The Contradictions of Liminal Legality: Economic Attainment and Civic Engagement of Central American Immigrants on Temporary Protected Status

Abstract

This study examines how Temporary Protected Status (TPS) may shape immigrants’ integration trajectories. Building on core themes identified in the immigrant incorporation scholarship, it investigates whether associations of educational attainment with labor market outcomes and with civic participation, which are well established in the general population, hold for immigrants who live in the “liminal legality” of TPS. Conducted in 2016 in five U.S. metropolitan areas, the study is based on a unique survey of Salvadoran and Honduran TPS holders, the majority of immigrants on this status. The analyses find that TPS holders with higher levels of educational attainment do not derive commensurate significant occupational or earnings premiums from their education. In contrast, the analysis of the relationship between educational attainment and civic engagement detects a positive association: more educated TPS holders are more likely to be members of community organizations and to participate in voluntary community service, compared to their less educated counterparts. These findings illustrate the contradictions inherent to TPS as it may hinder certain aspects of immigrant integration but not others. This examination contributes to our understanding of the implications of immigrants’ legal statuses and of immigration law and policy for key aspects of immigrant integration trajectories.

Immigration scholarship has emphasized the key role of legal status in the lives of immigrants (Bean, Brown, and Bachmeier 2015; Light, Massoglia and King 2014; Menjívar and Abrego, 2012; Waters and Gerstein Pineau 2015), increasingly conceptualizing it as an important axis of stratification (Asad and Clair 2017; Gee and Ford 2011; Greenman and Hall 2013; Massey 2007; Menjívar 2011; Rugh and Hall 2016). Legal status affects immigrants’ employment opportunities and choices, income, physical and mental health, access to education and services, housing and living arrangements, experiences of discrimination, and intentions to return (Agadjanian, Menjívar, and Zotova 2017; Asad and Clair 2017; Gee and Ford, 2011; Greenman and Hall 2013; Hall and Greenman 2015; Landale, Oropesa and Noah 2017; Massey and Gelatt 2010; McConnell 2015; Terrazas et al., 2020; Oropesa, Landale, and Hillemeier 2016). Yet, while research has shown that legal status affects almost all spheres of life, its effects intersect with other social cleavages to produce heterogeneous experiences within the immigrant population (Abrego 2014a; Cebulko 2018; Gonzales and Burciaga 2018; Terriquez, Brenes, and Lopez 2018). Importantly, legal status is not binary; it cannot be categorized easily into a legalized vs. illegalized (documented-vs-undocumented) dichotomy.¹ In fact, the legal production of migrant il/legality has created and expanded the gray area of in-between statuses (Cebulko 2014; Genova and Zontini 2020), or what Menjívar (2006) defined as “liminal legality.”

Statuses that are liminally legal, and, therefore, temporary and uncertain, have expanded significantly, and today they can be conceptualized as a spectrum (Joseph 2016). In Menjívar’s (2006) original conceptualization of liminal legality, Temporary Protected Status (TPS) was the emblematic liminal legal status in the United States. But other varieties of temporary statuses have existed, such as those resulting from long waiting times for visa approval, expanded temporary worker programs, and programs like DACA (Deferred Action for Childhood Arrivals), which share many of the characteristics of TPS albeit each with distintct relationships to the legal regime. And although temporary legal statuses are not new in the management of immigration flows and state crafting, within the past two decades or so they have expanded significantly around the world, especially in the large immigrant-receiving countries (Cook-Martin 2019). The similarities across these statuses regarding implementation and requirements may obscure unique legislative contexts behind their provenance, implementation, groups covered, and termination. For instance, some temporary statuses, like TPS, originate in legislative decisions and, thus, the status becomes part of the Immigration and Nationality Act, even as the decision to designate countries for the status depends on the political currents of the day. Others, such as DACA, are enacted through executive order and tethered to the political agendas of the administration in power, and because they are not legislative acts, they are not part of the Immigration and Naturalization Act and usually are conferred onto a specific group. Other temporary statuses, such as H-2A (agricultural) and H-2B (non-agricultural) employment visas, are grounded in labor force demands and allocated to employers, so that they can fill temporary labor shortages (Hernández-León 2020; Hernández-León and Sandoval Hernández 2017; Portes 2020). Indeed, the expansion of temporary statuses to fill labor needs, Portes (2020) argues, is so extensive that, it “has rapidly become the dominant component of the American immigration system, relegating stable permanent immigration to a secondary role” (2020: 13). While all these statuses position their holders in temporary, precarious legality, they vary in the degree of anchoring they provide, the uncertainty, the consequences of termination (see also Waters and Gerstein Pineau 2015), and the holders’ relationship to the legal regime.

Because such temporary statuses do not lead to permanent legal status, immigrants living in “liminal legality” can return to an illegalized status when the political currents that shape immigration policy decisions change direction (Menjívar 2006). This is exemplified most recently by the volatile climate that the Trump administration has created with its decision to end TPS designation for several countries, even those that have been designated for this dispensation for years, such as El Salvador, Haiti, and Honduras, as well as more recent designees, such as Nepal. Importantly, the in-between legality of immigrants in temporary statuses and the ambiguity embedded in such statuses lead to experiences that differ in many respects from those of illegalized as well as legalized immigrants (Bergeron, 2014; Cebulko, 2014; Coutin 2000; Farcy and Smit 2019; Hallett 2014; Menjívar 2006, 2017; Menjívar and Coutin 2014).

In this study, we investigate the integration outcomes of “liminally legal” immigrants using unique recent survey data collected from Honduran and Salvadoran immigrants who are TPS holders. Although these data do not allow for direct comparisons with other legal or citizenship statuses, they offer an opportunity to test whether general patterns of economic and social incorporation are present among this group. Specifically, our study is focused on the associations of educational attainment with labor market outcomes and with civic engagement, as these associations have been identified as core themes in the scholarship on immigrant integration (Jiménez, Park, and Pedroza 2018; Waters and Gerstein Pineau 2015).

There is ample evidence that in the general population educational attainment is positively associated with employment and earnings (Blau and Duncan 1967; Hout 2012; Tamborini, Kim, and Sakamoto 2015), and with civic participation (Campbell 2009; Egerton 2002; Rafail and Freitas 2016). Among immigrants, however, the relationship between education and labor market outcomes varies and depends greatly on legal status (Hall, Greenman, and Farkas 2010), especially when legal status is coupled with the enforcement regime. For instance, Gentsch and Massey (2011) find that prior to the passage of the Illegal Immigration Reform and Immigrant Responsibility Act (IIRIRA) in 1996, greater education correlated with increased employment rates and higher wages among Mexican immigrants. However, after 1996, the relationship of educational attainment with both job attainment and earnings disappeared (Gentsch and Massey 2011). Furthermore, wages for immigrants have stagnated, even among immigrants with higher levels of education (Massey and Gelatt 2010). This is despite the fact that since the 1970s, the educational level of immigrants, particularly from Mexico, has risen (Massey and Gelatt 2010). As for civic participation, even though immigrants have been found to have lower levels of civic engagement than the native-born population overall (Waters and Gerstein Pineau 2015), these levels can vary by legal status (Bloemraad 2013), institutional threats, local enforcement efforts (Ebert and Okamoto 2013), and extent of inclusionary context (Okamoto and Ebert 2010); they are affected by socioeconomic characteristics such as education (Dávila and Mora 2007; Foster-Bey 2008) and by an immigrant group’s home country conditions of exit (Terriquez 2012). Civic engagement among immigrants can produce critical political gains (see Brettell 2020), even when overall anti-immigrant sentiment is in high gear.

Building on this research, we investigate whether the associations of educational attainment with labor market outcomes and with civic participation hold for immigrants who live in the “liminal legality” of TPS. Our study thus contributes to understanding the implications of vulnerable legal statuses for key aspects of immigrant incorporation trajectories. Given the expansion of temporary statuses across groups and national contexts today, our examination is germane beyond its specific focus and geographic location, as TPS holders’ experiences in the United States can help illuminate and address the challenges faced by other liminal-status groups such as DACA recipients and other groups in similar statuses around the world. Significantly, our analysis, focused on long-term holders of TPS status, sheds light on the consequences of long-term temporariness for immigrant integration. Our study, therefore, exposes the significance of immigration law and policy for understanding immigrant integration more generally.

TEMPORARY PROTECTED STATUS

TPS was initially signed into law by President George H. W. Bush as part of the Immigration Act of 1990 (which adjusted the Immigration and Nationality Act of 1965) to provide temporary protection to immigrants who are unable to return to their origin countries due to an ongoing armed conflict, environmental disasters, or other extraordinary conditions. The Secretary of Homeland Security designates the countries whose nationals are deemed in need of this protection. This status provides relief from “removal” from the United States for a period between six and eighteen months, and those holding this status are also eligible for a work permit. To underscore the fundamental uncertainty of this status, renewals are announced just months before the designation is about to expire. In principle, holders of the status are eligible for travel authorization through advanced parole (USCIS 2020a), but many opt for staying put as policies may change and thus they may risk being unable to return. Since the status was created in 1990 it has been renewable provided that conditions in the designated origin country do not assure a safe return and that the protected individual registers for renewal, pays a $495 fee,² and complies with a series of requirements that include having a clean criminal record. No new applicants can be admitted into the program after a country’s first designation date and TPS is technically not considered “admission” for the purposes of adjustment of status. Indeed, this technical designation means that the TPS population is included in the estimated 10.6 million illegalized population in the country (Warren 2020). The Immigration Act of 1990 specifically prohibits the adjustment of TPS status to lawful permanent residence, with rare exceptions,³ as well as the enactment of a law that would allow such adjustment, unless it receives three-fifths supermajority approval in the Senate (Bergeron 2014:25).⁴ As of January 2017, there were an estimated 325,000 immigrants on TPS from ten designated countries: El Salvador, Haiti, Honduras, Nepal, Nicaragua, Somalia, South Sudan, Sudan, Syria, and Yemen, with approximately 78 percent being from El Salvador or Honduras (Warren and Kerwin 2017).⁵

The Trump administration has ended several of these country designations, with some set to end at various points in 2020 and 2021, coinciding with an election period and adding to the uncertainty about any future extensions. Immigrants from countries that have lost TPS designation under the current administration have been given between six and eighteen months to leave the country. However, several lawsuits have been filed to appeal the government’s decision, and, as of this writing, these lawsuits are still in the courts. Decisions to end or expand designations are tied to specific court cases.⁶ And even though TPS designation may end for specific groups, the legal status itself is part of the Immigration Act of 1990 and may be used for other groups or the same groups may be redesignated in the future. As of this writing, TPS designations for both Honduras and El Salvador are set to end on January 4, 2021.

Since the creation of TPS status in 1990, several countries have been designated and undesignated.⁷ However, those designations were brief, most lasting between a few months and three years, with one or two renewals,⁸ and they represented a small fraction of the TPS population. In contrast, Hondurans and Salvadorans have been on TPS for two decades, renewing their permit thirteen times. Importantly, the legal and social context has changed; in today’s climate of rising anti-immigrant hostility and ramped-up enforcement, the potential consequences of TPS termination are far greater and harmful than in the past.

Significantly, TPS holders’ experiences of living in uncertain legality parallel those of immigrants in similarly uncertain liminal legal statuses, as such statuses have become a common tool to control immigration in immigrant-receiving countries around the world today. Temporary legality is thus no longer exceptional; it has expanded and multiplied, emerging in different modalities across categories of admission in the United States and around the world (see Allerton 2017; Farcy and Smit 2019; Goldring and Landolt 2013; Kilpatrick 2020; Kubal 2013; Nissim and De Vries 2014; Thayer Correa, Stang and Abarca 2016). Our examination, therefore, is theoretically and analytically relevant beyond the TPS case in the United States.

A contradiction between the intention of temporary protection in the TPS program and its requirements and implementation on the ground has placed TPS holders in a quasi-permanent state of legal uncertainty. To be eligible for TPS renewal, immigrants must demonstrate continuous presence in the United States from the time their country was designated for TPS. This means that TPS holders from El Salvador and Honduras have lived “temporarily” in the United States, continuously since 2001 and 1999, respectively, often unable to travel to their country of origin.⁹ Salvadorans and Hondurans together constitute not only the largest group on TPS, but, except for a small group of Somalis who has held TPS since 1991, they also have the longest tenure on TPS: in fact, most Honduran and Salvadoran immigrants on TPS have spent half of their lives (or more) living in temporary legality. Although scholars have examined short-term consequences of holding temporary legal statuses such as DACA (Gonzales, Terriquez, and Ruszczyk 2015; Martinez 2014), the constrained opportunities for socioeconomic mobility that DACA creates (Hsin and Ortega 2018), and the effects of transitioning from illegalized to the temporary status of DACA (Patler and Pirtle 2017), little is known about long-term consequences of holding temporary legal statuses for immigrant incorporation. Examining the experiences of Salvadoran and Honduran TPS holders, who have held this status for two decades, can illuminate important aspects of the link between temporary, quasi, or liminal, legality and immigrant incorporation.

Although Salvadoran and Honduran TPS holders have spent, on average, between one third to one half of their lives in the United States and have contributed economically, socially, and culturally to local communities and U.S. society as a whole, their partial legal status has thwarted their integration (see also Waters and Gerstein Pineau 2015). Research has documented how the uncertainties of quasi-legal statuses reshape or redirect these immigrants’ socioeconomic integration (Bean et al. 2015; Bergeron 2014; Cebulko 2014; Coutin 2000; Hallett 2014; Menjívar 2006). While TPS holders are able to access economic opportunities they would otherwise be ineligible for and do relatively better in the job market than their iligalized counterparts (Menjívar 2017; Orrenius and Zavodny 2015), the temporariness of TPS hinders other forms of integration and mobility, producing contradictory experiences of incorporation (Hallett 2014), or its “incompleteness” (Brown and Bean 2006). As the National Academies of Sciences, Engineering, and Medicine’s report on immigrant integration observes, “TPS confers partial inclusion while simultaneously affirming (with periodic reminders) that this status is temporary and partial” (Waters and Gerstein Pineau 2015:140).

LEGAL STATUS, ACHIEVEMENT, AN D IM MIGRANT INTEGRATION

According to Linda Bosniak (2000), citizenship is composed of four different strands—legal status, rights, civic and political engagement, and collective identity and belonging (see also Bloemraad, Korteweg, and Yurdakul 2008). In Bosniak’s conceptualization, these strands do not necessarily align, and the various facets of integration may proceed at different paces and may not even match up. Thus, an individual can feel part of a society (in the sense of collective identity or solidarity), while lacking the formal legal recognition that national or legal status conveys (Bloemraad 2013). This in-congruence may be more acute for immigrants who have lived temporarily for an extended period, as their “in-between” position in the face of long-term residence can challenge normative conceptualizations of immigrant integration.

One key aspect of theorizing immigrant incorporation is in identifying not only factors that facilitate it, but also those that can derail or stall it. Different dimensions of integration, such as economic, political, or social, can vary in degrees of completeness within the same group, leading to “assimilation incompleteness” (Brown and Bean 2006). Importantly, as Brown and Bean (2006) observe, the type of incompleteness matters; analyzing factors that lead to incompleteness in one dimension but not in another is key for theorizing incorporation, a point that is also policy relevant. Temporary and uncertain statuses can influence the degree of (in)completeness of a particular aspect of integration as legal status can affect, for instance, whether immigrants acquire English language skills fast enough (see Brown and Bean 2006). Thus, the incorporation trajectories of immigrants in temporary statuses in many ways mirror the in-betweenness or liminality of this status, leading to in-corporation incompleteness. We focus on variations in incorporation in/completeness among TPS holders along two key dimensions—economic attainment and civic engagement—with implications for immigrant incorporation theorizing more generally.

Economic Attainment

Economic attainment is a key axis of immigrant integration (Massey, Durand, and Malone, 2002; Rumbaut and Komaie 2010; Telles and Ortiz, 2008) and central in theorizing immigrant assimilation (Jiménez et al. 2018). Studies have investigated various barriers to immigrants’ economic attainment (Portes 1995; Van Tubergen, Maas, and Flap 2004; Waters and Eschbach 1995). The scholarship that examines the role of legal status has revealed the negative effects of an illegalized status on income (Durand, Massey, and Pren 2016; Hall et al. 2010; Massey, Durand, and Pren 2016; Massey and Gellat 2010). Consistent with this scholarship, research on employment and earnings of workers on TPS has noted that TPS holders do better in the labor market compared with illegalized immigrants (see Orrenius and Zavodny 2015), but not as well as immigrants in more secure statuses (Menjívar 2017), thus reflecting TPS holders’ in-between status.

It also has been observed that the wages of illegalized immigrants or of those who hold temporary status are not commensurate with their levels of human capital, skills, or training (Abrego 2014b; Hagan, Leal, and Rodriguez 2015; Menjívar 2000). However, there is no systematic evidence on how the association between education and labor market outcomes among TPS holders may deviate from more general patterns.

Civic Engagement

Civic engagement is another important dimension of immigrant integration. Civic engagement can be a complementary, but also a compensatory, dimension to the economic sphere, particularly when groups are blocked out of other opportunities for advancement (Ebert and Okamoto 2013). For instance, perceived unfair treatment and discrimination from employers and generally poor labor conditions, which also hamper economic advancement, can propel compensatory civic engagement (Kasinitz et al 2009; Milkman 2006; Mora et al. 2018; Suárez-Orozco, Hernández, and Casanova 2015). Anti-immigrant hostility and restrictions may also encourage immigrant organizing and collective civic activities (Ebert and Okamoto 2013; Okamoto and Ebert 2010). Thus, blocked opportunities that arise from legal vulnerability can drive civic engagement among immigrants, even though participation in immigrant rights movements is uneven and can have different integrational outcomes depending on the objectives of the organization (Nicholls 2019).

Generally, immigrants tend to have lower levels of civic participation and volunteerism than the native-born population (Waters and Gerstein Pineau 2015), and some scholars have even argued that the diversity that immigration brings may play a role in the decline of overall civic engagement (Putnam 2007). However, the difference between immigrants and non-immigrants largely disappears when controlling for other factors, such as socioeconomic status and education (Dávila and Mora 2007; Foster-Bey 2008; Lopez and Marcelo 2008; Stepick, Stepick, and Labissiere 2008). Some scholars suggest caution with the debate on whether the increased diversity that comes with immigration undermines overall civic engagement; they note that we should instead attend to the various forms that civic engagement takes, including participation in and ethnic-serving organizations (Ramakrishnan and Bloemraad 2008:2). The length of time that immigrants have spent in the host society is another important factor for immigrants’ civic participation: as duration of residence in the United States increases, so does their civic engagement (Stoll and Wong 2007). However, civic engagement among immigrants also has been found to vary by legal status, with citizenship, in particular, positively affecting the likelihood of participating in civic life (Lopez and Marcelo 2008; Stoll and Wong 2007). Research also has documented extensive civic involvement among immigrants in precarious legal statuses (Bloemraad and Trost 2008; Delgado 1994; Leal 2002; Nicholls 2013, 2016; Varsanyi 2005). Yet, it is unclear how the patterns and correlates of civic engagement among these categories of immigrants differ from those among the general population. Specifically, although a strong positive association has been found between educational attainment and participation in civic activities among the general population (Campbell 2009; Rafail and Freitas 2016), this association has not been examined among immigrants in liminal legal status.

Hypotheses

Guided by the reviewed literature, we formulate two sets of alternative hypotheses to examine the association of education with labor market outcomes and with civic participation among TPS holders. We examine whether associations identified in the general population also hold for immigrants on TPS. In the labor market realm, we first look at occupation type—higher-skilled vs. lower-skilled. If liminal legal status does not undermine the employment premium of educational attainment, then higher levels of education among TPS holders will translate into a greater likelihood of employment in higher-skilled occupations. However, if liminal legality hinders the positive effects of human capital on employment, more educated TPS holders will not significantly differ from the less educated ones in the type of occupational allocation. Second, we look at the association of education with earnings of those employed. Again, if liminal legal status does not hamper the premium of education, then TPS holders will display a positive association between educational level and earnings. If, on the contrary, liminal legal status does matter, then no educational gradient in earnings among TPS holders will be present. Accordingly, the two sets of alternative hypotheses are:

Hypothesis 1–1.A:

Higher levels of education among TPS holders will translate into a greater likelihood of having a higher-skilled occupation compared to a lower-skilled one.

Hypothesis 1–1.B:

Better educated TPS holders will not be more likely to hold a higher-skilled occupation than a lower-skilled one.

Hypothesis 1–2.A:

TPS holders will display a positive association between educational level and earnings.

Hypothesis 1–2.B:

The educational gradient in earnings will not be present among TPS holders.

For civic engagement, we examine two outcomes – membership, formal or informal, in civic organizations and participation in voluntary community service. As with labor market outcomes, we test two sets of alternative hypotheses. If the liminal legality of TPS holders does not impinge on their civic engagement, we expect them to follow the association identified more generally, that is, greater participation in civic organizations and voluntary community service among the more educated. Alternatively, if the liminal legality of TPS holders constrains their civic engagement, no significant differences between more and less educated TPS holders in levels of participation in civic organizations and in voluntary community service would be observed. Hence, the hypotheses:

Hypothesis 2–1.A and 2–2.A:

More educated TPS holders will have higher levels of membership in civic organizations (2–1.A) and of participation in voluntary community service (2–2.A) than less educated ones.

Hypothesis 2–1.B and 2–2.B:

There will be no significant differences between more and less educated TPS holders in levels of membership in civic organizations (2–1.B) and of involvement in voluntary community service (2–2.B).

DATA AND METHOD

Data

We use a unique dataset from a multi-site survey of TPS holders conducted between March and October 2016. This first ever survey of immigrants on Temporary Protected Status included mainly immigrants from El Salvador but also from Honduras, the two largest groups of TPS holders with the longest time on TPS.

Using U.S. Census data, we selected five metropolitan areas with the highest concentrations of immigrants from the two Central American countries: Los Angeles, San Francisco, Houston, Washington, D.C., and New York/New Jersey.¹⁰ The survey was conducted through the University of Kansas Center for Migration Research¹¹ and was directed by Menjívar in collaboration with organizations working in the Central American community in each metropolitan area. Obviously, no national sampling frame for TPS holders is available. To build the sample, we used the lists of TPS holders supplied by community organizations that provide legal support for TPS applications and renewals. The Salvadoran consulates also offered their lists of TPS holders to assist in selection. Participants were randomly selected from those lists. This selection process yielded a sample of 2,094 (1,123 men and 971 women), 1,936 of whom were from El Salvador and 157 from Honduras.¹²Well-trained survey interviewers, themselves TPS holders, carried out the survey interviews in Spanish in all sites.¹³ Most interviews were conducted by telephone, but a small number were done in person at the offices of the community organizations. The survey was approved by the Institutional Review Board of the University of Kansas.

It is, of course, impossible to fully ascertain the degree to which our sample is representative of the Salvadorans and Hondurans on TPS. However, we should note that key sociodemographic characteristics of our sample generally resemble those based on the estimates of the TPS population using American Community Survey (ACS) data (see Warren and Kerwin 2017), thus instilling confidence in our sample’s representativeness.

Method

To test our hypotheses on the association between education and labor market outcomes we use two measures for respondents who were employed at the time of survey. The first measure is relative occupational status, coded as a dichotomy: higher-skilled (coded 1) vs. lower-skilled (0) occupations. The higher-skilled category includes nurse, electrician, stylist/barber, car mechanic, plumber, and similar occupations; the lower-skilled category consists of dishwasher, cleaner, janitor, and comparable occupations. This occupational dichotomy modifies the standard occupational classification by deliberately avoiding the frequently used demarcation of “skilled” vs. “unskilled” labor. As Hagan, Hernández-León, and Demonsant (2015) have argued, such framing hides and devalues the important skillset and work experience that immigrants labeled as “unskilled” often bring to their work. The second labor market outcome is respondent’s income. We use log-transformed earnings to smooth out the distribution.

We test the hypotheses on civic engagement using the following two dichotomous outcomes: respondent’s membership (regardless of its formalization) in neighborhood, children’s school, church, work-related, sports, or other type of community organization in the past 12 months (was a member of at least one such organization = 1, was not = 0); and participation in voluntary activities benefiting the respondent’s community, such as donating blood and participating in neighborhood street cleaning (participated at least once = 1, never participated = 0). The sample size is larger in models for civic engagement than it is for the labor market outcomes: unemployed TPS holders (men: 71, women: 166) are excluded from the occupational models, and those employed who did not report their earnings (men: 83, women: 55) are additionally excluded from the earnings models. These respondents are retained in the civic engagement models if they answered the questions on organizational membership or voluntary community service.

Our main explanatory variable is educational attainment, which is classified into three levels: less than high school, high school or GED, and at least some university or higher. Due to the already small fraction of those with at least some university education (9 percent of men and 10 percent of women), we do not disaggregate them further based on whether they completed university education (sensitivity analysis separating those with complete and incomplete tertiary education yields similar results as the analysis using the three educational levels).

The statistical models control for the country where the highest level of education was received – the United States vs. the country of origin – because U.S.-based education is an important marker of immigrant integration and may affect our outcomes of interest (cf. Zeng and Xie 2004). The models also control for demographic and other characteristics potentially related to immigrants’ socioeconomic and civic practices. These controls are: age in linear and quadratic forms, marital partnership characteristics (no partner, partner living in the United States, partner living in the country of origin), number of children, country of origin (El Salvador or Honduras), number of years respondent has lived in the in the United States, and place of residence (Houston, Los Angeles, New York/New Jersey, San Francisco, and Washington, D.C.). In addition, the earnings model controls for occupation types (lower-skilled vs. higher-skilled).

We fit binary logistic regression for dichotomous outcomes: occupational status, organizational membership, and voluntary community service. Linear regression is used for log-transformed earnings. All models are estimated separately by gender, because women’s and men’s employment patterns differ vastly (Bielby and Baron 1986), and gender segregation in the labor market remains universally pervasive (Blau, Brummund, and Liu 2013; Blau, Ferber, and Winkler 2001; Del Río and Alonso-Villar 2015; Reskin 1993). Likewise, there is evidence that patterns and levels of civic engagement also vary by gender (Burns, Schlozman, and Verba 2001; Wilson and Musick 1997).

Table 1 summarizes the definition and distribution of the outcomes and covariates by gender. The vast majority of both women (88 percent) and men (96 percent) were employed.¹⁴ Two thirds of employed men but just 45 percent of employed women worked in what we defined as higher-skilled occupations. Men’s median monthly earnings were $3,050, while the median monthly earnings of women were much lower, $1,800. Both figures, but especially those for women, are lower than the estimates for the general U.S. population in 2016, $4,303 and $3,463, for men and women, respectively (Hegewisch and Williams-Baron 2017). A third of female respondents reported membership in civic organizations, compared to 28 percent of their male counterparts.¹⁵ Similar shares of both women and men (20 percent and 22 percent) reported being involved in a voluntary community service activity at least once in the twelve months preceding the survey.

Table 1.

Sample Characteristics

	Men		Women
	Mean (Median)	S.D.	Mean (Median)	S.D.
Employment status
Not employed	0.038	0.190	0.122	0.327
Employed	0.962	0.190	0.878	0.327
Occupational status
Lower-skilled	0.249	0.433	0.550	0.498
Higher-skilled	0.751	0.433	0.450	0.498
Monthly income
Log-transformed earnings	8.021	0.954	7.434	0.851
(Median earnings)	($3,050)		($1,800)
Membership in a civic organization
No	0.723	0.448	0.672	0.470
Yes	0.277	0.448	0.328	0.470
Voluntary community service
No	0.784	0.412	0.798	0.402
Yes	0.216	0.412	0.202	0.402
Educational level
Less than high school	0.648	0.478	0.650	0.477
High school or GED	0.266	0.442	0.253	0.435
Some university or higher	0.086	0.281	0.098	0.297
Highest education obtained in the U.S.
No	0.883	0.322	0.889	0.315
Yes	0.117	0.322	0.111	0.315
Age	43.758	9.384	45.621	10.094
Years in the U.S.	19.832	4.252	20.760	4.707
Marital partnership status
None	0.359	0.480	0.508	0.500
Partner in the U.S.	0.581	0.494	0.467	0.499
Partner outside the U.S.	0.061	0.239	0.025	0.155
Number of children	2.184	1.621	2.655	1.561
Country of origin
El Salvador	0.932	0.251	0.916	0.277
Honduras	0.068	0.251	0.084	0.277
City of current residence
Houston	0.316	0.465	0.252	0.435
Los Angeles	0.295	0.456	0.395	0.489
New Jersey and New York	0.076	0.265	0.059	0.235
San Francisco	0.098	0.297	0.091	0.287
Washington, D.C.	0.215	0.411	0.203	0.402

Note: The total sample size is 1,123 for men and 971 for women (missing cases for each variable are excluded).

Almost two thirds of both women and men had less than high school education and fewer than ten percent of respondents of both genders had at least some university education. This falls below the average in corresponding statistics among the general U.S. population, where 86 percent of men and nearly 88 percent of women hold a high school diploma or higher and about 30 percent of both men and women hold bachelor’s degrees or higher (U.S. Census Bureau 2016). Just over one tenth of both male and female respondents received their highest education in the United States.

The vast majority of respondents were from El Salvador: only seven percent of men and eight percent of women were from Honduras. The average age of respondents was 44 years for men and 47 years for women. They had lived in the United States for an average of 20 years (men) and 21 years (women). Thirty-six percent of men and 51 percent of women had no marital partner; the overwhelming majority of those in marital relationships had their marital partners also living in the United States. Male respondents reported 2.2 children on average, while the average number of children reported by female respondents was 2.7.

RESULTS

Bivariate associations between predictors and outcomes by gender are presented in Table 2. As the educational level increases, the share of those employed in higher-skilled occupations rises slightly among men, but the difference between respondents at both ends of the educational attainment spectrum is not statistically significant. Among women, who generally have a much smaller share in higher-skilled occupations, this share also rises with education and the difference between those with less than high school education and those with at least some university degree is statistically significant. Log-transformed income shows a similar ascending trend across the educational spectrum, but the differences are not statistically significant among either women or men. In contrast to relatively small and mostly non-significant associations between education and labor market outcomes, the associations between education and civic engagement are quite pronounced. At each educational level both the share of those reporting membership in civic organizations and the share of those reporting participation in voluntary community services increases markedly and significantly (p<.01).

Table 2.

Bivariate Associations

		Total				Difference
			Less than high school	High school or GED	Some university or higher	Pearson Chi2	T-value
Men
Occupational status	Lower-skilled	24.9%	26.5%	22.4%	19.3%	3.236
	Higher-skilled	75.1%	73.5%	77.6%	80.7%
Log-transformed earnings	Mean	8.021	7.984	8.065	8.151		1.889
	(S.D.)	(0.954)	(0.782)	(1.360)	(0.660)
Membership in a civic organization	No	72.3%	75.2%	67.7%	58.5%	14.448**
	Yes	27.7%	24.8%	32.3%	41.5%
Voluntarycommunity service	No	78.4%	83.5%	71.8%	59.1%	38.097***
	Yes	21.6%	16.5%	28.2%	40.9%
Women
Occupational status	Lower-skilled	55.0%	58.3%	52.3%	44.6%	6.368*
	Higher-skilled	45.0%	41.7%	47.7%	55.4%
Log-transformed earnings	Mean	7.434	7.396	7.489	7.553		1.804
	(S.D.)	(0.851)	(0.776)	(0.863)	(1.158)
Membership in a civic organization	No	67.2%	71.0%	63.1%	54.3%	12.672**
	Yes	32.8%	29.0%	36.9%	45.7%
Voluntary community service	No	79.8%	83.8%	76.2%	62.9%	23.700***
	Yes	20.2%	16.2%	23.8%	37.1%

Significance level:

^*p < .05;

^**p < .01;

^***p < .001 (two-tailed tests).

Notes: The sample size is 1,123 for men and 971 for women (missing cases for each variable are excluded). Pearson Chi² tests were used for the differences among three educational groups. T-tests were used for the differences between those with some university education or higher and those with less than high school education.

Table 3 reports the results of multivariate models for occupational type (Section 3.A) and earnings (Section 3.B). As can be seen in Section A, the coefficients for high school/GED and at least some university education are in the predicted direction but are not statistically significant in both women’s and men’s models. Hence, Hypothesis 1–1.B is supported. In the linear regression predicting earnings, occupational level is added as a control. Again, we observe the expected trend in the effect of education, but similar to the occupational-level model, the effect of educational attainment on earnings among both women and men is not statistically significant. The results support Hypothesis 1–2.B.

Table 3.

Occupation Status (logistic regression) and Earnings (OLS regression), Parameters Estimates and Standard Errors

	3.A Occupational status: Higher -skilled vs. lower-skilled				3.B Earnings (log-transformed)
	Men		Women		Men		Women
	B	S.E.	B	S.E.	B	S.E.	B	S.E.
Educational level (reference = Less than high school)
High school or GED	0.079	(0.198)	0.034	(0.196)	0.103	(0.080)	0.117	(0.084)
Some university or higher	0.313	(0.309)	0.201	(0.277)	0.208	(0.117)	0.210	(0.118)
Highest education obtained in the U.S. (reference = Not in the U.S.)
Yes	0.221	(0.326)	0.525	(0.320)	0.118	(0.123)	−0.146	(0.139)
Age	0.001	(0.064)	−0.072	(0.060)	0.026	(0.026)	0.035	(0.026)
Age2	0.000	(0.001)	0.001	(0.001)	0.000	(0.000)	0.000	(0.000)
Years in the U.S.	0.007	(0.020)	−0.001	(0.019)	0.009	(0.008)	0.004	(0.008)
Partner (reference = No partner)
Yes, in the U.S.	0.459	(0.165)**	−0.093	(0.157)	0.240	(0.069)***	0.029	(0.067)
Yes, outside the U.S.	0.197	(0.329)	−0.011	(0.513)	0.201	(0.139)	0.131	(0.227)
Number of children	−0.078	(0.051)	−0.012	(0.054)	0.037	(0.022)	−0.023	(0.023)
Country of origin (reference = El Salvador)
Honduras	0.187	(0.325)	−0.030	(0.275)	−0.209	(0.134)	−0.045	(0.125)
City (Reference = Houston)
Los Angeles	−0.079	(0.200)	0.213	(0.201)	−0.211	(0.083)*	0.072	(0.088)
New Jersey and New York	−0.304	(0.292)	0.736	(0.345)*	0.006	(0.124)	0.220	(0.146)
San Francisco	−0.213	(0.269)	0.038	(0.287)	0.079	(0.116)	0.280	(0.127)*
Washington, D.C.	0.424	(0.229)	−0.573	(0.241)*	0.035	(0.087)	0.205	(0.100)*
Occupational status (reference = Lower-skilled)
Higher-skilled					0.180	(0.073)*	0.038	(0.066)
Intercept	1.266	(1.467)	1.256	(1.386)	7.203	(0.587)***	6.619	(0.596)***
N	979		748		897		693
−2LL	1059.764		993.934
Pseudo R2	0.028		0.034
R 2					0.069		0.033
Adjusted R2					0.053		0.012

Significance level:

^*p < .05;

^**p < .01;

^***p < .001 (two-tailed tests).

Among other covariates, the country where the highest level of education was obtained is not significantly associated with earnings. Length of time in the United States is not significant either.¹⁶ Occupational level and earnings of both women and men do not seem to vary by age and the number of children.¹⁷ Having a partner in the United States increases men’s likelihood of holding a higher-skilled job as well as their earnings, but neither relationship is significant among women. There are no significant differences between Salvadorans and Hondurans in either occupational level or earnings. Some variation in the type of occupational allocation across survey sites seems to exist, but it requires additional investigation that lies beyond the aim of the present study. Finally, holding a higher-skilled occupation leads to higher earnings among men; among women, however, this association is absent.

Table 4 presents the results of the models predicting civic engagement. Echoing the bivariate associations, these results point to a positive association of education with both organization membership (Section 4.A) and voluntary service (Section 4.B). Thus, women and men with some university education are significantly more likely to report membership in civic organizations, compared to their counterparts with no high school diploma (corresponding odd ratios, obtained by exponentiating the regression coefficients, are OR=exp(.744)=2.104 for men and OR=exp(.868)=2.382 for women). Among women, the difference between those with a high school diploma/GED and those without a high school diploma is also statistically significant. Hypothesis 2–1.A is, therefore, supported. Similarly, male respondents with at least some university education are much more likely than their least educated counterparts to report participation in voluntary community activities (OR=exp(1.085)=2.959). In fact, men holding high school diplomas or GEDs are also significantly different from the reference group (OR exp(0.506) 1.659). Among women, this difference is much smaller and not statistically significant, but the difference between the two extreme educational categories is significant (OR exp(.755) 2.128).¹⁸ The results provide support to Hypothesis 2–2.A.

Table 4.

Membership in Civic Organizations and Participation in Voluntary Community Service (Logistic Regression, Parameter Estimates and Standard Errors)

	4A Membership in a civic organization				4.B Voluntary community service
	Men		Women		Men		Women
	B	S.E.	B	S.E.	B	S.E.	B	S.E.
Educational level (reference = Less than high school)
High school or GED	0.282	(0.195)	0.424	(0.196)*	0.506	(0.207)*	0.117	(0.236)
Some university or higher	0.744	(0.267)**	0.868	(0.269)**	1.085	(0.272)***	0.755	(0.299)*
Highest education obtained in the U.S. (reference = outside the U.S.)
Yes	0.434	(0.280)	0.177	(0.305)	0.175	(0.287)	0.685	(0.326)*
Age	−0.005	(0.059)	−0.047	(0.054)	−0.018	(0.062)	−0.116	(0.058)*
Age2	0.000	(0.001)	0.000	(0.001)	0.000	(0.001)	0.001	(0.001)
Years in the U.S.	0.013	(0.020)	0.018	(0.018)	0.044	(0.021)*	0.039	(0.021)
Partner (reference = No partner)
Yes, in the U.S.	0.342	(0.167)*	0.044	(0.156)	0.256	(0.184)	0.192	(0.187)
Yes, outside the U.S.	0.179	(0.349)	−0.412	(0.552)	0.127	(0.406)	0.161	(0.605)
Number of children	0.090	(0.054)	0.092	(0.053)	0.035	(0.059)	−0.043	(0.064)
Country of origin (reference = El Salvador)
Honduras	0.667	(0.322)*	0.200	(0.291)	0.566	(0.342)	0.426	(0.316)
City (reference = Houston)
Los Angeles	0.958	(0.222)***	0.879	(0.219)***	1.389	(0.232)***	0.780	(0.243)**
New Jersey and New York	2.141	(0.293)***	2.181	(0.337)***	1.195	(0.331)***	0.054	(0.445)
San Francisco	2.053	(0.272)***	1.662	(0.293)***	0.851	(0.320)**	0.233	(0.379)
Washington D.C.	1.416	(0.228)***	1.066	(0.244)***	0.707	(0.264)**	0.292	(0.297)
Intercept	−2.121	(1.340)	−1.182	(1.258)	−2.618	(1.400)	0.045	(1.341)
N	1,028		897		1,014		880
−2LL	1108.734		1049.426		957.702		809.336
Pseudo R2	0.108		0.074		0.090		0.075

Significance level:

^*p < .05;

^**p < .01;

^***p < .001 (two-tailed tests).

Among other covariates, Honduran men are more likely than their Salvadoran counterparts to report membership in a civic organization. Interestingly, among men, having a partner in the United States increases the likelihood of civic organization membership. The membership model also shows a lower level of membership in Houston, compared to other sites. Having obtained the highest level of education in the United States has a net positive association with the likelihood of voluntary service among women but not among men. In comparison, time spent in the United States is a significant predictor of involvement in volunteer work among men but not among women. Volunteer work participation rate also shows some variation across study sites. These patterns require further exploration.

DISCUSSION AND CONCLUSION

Using unique data from a recent multi-site survey of Honduran and Salvadoran TPS holders, we examined whether the associations of educational attainment with labor market outcomes and civic engagement, i.e., the associations that have been well established in the general population, are present among this group of temporary protected immigrants. Although with our data we could not directly compare TPS holders with immigrant groups in other legal statuses or with the non-immigrant population, our findings shed important light on the experiences of liminal legality as they allow us to infer, even if indirectly, how these experiences may imprint the universal patterns of economic and civic incorporation.

We found no net occupational or earnings premium for the educational attainment of TPS holders. These results suggest that TPS holders with high levels of educational attainment do not derive commensurate significant occupational and economic benefits from their education. Although we cannot explain this finding with our data directly, we suggest that reduced economic returns to education among this group reflect the specifics of their legal status. The very status that authorizes TPS holders to work legally may also constrain their economic mobility. Thus, because their work permits must be renewed every eighteen months, employers’ concerns about potential delays and other problems with renewal (and more recently, about possible discontinuation of the entire program) may keep them from promoting these immigrants to higher paid positions even if employers value their TPS workers and some industries have come to depend heavily on them (Yee, Robbins, and Dickerson 2018). Temporary legal protection does not shield immigrants from pernicious employment practices (Chacón 2015). Importantly, the longer TPS workers remain in lower-paid positions, the more employers rely on them and associate them with those positions, resulting in the racialized notion that these workers are inseparable from those economic niches because they do the “work that no one else wants” or because they have acquired a reputation for being the “best” workers for those jobs. Such racialized associations shape perceptions of these immigrants, with potential long-term effects for this group.

In contrast to our findings regarding labor market outcomes, our analysis of the relationship between educational attainment and civic engagement and voluntary community service detected a significant positive association. Thus, despite their liminal legal status, more educated TPS holders tend to be more engaged civically than their less educated counterparts—both in terms of organizational participation and in terms of voluntary community service. The limited legality of TPS status does not seem to hinder the typical mechanisms that shape civic engagement. Hence, civic engagement may not be solely complementary to economic integration, but it may also be compensatory as immigrants engage in community organizations and rights movements to advocate for improving the conditions that block them in other spaces of integration (see also Ebert and Okamoto 2013; Okamoto and Ebert 2010; Terriquez 2012).¹⁹ A meaningful example is the National TPS Alliance, the nation-wide organization that formed in the wake of the Trump administration’s announcement to terminate TPS. In just two years, TPS holders have organized at least sixty committees throughout the country, holding demonstrations, lobbying legislators, filing key lawsuits, mounting campaigns to provide information about TPS status, establishing a media presence, and advocating for TPS continued renewal and for the long-term goal of a more permanent status.

In sum, our findings illustrate the complexity and contradictions of living in liminal legality. Reflecting its in-betweenness, this status hinders certain aspects of immigrant integration but not others, producing seemingly inconsistent experiences for immigrants in this status (cf. Hallett 2014). Perhaps more than other immigrants, those with temporary legal status display patterns of incorporation incompleteness (Brown and Bean 2006), paralleling the unevenness often found among the different strands of citizenship as legal status, as rights, and as belonging (see Bosniak 2000). On the one hand, their socioeconomic integration diverges from paths typically found among immigrants with more secure statuses (as well as among the non-immigrant population). On the other hand, their civic engagement trajectories, in general, follow a more typical pattern, demonstrating engagement in community organizations, neighborhood associations, but also in organizations that seek to change the conditions that may hinder incorporation. Our findings thus underscore the distortions in conventional associations between education and integration outcomes that come from limited access to socioeconomic mobility (see also Hsin and Ortega 2018 for DACA).

These findings suggest that the barriers to labor force integration and economic mobility are steeper for liminally legal immigrants than those that stand in the way of their civic engagement. At the same time, for better educated immigrants civic engagement may provide a channel for integrating more fully in their communities and neighborhoods. Additional, especially qualitative, research is needed to fully understand the mechanisms behind the detected patterns. Direct comparisons of TPS holders with immigrants in other legal statuses would also be necessary to fully ascertain the unique implications of TPS. Although our findings do not offer all the answers, they, nonetheless, contribute to an understanding of legal status as a new axis of stratification. As more immigrants are moved to uncertain and temporary legal statuses in the United States and in other major immigrant-receiving countries (see Agadjanian et al. 2017; Gonzales and Burciaga 2018; Massey 2007), legal status stratifies immigrant populations into hierarchical classes with unequal access to society’s resources, rights, and rewards. Our results are also highly relevant for policy, as government initiatives today aim at ending these forms of temporary legality for some groups (such as those on whom we focused) and create and expand temporary legality for others, while alternative models to TPS are also being proposed (Frelick 2020). Temporary legality, granted for just a few months at a time, thwarts potential, and limits plans and aspirations (see Menjívar 2008) with detrimental consequences for socioeconomic mobility in the long term. Our findings reveal the critical place of immigration law and policy for understanding immigrant incorporation across immigrant-receiving societies today.