (FTC). I find that welcoming counties have more immigration scam complaints, while counties with exclusionary contexts tend to have fewer complaints. The results do not suggest that scams are more prevalent in welcoming contexts, because the actual number of scams is unknown. Instead, we can conclude that noncitizens tend to come forward to report immigration scams in welcoming contexts of reception, even after accounting for exclusionary policies. A robust safety net proved the most reliable predictor of immigration scams reported to the FTC. The concentration of immigration attorneys, legal aid services, and language access was also positively associated with the number of FTC scam reports. Taken together, these results suggest that immigrant-serving capacity and access to key services support noncitizens who report immigration scams, while hostility toward immigrants may deter them from exercising those same rights.

implemented initiatives that are either more welcoming toward immigrants or less so (García, 2019;Prieto, 2018;Steil & Vasi, 2014;Valdivia, 2019).Past research has documented the role that local contexts of reception-whether welcoming or not-can have on immigrant reports of employmentrelated crime and discrimination (Fussell, 2011;Gleeson, 2009Gleeson, , 2015) ) and domestic violence (Amuedo-Dorantes & Arenas-Arroyo, 2021; Muchow & Amuedo-Dorantes, 2020).In this study, I investigate where noncitizens i are willing to denounce immigration scams by examining all county contexts across the United States.When noncitizens make claims to justice by reporting immigration scams, are immigration scams reported to the Federal Trade Commission (FTC) positively associated with welcoming contexts of reception, even after accounting for immigration restrictionism?
To answer this question, I focus on immigration scam reports submitted to the FTC.To be clear, no data capture the actual number of scams.Absent such a prevalence measure, I examine where noncitizens willingly denounce scams.I am interested in whether restrictionism offsets an expected positive association between FTC scam reports and welcoming contexts of reception.On the welcoming end of the spectrum, noncitizens' claims to justice should coincide with local webs of support-namely meaningful access to the publicly funded safety net as well as access to justice facilitated by social helpers.By contrast, scam reports should be less common in restrictionist counties (i.e., counties with high deportation rates under the Secure Communities immigration enforcement program and counties in states with policies criminalizing immigrants).
Contributing to existing research on the relationship between contextual factors and where vulnerable groups denounce under-reported crimes (Gleeson, 2015;McVeigh et al., 2003), I find that reports of immigration scams are more common in welcoming contexts.In addition, measures of welcoming contexts remain reliable predictors of scam reporting to the FTC, net of restrictionism and other factors.Counties that deliver safety net services to poor Hispanics ii most consistently account for the difference between (a) higher-than-expected reports of immigration scams in more welcoming counties and (b) the scant reporting of such crimes in otherwise comparable but less welcoming counties.The positive relationship between the safety net and FTC scam reports remains consistent across a series of regression models (both cross-sectional and fixed effects) and robustness checks.
I also discuss the implications of these results for research on immigrant rights-claiming and civic inequality (Bloemraad, 2018;Bloemraad et al., 2020;Leitner & Strunk, 2014a, 2014b).With respect to the question of where noncitizens make claims to justice, welcoming contexts of reception provide clues as to where immigrants choose to exercise their rights.Denouncing scams amounts to civic action and reveals where noncitizens have spoken up to "stretch the boundaries of liberal democratic citizenship" by enacting a form of insurgent citizenship (Leitner & Strunk, 2014b, p. 354).Far from being evenly distributed across the country, these actions are most common wherever welcoming webs of support abound.For instance, when street-level bureaucrats deliver safety net services, they reach immigrant communities (Marrow, 2009), and such access to mainstream institutions may translate into knowledge about and trust in public institutions (Cook, 2001) like the FTC.Given the mismatch between the current location of the nation's noncitizen population and the uneven concentration of welcoming organizations (de Graauw et al., 2013)-as well as experienced lawyers (Sandefur, 2008;Sandefur & Smyth, 2011) and efforts to reduce language barriers in law enforcement contexts (Williams, 2015)-the publicly funded safety net may represent the lone reliable source of support for immigrants living outside established immigrant destinations.Absent these webs of support, we observe fewer people willing to file reports with the FTC.When noncitizens seek protections as consumers in welcoming contexts, doing so can help bring people into the fold of mainstream institutions.Currently, if noncitizens' rights hinge on where they live, then uneven contexts of reception could create relative winners and losers among noncitizens and exacerbate civic inequality.

| BACKGROUND ON IMMIGRATION LEGAL SERVICES CRIME
Immigration legal services crimes are not new.Accounts of legal services fraud appear throughout ethnographic accounts of immigrant communities (Coutin, 2003;Hagan, 1994;Mahler, 1995;Menjívar, 2000).The enactment of the 1986 Immigration Reform and Control Act (IRCA) resulted in a surge in demand for legal services.Given the large number of immigrants looking to adjust their legal status, trained legal professionals could not meet the rising demand.Community organizations, "lay lawyers," and notaries each played a pivotal role in helping immigrants prepare for IRCA (Coutin, 2003;Hagan, 1994;Longazel, 2018).
The rise of notaries and the attendant demand for legal services has downsides.Although notaries provide valuable services, the line between notary services and the unauthorized practice of law can blur.The most notorious schemes involve notarios (Spanish for "notaries"), who take advantage of immigrant clients by promising naturalization and work permits.Clients unfamiliar with US law typically assume that notarios know best, in part because the word "notario" is a false cognate commonly understood in Latin America as a synonym for a seasoned lawyer (Cohen et al., 2013;Longazel, 2018).
Over time, notary and related services have become increasingly elaborate.In the 1990s, multipurpose agencies offering a range of legal, remittance, tax, and financial services emerged.Unsuspecting clients seeking legal advice from lawyers could end up consulting with untrained or misleading service providers.As Mahler (1995) observed, "Itinerant translators and self-professed 'lawyers' travel from town to town like nineteenth century quacks selling home-made remedies" (pp.141-142).Similarly, there have been cases in which notaries have "charged exorbitant sums to fill out applications for political asylum or Temporary Protected Status," as well as "cases in which these 'legal workers' have disappeared with money that people have paid in advance for their services" (Menjívar, 2000).Survey data from the mid-1990s suggest that sources of legal assistance varied, and unauthorized immigrants and asylum seekers (unlike US citizens, permanent residents, and refugees) were more likely to rely on notarios than on private attorneys (Bach, 1993).A qualitative study of fraud targeting immigrants noted the persistence of notario fraud (Marlowe & Atiles, 2005).By the 2000s, half of immigrant respondents had adjusted their status with help from lawyers or other legal practitioners (Jasso et al., 2006).Immigrants without access to legal services tend to rely on family members to navigate legal issues (Delgado, 2020).
Data on the prevalence of immigration scams do not exist.However, we do know where immigrants have willingly reported scams to the FTC.Since 2011, the FTC began tracking incidents such as notario scams and related crimes.Anyone can submit a complaint-either in English or Spanish-via an online "FTC Complaint Assistant" or a toll-free Consumer Response Center.In order to spread the word to consumers, the FTC conducts media outreach and holds meetings with legal services partners.Targets of scams, family members, and intermediaries reporting an incident on behalf of a crime victim can submit complaints.When victims of crime come forward to report scams, they do so in order to prevent others from becoming scam victims.For example, a complainant submitted the following to the FTC (translated from Spanish): A friend recommended them [the scammer] to me because they supposedly had contacts with immigration services, and they could help me obtain a work visa, and also speed up a friend's [immigration] case, who was also scammed, and another friend of mine, too, and they keep doing the same thing and someone needs to put a stop to this, since they keep doing the same thing to other people.I have another acquaintance who also lost $5,000-please someone do something so they don't keep scamming people.
The FTC then maintains a record of each complaint and shares scam reports with state and local law enforcement to identify recurring scams.When reporting a scam to the FTC, complainants often noted that they lost income (median: $300-500).In cases where complainants reported losing thousands of dollars, the amount lost was much larger (median: $3000-4000).In select cases, the FTC helps customers to recoup lost expenses.For example, following litigation brought by the FTC, a court ordered a firm to repay $616,000 to customers and to cease providing immigration services.As reflected in FTC claims data, immigration scams also siphon immigrants' resources and social capital, further exposing those at risk of deportation (Hagan et al., 2015;Valdivia, 2019).
I analyze these FTC complaints to examine where various kinds of immigration scams come to light.Based on complaints submitted to the FTC, reported immigration scams come in many forms.In addition to notario scams and instances where lawyers mishandle an immigration case, noncitizens report telemarketers impersonating immigration authorities and demanding payment under threat of deportation.Fake enforcement (or citizenship services) agents call people at home and sometimes secure credit card information or electronic payments from scam victims who believe that they are speaking with government officials.Online schemes include fake and misleading organizations selling green card lotteries and misrepresenting services as immigration benefits.Often, scammers create fake websites designed to resemble either federal government agencies or legitimate immigration service providers.Such scams target noncitizens who lack knowledge of immigration law, fear the threat of deportation, are willing to pay for schemes that later turn out to be "too good to be true," or rely on incompetent or misleading advice regarding their immigration situation.

| Contexts of reception
Contexts of reception help shape immigrant and noncitizen integration and can make the difference between welcoming versus exclusionary responses to new arrivals.Whether and how noncitizens exercise their rights depends on a dynamic set of constraints (Bloemraad, 2021;Gleeson, 2010), within which noncitizens can make resonant claims based on their "structured agency" (Bloemraad, 2018).I explore meso-level pathways across welcoming contexts of reception that may provide "access to receiving society institutions to resolve issues," as Martínez-Schuldt (2020) has theorized, as well as whether rights-claiming actions in otherwise welcoming locations can withstand hostilities toward immigrants codified in restrictionist immigration initiatives.I leverage variation in welcoming contexts across multiple domains: the social safety net, language services, community organizations, and immigration lawyers.Such "assemblages" (Leitner & Strunk, 2014a) of entities and actors have their own agendas and interests, but they nevertheless operate as webs of support conducive to immigrant rights-claiming actions.Scam reports thus reveal the extent to which various webs of support predict acts of "insurgent citizenship" (Leitner & Strunk, 2014b) in the form of noncitizen rights-claiming.
Variation in contexts of reception can stratify immigrants' interactions with US institutions.Practically speaking, relatively welcoming public service agencies deliver tangible benefits and demonstrate how US bureaucracies can be accessible and trustworthy sources of support.For instance, public entities can shape immigrants' context of reception via access to the safety net (Fortuny & Chaudry, 2012;Gelatt et al., 2014;Marrow, 2009) and municipal resources ( de Graauw, 2016).Transnational entities such as consulates can also soften the landing by providing assistance, especially where unauthorized and recent Mexican immigrants settle (Martínez-Schuldt, 2020).Welcoming contexts can likewise lower barriers to interactions between noncitizens and mainstream institutions by opening lines of communication.For example, law enforcement agencies can attempt to address language barriers in crime reporting via outreach in immigrant communities (Provine et al., 2016;Williams, 2015).
Furthermore, for those seeking legal advice and assistance regarding immigration or immigration scams, webs of support also include community-based organizations and experienced immigration lawyers.Welcoming contexts, where such entities are most common, can facilitate access to justice.For example, immigrant-serving organizations facilitated naturalization applications in Houston following federal immigration reform in the 1980s (Hagan, 1994).Decades later, organizations also assisted immigrants applying for the Deferred Action for Childhood Arrivals (DACA) program and its immigration benefits (Wong & García, 2016).The concentration of such organizations may lighten the burden on Mexican consular offices tasked with serving Mexican immigrants (Martínez-Schuldt, 2020).Recent evidence also suggests that deportations are lower in counties with nonprofit capacity to assist with immigration cases (Chand et al., 2020).Similarly, Gleeson (2015) highlights the role of labor unions and state Equal Employment and Opportunity Commissions in identifying cases of wage theft and workplace discrimination against immigrants.These findings echo related work on an influx of legal aid funding dating back to the 1960s, which helped individuals report previously underreported crimes and lowered riot propensities (Cunningham, 2016;Cunningham & Gillezeau, 2018).Similarly, places with a history of civil rights organizing were shown to outperform other parts of the country in hate crime reporting (McVeigh et al., 2003).
Experienced immigration lawyers can also play a role in integrating and orienting noncitizen clients by shaping noncitizens' awareness of US laws and institutions (Menjívar & Lakhani, 2016), and can thus alter their "legal-spatial consciousness" amidst a shifting immigration policy landscape (Flores et al., 2019).However, as immigration law has become increasingly complex and noncitizens have moved to new US destinations, the number of available legal service providers has lagged behind demand (de Leon & Roach, 2013;Legal Services Corporation, 2017;Sandefur & Smyth, 2011).In fact, although access to legal representation can make a difference for noncitizen detainees' outcomes in US immigration custody (Ryo, 2018;Ryo & Peacock, 2018, 2019), experienced immigration attorneys are mostly located in established immigrant destinations.

| Immigration enforcement inhibits a welcoming context
The reputation of US public sector entities among noncitizens must also contend with the proliferation of a patchwork of immigration policymaking (Varsanyi et al. 2012) at the state and local level via laws, ordinances, and administrative partnerships between immigration authorities and other subnational authorities (Armenta, 2012(Armenta, , 2017;;Coleman, 2012;García, 2014;Pedroza, 2019;Steil & Vasi, 2014).As restrictionism has increased, the negative consequences of deportations and noncitizen criminalization have followed (Vargas et al., 2019), including mistrust of mainstream institutions (Cruz Nichols et al., 2018).This mistrust extends to local law enforcement (Menjívar et al., 2018;Nguyen & Gill, 2015;Rocha et al., 2015;Theodore & Habans, 2016;Wong et al., 2021), likely due to cynicism regarding police and the legal system (Kirk et al., 2012;Ryo, 2016) and fear of deportation among, for instance, victims of wage theft (Fussell, 2011) and immigration scams (Marlowe & Atiles, 2005).Notably, the erosion of trust associated with restrictionism may not be insurmountable: local-level public servants may help clients overcome the fear of deportation if, for example, community members believe they will be treated fairly (Messing et al., 2015).

| Scam reports across contexts of reception
As the first empirical study of immigration legal services crime, this paper highlights where claims to justice come to the attention of the federal government after noncitizens willingly report scams to the FTC.When choosing to make claims to justice with the FTC, immigration scam victims vary with respect to the possible webs of support at their disposal, depending on where they live.Examining scam reports as a case of making claims to justice, I extend what we know about the relationship between divergent contexts of reception and immigrant rights-claiming.Following research on the role of contextual factors as a predictor of rights-claiming activity (Gleeson, 2015;McVeigh et al., 2003), I measure variation across contexts of reception to analyze where immigration scams are coming to light in the form of complaints filed with the FTC, an entity tasked with helping consumers who have lost (or were at risk of losing) income to unscrupulous actors.The analyses below examine whether we observe reports to the FTC in welcoming counties more than in other contexts, net of the association between immigration restrictionism and scam reports.

| EXPECTED ASSOCIATIONS BETWEEN FTC REPORTS AND MAIN INDEPENDENT VARIABLES
Based on prior research, I expect more scams to come to light in welcoming contexts of reception.I expect that FTC reports will be positively associated with (a) access to the safety net; (b) legal aid capacity; (c) immigration lawyer capacity; and (d) language services capacity via local law enforcement.Conversely, immigration restrictionism should predict a fewer-than-expected number of such reports.A positive relationship between welcoming contexts and FTC complaints-and a corresponding negative relationship between restrictionism and FTC complaints-would echo existing research on how social helpers tend to encourage noncitizens to claim their rights.I am interested in whether restrictionist contexts offset these positive relationships.

| DATA
In this paper, I track immigration scam complaints known to the FTC.I leverage these data to examine where noncitizens come forward to claim rights by comparing the volume of complaints across contexts of reception.Table 1 details data sources and descriptive statistics for the variables described below.

| Dependent variable
I obtained FTC data on immigration scam reports through an FOIA request.iii According to FTC reports, between 400 and 700 immigration scams were reported annually between 2011 and 2014 (2314 total), which makes them about as common as anti-Hispanic hate crimes reported to the Federal Bureau of Investigation (FBI) over the same time period (1709 total, or 300-600 annually).Complaints are unevenly distributed across the country: 2650 counties reported no complaints during the study period, and few reported more than 10 complaints (Table 2).The number of FTC scam reports in California between 2011 and 2014 eclipsed similar complaints submitted to the California Attorney General between 2010 and 2015 (340 total, 50-64 each year).iv  Using FTC data, we can analyze where complaints are located.The FTC administrative data at the core of this paper required extensive data management and verification, including a case-by-case review of each complaint.I reviewed and geocoded each complaint recorded between 2011 and 2014 and supplemented location information with contextual data from secondary sources (Table 1).Nearly two-thirds of immigration services complaints are submitted directly to the FTC online (65%), with the remainder being referred to the FTC by organizations (e.g., the Better Business Bureau, state attorneys general, and nonprofit organizations) on behalf of consumers.In addition, a description of the incident accompanies each complaint, which allows us to review the content of each complaint.v

| Access to the safety net
In this paper, each of the main independent variables is continuous and varies across counties and over time.First, I measure access to the safety net among Hispanics, the largest immigrantorigin population.I use county-level data from the American Community Survey (ACS) to measure access to the safety net by accounting for the ratio of (a) Hispanic households receiving food stamp benefits and (b) the local Hispanic population living in poverty.In this paper, access to food stamps signals an inclusive approach to public benefits among local entities charged with administering the safety net because households receiving food stamps receive tangible support from a public program.Previous research shows that Hispanic households receiving food stamps include mixed-status households with unauthorized parents and US-born children (Hanson et al., 2014), and that select segments of the noncitizen population (e.g., pregnant mothers, refugees, long-term residents, and lawfully present residents) can also access the safety net (Fortuny & Chaudry, 2012).
Of course, Hispanic access to food stamps is not the only measure of safety net access.However, county-level ACS estimates covering the entire country are only available for one type of public program (food stamp receipt) and for one group relevant to this paper (Hispanics, but not immigrants or specific types of Hispanic households).Using available ACS microdata on a subset of US counties, I conduct robustness checks to examine whether a similar relationship holds when employing alternate measures of access to the safety net.

| Access to justice
Legal aid capacity is a proxy for access to justice in immigrant communities.Following work on nonprofit capacity (Chand et al., 2020;Gleeson, 2009Gleeson, , 2015;;Gleeson & Bloemraad, 2013;Joassart-Marcelli, 2013;Martínez-Schuldt, 2020;McVeigh et al., 2003;Yasenov et al., 2020), I manually construct a nationwide county-level measure of legal aid capacity specific to immigrant-serving nonprofits.I use a list of organizations that represent immigrants from the Department of Justice's (DOJ's) "Recognized Organizations and Accredited Representatives Roster" (Executive Office for Immigration Review, 2021) and supplement it with additional  2020), legal aid capacity in this paper captures variation in nonprofit resources, which I measure as county-year salaries for non-executive staff (divided by county size, in 2013 dollars, and lagged one year).I anticipate that greater capacity will mean that nonprofit staff will be able to spot crimes targeting vulnerable groups like noncitizens.In an FTC complaint, a nonprofit worker providing services to a client recounted how an immigrant victim of domestic violence had visited an informal organization selling immigration legal services.The worker helped the client file an FTC complaint against the informal provider, which had previously requested thousands of dollars to represent her case and then refused to turn over the client's personal documents after nonprofit staff became involved in the matter.
The analyses also account for the concentration of immigration lawyers (per 100,000 noncitizens in each county) listed as members of the American Immigration Lawyers Association (AILA) who also have experience working with deportation cases (American Immigration Lawyers Association, 2021).I anticipate that the concentration of these experienced lawyers will predict more reports of immigration scams to the FTC.Previous research has found that the proximity of immigration lawyers to immigrant detainees translated into positive outcomes for immigrant detainees (Ryo & Peacock, 2018, 2019).
Finally, the analyses also account for variation in language interpretation services.When local law enforcement aims to provide language access, these entities attempt to reach non-English speakers who may want to report a crime (Shah et al., 2007).Using law enforcement survey data (Department of Justice, 2011), I create an index of each county's total number of language interpretation services provided by a range of personnel and partners (per 100,000 noncitizens).Scam reports should be more common in counties with language access to law enforcement compared with counties without such services.

| Restrictionist immigration policy
Prior work has identified immigration restrictionism as an important predictor of the context of reception.State and local policies and programs designed to criminalize and deport immigrants escalated during the study period, albeit unevenly across states (De Trinidad Young & Wallace, 2019) and local communities (Moinester, 2018;Pedroza, 2019).I construct an index based on two existing measures of immigration restrictionism: a state-level measure of laws and policies that criminalize immigrants (minimum: 1; maximum: 6) as developed by De Trinidad Young and Wallace (2019), plus a county-level rate of deportations carried out by US Immigration and Customs Enforcement (ICE) under the Secure Communities program (Pedroza, 2019).The deportation rate equals: where deportations equal the number of county-level deportations under the Secure Communities deportation program adjusted for the size of the noncitizen population (n) and the number of days each county had the program (d).Deportation rates are lagged one year and logged to reduce the influence of outliers.Each of the two measures is first standardized (mean: 0; standard deviation: 1), then summed, and finally restandardized.Both measures capture a range of contexts, including the most restrictionist counties in the nation as well as locations where local communities have challenged hostile immigration laws (Steil & Vasi, 2014) and the Secure Communities program (Strunk & Leitner, 2013).
5.5 | Other correlates of crime report activity

| State contexts
My unit of analysis is county-years between 2011 and 2014.However, state-level variations in deportation rates (Pedroza, 2013), access to the safety net (Hero & Preuhs, 2007), and the concentration of legal aid (de Leon & Roach, 2013) and lawyers (Sandefur & Smyth, 2011) are all well known.In the regression models presented below, I include state fixed effects and cluster standard errors at the county level.

| Demographics
Echoing the negative relationship between immigrant concentration and crime reporting (Gutierrez & Kirk, 2017), reported consumer crime is typically lower in locations with high concentrations of Hispanics (Raval, 2016).As such, I adjust for key characteristics: (a) total population (logged); (b) total noncitizen population (logged); and (c) population density (per square mile).The analyses also account for a series of variables that might affect scam reports.

| Legal services market
In order to account for local variation in potential legal service needs, I use an index of internet search activity for Spanish and English language terms for immigration legal services, following the Google Trends method developed by DiGrazia (2017).vii

| Republican voter support
Support for restrictionist immigration policies tends to follow partisan divides (Chand & Schreckhise, 2014).I measure Republican voter support, or the proportion of votes for Republican Presidential candidates, as a percentage of those cast for the two major parties.I use 2008 results as predictors of FTC complaints in 2011 and 2012 and results from 2012 as predictors of complaints in 2013 and 2014 (Leip, 2012).

| Criminal justice context
Differences in the number of reports related to immigration scams may also stem from local variation in other rights-claiming actions.As such, DOJ data on anti-Hispanic hate crimes (logged) are included as a proxy for the positive relationship between civil society capacity and rights-claiming (McVeigh et al., 2003).Finally, it is possible that FTC complaints are merely a function of other crime reports.I account for five different arrest rates to determine whether FTC reports are associated with local arrest activity among a county's general population (per 100 residents).

| METHODS
FTC complaints capture incidents of people coming forward to denounce immigration scams.
To be clear, the underlying prevalence of such criminal activity is unknown, and the vast majority of counties had no reported scams during the study period (Table 2).Absent data on the location and prevalence of unreported immigration scams, the current study is limited to known cases and may not apply to the full universe of immigration scams.The challenge of analyzing known crime rather than all crime is not unique to this paper.Past work analyzing underreported crime, especially crime against vulnerable groups (e.g., hate crimes, sexual assault, wage theft, and workplace violations) has likewise focused on predicting known and reported incidents of crime against vulnerable groups.In doing so, I rely on count models that are ideally suited to account for variation in known incidents, especially rare events (Gleeson, 2015;McVeigh et al., 2003).The approach below is designed to differentiate between counties where we expect to find FTC complaints of immigration scams and locations where they are unlikely to be found given known county characteristics.
The outcome is a county-year total of events that features an excess number of zeros with a non-random distribution across the country.Count models can account for excess zeros, or counties with no complaints where at least one complaint is expected due, for example, to a sizeable immigrant presence.I employ zero-inflated negative binomial (ZINB) regressions to estimate two separate processes: whether we expect any complaints (0, 1), and the number of complaints expected given a set of county characteristics.The first process is estimated as: where ω represents the structural zeros, or zero counts not expected, and μ represents expected scam reports.The process follows a binomial distribution and assumes that some counties with no complaints also have a very low probability of immigration scams, such as small counties with few immigrants.
There are three reasons why we might expect zero scam reports: (a) consumer crime goes unreported; (b) no consumer crime occurred; or (c) there are no potential victims in the county.The first source of excess zeros remains unknown.In order to partially account for the second source of zeros, the logistic function measures anti-Hispanic hate crimes.If no hate crimes are reported, then the likelihood of observing scam reports may also be near zero if we assume that rights-claiming among victims of immigration scams is closely related rights-claiming among Hispanic victims of hate crimes.The third source of structural zeros (i.e., zero scams in places with no prospective victims) should be a function of population size.State fixed effects are included to account for unmeasured variation across states.
The probability of observing structural zeros is then used to weigh a second process, which follows a Poisson distribution for event outcomes (Long & Freese, 2006;Tutz, 2011) such as FTC complaints: The count process is a function of the independent variables described earlier as well as year fixed effects (reference year: 2011), and standard errors are clustered at the county level.

| RESULTS
The noncitizen proportion of a county (mean: 2.8%; standard deviation: 3.7%) is strongly related to the number of FTC complaints, but notable exceptions exist.Between 2011 and 2014, Los Angeles County, for example, reported the most immigration legal services crimes to the FTC (150), followed by San Diego County (56) and Miami-Dade County (53).The noncitizen proportion of these three counties' respective populations was high (between 12% and 23%) compared with the national average (<3%).In select cases, smaller counties contributed more than their anticipated share of scam reports given their county size; for instance, there were 20 combined scam reports filed in three counties (eight reports in Cobb County, Georgia; six reports in Marion County, Oregon; and six reports in Santa Cruz, California) with noncitizen populations of 9%-12%.However, other places do not necessarily report a comparable number of complaints given their respective share of noncitizens.Consider Madera County, California, and Cameron County, Texas, which have high percentages of noncitizens (16% and 17%, respectively) and over 25,000 and 70,000 noncitizens, respectively.Despite having such relatively high noncitizen populations, these counties reported fewer complaints (zero and three, respectively) than Denver, Colorado (10% noncitizen; 70,000 noncitizens), which had 14 complaints between 2011 and 2014.The above examples are emblematic: the number of FTC complaints is not uniformly associated with the size of a county's noncitizen population.Next, I examine whether scam reports are more common in welcoming contexts, net of immigration restrictionism and other predictors (Table 1 and Figure 1).Welcoming contexts of reception should predict a higher-than-expected number of FTC reports of immigration scams.The results are adjusted for differences in counties with no scam reports compared with those with scam reports.The models first compare whether counties without complaints differ from those with any FTC reports in terms of county size, population density, noncitizen population, anti-Hispanic hate crimes, and state context.The results below then help answer the main question: are indicators of welcoming contexts of reception reliable correlates of county scam reports once we account for variation in immigration restrictionism?
First, I examine the role of contexts of reception as correlates of FTC reports to determine whether scam reports are more common in counties with social helpers and sources of support.I differentiate between multiple proxies for welcoming and exclusionary contexts that are rarely measured when predicting an immigration-related outcome.It is worth noting how unevenly distributed different types of social and legal services are across the nation.For example, only 4% of counties had multiple nonprofit organizations offering legal aid to immigrants, including F I G U R E 1 Total FTC complaints related to immigration legal services crime (by noncitizen share of county's population).Source: Author's tabulations of FTC data.Circle size equals total number of FTC complaints (0 to 150) within each county for the entire study period (2011)(2012)(2013)(2014).Among counties with any complaints, the mean is 4.8 complaints (standard deviation: 10.2 complaints; 25th percentile: 1 complaint; 75th percentile: 4 complaints).Noncitizen population shares also vary across the nation (mean: 2.8%; standard deviation: 3.7%), especially among counties where noncitizens comprise more than 10% of the county (mean: 14.7%; standard deviation: 4.6%) those who cannot afford lawyer fees.Similarly, 5% of counties had multiple immigration lawyers with deportation experience.Compared with legal services, public sector services like food stamps and language interpretation services were-albeit far from universal-less unevenly distributed.Specifically, the ratio of Hispanic households receiving food stamps to poor Hispanics was robust (i.e., above one-to-three) in one-fifth of US counties.In addition, about one-eighth of counties surveyed in 2007 reported having multiple language interpretation options.With respect to immigration enforcement, about one-seventh of counties reported 15 or more deportations during this time period, which coincides with a nationwide increase in mass deportations.The results confirm that restrictive and welcoming contexts predict FTC complaints in the expected opposite directions: more reports of known scams are filed in counties with welcoming contexts and fewer come to light in counties where immigration restrictionism is common.
Table 3 displays results from ZINB models predicting changes in the expected number of complaints after adjusting for a range of county-level factors.Each coefficient represents the expected change in FTC complaints per standard deviation increase in each measure of the context of reception.In the first model, I find support for the idea that a robust safety net is associated with more reports of immigration scams.Specifically, a standard deviation increase in access to the safety net coincides with an 18% rise in FTC reports.For reference, a standard deviation increase in the safety net measure (going from 0.23 to 0.45) is akin to going from a county where poor Hispanics outnumber Hispanic households receiving food stamps four-toone to a county where the latter are outpaced by only two-to-one.Supplementary analyses using ACS microdata of identifiable counties (15% of the nation) confirm that safety net access predicts more reports of known scams when we measure safety net access to either food stamps or Medicaid among different subpopulations (e.g., Hispanics, immigrants, Hispanic households with US-born children, or immigrant households with US-born children).Furthermore, the association between food stamp access and scam reports is not explained by states' safety net generosity for immigrants (Hero & Preuhs, 2007).viii  Webs of support can also come in the form of legal aid assistance for immigrants, experienced immigration lawyers, and local law enforcement agencies providing language services.I find evidence that each of these also predicts more FTC reports denouncing immigration scams.In Model 2, a standard deviation increase in legal aid capacity coincides with a 1.8% rise in the expected number of complaints.Substantively, the results suggest that increasing legal aid resources from the national average of $1.12-19.38 per capita (or two standard deviations higher) would be associated with nearly a 4% increase in expected complaints.In addition, counties with access to experienced lawyers and language services also tend to be counties in which noncitizens file a higher-than-expected number of FTC complaints.In Model 3, a standard deviation increase in immigration lawyers per capita is associated with a 20% increase in expected complaints.Similarly, in Model 4, a standard deviation increase in the concentration of law enforcement agency language services predicts an 8% increase in total complaints relative to what we would otherwise expect.In sum, each proxy for welcoming contexts predicts an increase in the number of people willing to report scams.
If welcoming contexts positively predict the number of FTC reports, scam reports should be inversely related to restrictionist immigration policies.As expected, restrictionist intensity, as captured by state laws criminalizing immigrants as well as Secure Communities deportation rates, is associated with fewer-than-expected FTC complaints.Specifically, in Model 5, a standard deviation increase in restrictionism is related to an 8% drop in complaints.
So far, the results confirm a relationship between FTC complaints and proxies for local contexts of reception.But are reports still more common in welcoming contexts when adjusting for immigration restrictionism?Net of all factors (Model 6), I find that welcoming contexts predict a higher-than-expected number of scam reports, while restrictionism and FTC complaints remain inversely related.Among the proxies of welcoming contexts, the magnitude of each of the relationships remains largely the same (see Model 6 compared with other models in Table 3).
Among other correlates of scam reports, hate crime reports and a county's concentration of noncitizens both predict more complaints than otherwise expected.Google search activity for immigration legal services and Republican voter support predict FTC reports in opposite directions, although neither measure is statistically significant in the full model.None of the five arrest rates is statistically significant.

| Robustness checks
Next, I discuss whether the results hold if we examine (a) variation in FTC complaints over time, (b) variation by the source of FTC complaints, and (c) possible interactive relationships between restrictionist and welcoming indicators.ix First, do the results hold if we predict yearto-year changes in FTC complaints?In fixed-effects count models, the availability of immigration lawyers, language services, and legal aid is not sufficient to predict a year-to-year rise in FTC reports.Meanwhile, access to the safety net remains just as reliable a predictor of complaints in fixed-effects or multilevel negative binomial regressions as in the above results.
Second, what if we analyzed the subset of reports filed by organizations on behalf of individuals?In ZINB models predicting FTC reports filed by organizations (N = 799, or 35% of all immigration scam complaints), each welcoming context measure is positively related to FTC reports.Similar to the above results, safety net access is most positively correlated with FTC reports filed by organizations, as compared with other indicators of welcoming contexts.In fixed-effects models, each measure of welcoming contexts predicts more scam reports filed by organizations.Notably, the relationship between nonprofit legal aid and scam reports facilitated by organizations was more pronounced than when analyzing the association between legal aid capacity and all complaints.
Third, the main results predicting FTC complaints above assume that indicators of welcoming versus restrictive contexts operate independently of each other.But what if these relationships intersect with each other to predict scam reports?Not surprisingly, welcoming immigrant destinations with a longstanding history of integrating immigrants (e.g., Chicago, New York, Los Angeles) have more FTC reports than expected, while new immigrant destinations with restrictionist policies (especially in the Southeast and Southwest) tend to have few scam reports.Other locations, however, have a mix of both inclusionary and exclusionary contexts.For instance, select counties with exclusionary contexts (Bexar and El Paso, Texas; Durham and Wake, North Carolina; Fulton, Georgia; Pima, Arizona) report higher-than-expected numbers of scam reports, perhaps because their county seats (San Antonio and El Paso; Raleigh-Durham; Atlanta; and Tucson, respectively) also have higher-than-average access to the safety net and/or legal services.However, interactive models are inconclusive.In cross-sectional ZINB models, only the interaction between restrictionism and access to the safety net is positive, which suggests that Hispanic access to food stamps can outweigh the negative association between restrictionism and FTC complaints, although the relationship is not statistically significant.Other interactive models are also not statistically significant, suggesting that neither restrictionist nor welcoming indicators reliably outweigh each other when predicting scam reports.

| Limitations
These results should be interpreted with caution.Similar to studies of hate crimes (McVeigh et al., 2003) and workplace violations (Gleeson, 2015), the analyses performed here cannot assess the extent of underreported crime; they work to predict the numerator (i.e., those willing to report scams) because the denominator (i.e., the actual number of scam incidents) is unknown.In addition, it remains possible that unmeasured factors affect both (a) where noncitizens lodge FTC complaints and (b) variation in the contexts of reception.Scam reports may be influenced by variation in immigrant civic participation (Ebert & Okamoto, 2013;García Bedolla, 2014;Terriquez, 2012) or other determinants of engagement such as economic downturns, evictions, and foreclosures (Rugh, 2020;Slee & Desmond, 2021).In analyses not presented above, proxies of these measures (e.g., protest size per noncitizens during the spring 2006 immigration marches, lagged unemployment rates, etc.) were not reliable predictors of scam reports.However, other civic and economic factors might predict scam reporting.Moreover, the analyses rely on one important source of information on immigration scams (the FTC), but results may differ when examining other sources of information on scam reports at the federal, state, or local level.
Furthermore, the contextual analyses are constrained by available data sources.Ideally, the analyses could account for individual-level variation across scam reports to examine, for example, reports filed by unauthorized versus lawfully present immigrants, especially since these immigrants' experiences as "semi-citizens" are not uniform (Cohen, 2009).Nevertheless, this study leverages variation across and within US counties to identify reliable predictors of scam reports.Despite the shortcomings of the FTC data, this paper provides valuable insights into where noncitizens report consumer crime in welcoming contexts using a novel nationwide dataset on immigration scams reported to the FTC.

| DISCUSSION
When weighing whether to report scams, noncitizens must decide whether making claims to justice might do them more harm than good.The conflicting commitments (Gleeson, 2012) of rights-affirming US institutions and immigration enforcement complicate the decisions of noncitizen victims of crime, who must weigh the possible benefits of reporting a scam against the potential complications of reporting crimes that may expose their legal status.When noncitizens seeking legal services are misled and decide to report such scams, geographic variation in levels of welcomingness versus restrictionism toward immigrants (Provine et al., 2016;Varsanyi et al., 2012;Williams, 2015) predicts where noncitizens choose to report immigration scams to the FTC.Specifically, welcoming counties have more immigration scam complaints, and counties with exclusionary contexts tend to have fewer complaints.Taken together, the results suggest that immigrant-serving capacity and access to key services support noncitizens who report immigration scams, while hostility toward immigrants may deter them from exercising those same rights.Crucially, although we observe fewer-than-expected FTC complaints in counties where immigration restrictionism is common, exclusionary policies do not account for the higher-than-expected number of FTC reports in more welcoming contexts.
Next, I discuss avenues for promoting welcoming initiatives, with an emphasis on the safety net, legal aid capacity, access to immigration lawyers, and language services.Since these webs of support can promote noncitizen rights-claiming, the uneven distribution of such services may hinder acts of insurgent citizenship (Leitner & Strunk, 2014a, 2014b) such as denouncing immigration scams.Currently, the gap between where noncitizens do and do not exercise the power of claims-making could have potentially long-term consequences for "civic inequality" (Bloemraad et al., 2020).Investing in welcoming initiatives will require addressing the mismatch between the concentration of support services and the concentration of noncitizen residents, especially in contexts where restrictive immigration policies remain common and where access to supportive services remains scarce.

| Investing in welcoming initiatives
The evidence presented here collected from an examination of FTC complaints provides mesolevel insights into the degree to which different types of welcoming contexts might matter for insurgent citizenship.A central lesson of this paper is the reliable role of access to the safety net as a predictor of where immigrants feel empowered to enact an insurgent citizenship in the form of rights-claiming acts.Expanding access to safety net services helps families to make ends meet and, in turn, may also assist immigrants in navigating public institutions.Based on FTC data alone, we cannot know for certain that experience with the safety net translates into heightened trust of institutions in general.However, it seems plausible that immigrants applying for social services might also gain such experience when engaging other public institutions.
My results echo qualitative work on how the safety net can function as a web of support for immigrants.Social helpers in the public sector (Marrow, 2009) can relay information about how US bureaucracies work.Households seeking food stamp assistance undergo a rigorous application process.Crucially, noncitizen applicants applying on behalf of eligible US citizen children need to trust that the information they share will not expose them to immigration authorities.Safety net providers must overcome these and related barriers (e.g., limited access to bilingual staff, complex applications and eligibility rules, etc.) when promoting immigrant families' access to safety net services (Perreira et al., 2012).In order to secure and retain benefits, families eligible for safety net services also periodically renew benefits, which can require assistance from community organizations (Crosnoe et al., 2012;Goodman-Bacon & Cunningham, 2019;Yoshikawa et al., 2014).In sum, promoting and funding welcoming, culturally responsive, and meaningful access to the safety net-alongside the public sector's respective nonprofit, community-based partners-can help immigrants interact with trusted government entities.
The results also point to immigration attorneys, translation services, and nonprofit legal aid as indicators of welcoming contexts, whose representatives empower or enable noncitizens to make their rights real (Bloemraad, 2006;Gleeson, 2009;Hagan, 1994).Local governments and nonprofit funders should invest in ways to improve access to experienced low-cost or pro bono legal and translation services.For example, the data sources I used can help to match immigrants looking for legal services with reliable organizations and experienced service providers.Immigrants seeking legal services can search for service providers in their communities-by location, language, populations served, and immigration specialties-using online directories.The data I analyze is derived from a directory of AILA members (American Immigration Lawyers Association, 2021) with experience handling deportation cases, the Immigration Advocates Network's directory of community organizations (Immigration Advocates Network, 2021), and the Executive Office for Immigration Review's catalog of DOJ-accredited representatives (Executive Office for Immigration Review, 2021).The Immigration Advocates Network has also begun cataloguing remote and pro bono legal service providers that can meet legal service demand in places with limited local-level capacity.Some data sources analyzed in this paper, however, are not accessible in a user-friendly manner, such as DOJ data on enforcement agencies providing language interpretation services.If made publicly available in a searchable format, this information could also assist immigrants in search of assistance.Efforts to document which communities and organizations provide welcoming (Huang & Liu, 2016) and immigrantfriendly (Smith & Schmitt-Sands, 2021) services for local communities in the nonprofit (Yasenov et al., 2020) and public sectors (Williams, 2015) could also help identify gaps in assistance and catalog programs and organizations that warrant investment.
I also find differences in how noncitizens choose to report immigration scams.Specifically, scam reports filed by intermediary organizations are similarly related to each measure of welcoming contexts, with one exception: nonprofit legal aid.In supplementary analyses, I examined FTC reports filed by organizations (N = 799; one-third of reports) and found that these scam reports were more strongly related to legal aid capacity than those filed by individuals.Although I do not aim to identify which services legal aid staff, especially non-attorney practitioners (Han, 2019;Shannon, 2011), can and should deliver, it seems that nonprofit legal aid can play a tangible role in supporting immigrants' insurgent citizenship.
Although the main results and the data sources noted above point to the supportive role of trusted and experienced service providers, daunting challenges remain.To begin, even the most welcoming contexts may not have enough experienced service providers to accommodate all immigrant populations.As a result, acts of insurgent citizenship depend on whether webs of support have the capacity to assist individuals whose questions or language needs reflect service providers' experience.Complicating matters further, experienced providers respond to a rising demand for services by queuing immigrant clients, which related work has found can result in lawyers deciding not to pursue a case (Galli, 2020).Moreover, long waitlists contribute to an unmet need for legal services (de Leon & Roach, 2013;Legal Services Corporation, 2017), especially since only a few counties are home to multiple service providers.Recall that only 4% of counties had multiple legal aid options, and only 5% had multiple AILA lawyers.Given such a mismatch, access to support services in large immigrant destinations may not resemble access in new destinations, suburbs, small towns, and rural areas (Carrillo, 2016;Cordero-Guzm an et al., 2008;Panchok-Berry et al., 2013).
While access to experienced immigration lawyers and legal aid was concentrated in a small proportion of counties, two sources of support were more common: access to the safety net and the availability of language interpretation services among law enforcement agencies.As noted in the results, one-fifth of US counties reported robust, higher-than-average access to food stamps, and one-eighth of counties surveyed had multiple language interpretation options.Immigrants looking to report immigration scams who live in counties without lawyers or legal aid may then turn to these public sector sources of assistance.In small or remote counties where the safety net and law enforcement are immigrants' main (or lone) points of contact, investing in welcoming initiatives means improving access to publicly funded outreach and service delivery.Absent local sources of support, immigrants may instead rely on nearby immigrant destinations if they prefer to receive assistance from longstanding, trusted community organizations located in large urban centers (de Graauw et al., 2013;de Leon et al., 2009).
Although all scam victims-regardless of legal status-have both the same rights "on the books" and comparable opportunities to report consumer complaints to the FTC, county-level variation in how many actually do so reveals where noncitizens trust the FTC enough to report scams.If webs of support and experience working with trusted intermediaries can promote noncitizens' claims to justice, the evidence presented here points to multiple promising avenues for assistance at the local level.Such support can help ensure that the onus of identifying scams does not fall on the targets of consumer crime-as cautioned in prior research on notario fraud (Longazel & Fleury-Steiner, 2013)-but rather on immigrant-serving and public safety organizations tasked with protecting everyone's rights.Otherwise, if large segments of the noncitizen population do not exercise their rights as consumers, or if doing so remains dependent on an uneven distribution of social helpers, then US society will reinforce existing inequalities between noncitizens and everyone else.In response, scam artists will have ample incentive to innovate in how they take advantage of immigrants, which undermines equal protection under the law.

| Open questions about scam reports and the unknown prevalence of scams
Future research can address two open questions: why are scams reported more often in certain locations, and what is the prevalence of scams?On the first issue of scam reporting, the insights into FTC scam reports discussed so far cannot conclusively test which mechanisms explain the results.Although organizations can help individuals make claims to justice in welcoming contexts, two other explanations are also possible.For instance, news stories have detailed how nonprofit staff can mishandle cases or mislead clients (Fertig, 2017), which has led some to argue that only lawyers can properly serve clients (Das, 2016).If nonprofits do more harm than good, then the positive association in this paper might mean that scams are more common where nonprofits operate.But this interpretation, which characterizes the dependent variable as a measure of scam prevalence, would also have to explain why immigration restrictionism predicts fewer scam reports.Alternatively, if scam artists focus on traditional immigrant destinations when searching for potential scam victims, then the results might suggest that welcoming contexts draw the attention of scammers.These results cannot rule out such an explanation.Nevertheless, we can safely conclude that the number of scam reports in a given location of scam reports is correlated with the number of reported hate crimes, not arrests for a range of other offenses, suggesting that scam reports are a proxy for rights-claiming rather than prevalence of arrests.Future research can disentangle the exact mechanisms that may explain why more scam reports get filed in welcoming locations and fewer-than-expected complaints are reported where immigration restrictionism abounds.
The second issue concerns the actual prevalence of scams.To the best of my knowledge, no dataset captures the prevalence of scams beyond those incidents reported to law enforcement.To arrive at a prevalence estimate, we could gather information concerning where scam artists are most likely to place advertisements for legal services across locations.Based on the number of generic Craigslist ads for immigration-related legal services (in the 20 largest immigrant cities between 2016 and 2018), the nine least welcoming cities averaged five times more ads for immigration-related help than the five most welcoming cities.Similarly, the Craigslist sites for six cities with a mix of welcoming and restrictive policies saw double the number of these ads than the sites for the most welcoming cities.In other words, these ads were much more common outside of welcoming immigrant destinations like Chicago, San Francisco and the Bay Area, and New York, where FTC reports are concentrated.Consistent with these patterns, recent work suggests that unlicensed and inexperienced legal service providers seek captive markets in places where immigrant-serving organizations are sparse ( de Graauw & Gleeson, 2020).
Future empirical research could also sample a range of immigrant destinations and audit a diverse group of legal service providers to estimate the risk of fraud.If we knew where scams are most likely to happen, then such information could help identify locations and groups most likely to be targeted by scam artists.A prevalence study could also determine whether and how much the universe of alleged scams-including those that go unreported-differs from those known to law enforcement agencies such as the FTC.An audit study could help establish a sampling frame for additional work on immigrants' unmet needs for legal representation, which could mushroom if the federal government reforms the nation's legalization processes.

| CONCLUSION
This paper provides the first nationwide analysis of immigration legal services crime, providing insights into where noncitizens make claims to justice.Currently, scam reports remain uneven and patterned after local contexts of reception.According to the findings in this paper, reports of immigration scams lag behind in counties where these restrictionist initiatives have taken hold.These policies may end up exchanging escalating enforcement for trust in public institutions.It seems that we may have much to learn from locations that provide robust services for the poor, immigrants, and non-English-speakers, as these are the places where noncitizens most often report immigration scams to the FTC.If extending such webs of support across local immigrant communities can encourage immigrants to make claims to justice, then we know where to start: access to the safety net, access to legal counsel and representation, and translation services.
enforcement initiatives exacerbate these inequalities and leave imprints in our local communities.How to cite this article: Pedroza, Juan Manuel.2022."Making Noncitizens' Rights Real: Evidence from Immigration Scam Complaints."Law & Policy 44(1): 44-69.https://doi.org/10.1111/lapo.12180 Data sources, coding, and descriptive statistics T A B L E 1Note: Each variable is measured at the county level, except for Google Trends data (state level).Multivariate analyses standardize all variables (mean: 0; standard deviation: 1).
de Leon and Roach (2013)mplaints and counties (by year and range of FTC complaints) : National Center for Charitable Statistics (NCCS) data queried byde Leon and Roach (2013), the Immigration Advocacy Network (Immigration Advocates Network, 2021), and a list of Catholic Legal Immigration Network members.viI also include nonprofits that filed 1099 forms with the Internal Revenue Service.As suggested byBloemraad et al. ( sources negative binomial regression results (incident rate ratios predicting expected count of FTC complaints per standard deviation change in independent Standard errors in parentheses.Robust standard errors clustered at the county level.ZINB results control for excess zeros as a function of a county's total population (logged), noncitizen population (logged), population density, hate crimes, and state fixed effects.Count model variables are displayed above, except for the five arrest rates (which are introduced in Model 7, although none are statistically significant) and year fixed effects (included in all models with a reference year of 2011).Data sources described in Table1.The results reflect standardized independent variables.