Race Relations, Black Elites, and Immigration Politics: Conflict, Commonalities, and Context

Race Relations, Black Elites, and Immigration Politics: Conflict, Commonalities, and Context Abstract This paper investigates the question of whether and how race and race relations affect state legislators’ support for restrictive immigration policy. Focusing on the nine new immigration destinations in the US Southeast, we compare the roll call votes cast by African American and White Democratic state legislators on 196 restrictive immigration measures proposed between 2005 and 2012. Overall, the majority of Black Democratic legislators support restrictive immigration legislation, although at a consistently lower rate than White Democratic legislators. We test hypotheses that Black-White differences in support for restrictive immigration legislation depend upon the material resources at stake and the symbolic significance associated with specific legislation. Using bill topic to measure these conditions, we find that Black and White Democrats equally support immigrant “competition” bills on employment topics. Black Democratic legislators are less likely than their White counterparts to support immigration bills on Black-immigrant “commonalities” topics related to civil rights issues: voter ID regulations, education, police, and omnibus restrictions. Further, for civil rights bills only, district threat indicators such as a rapidly growing district Latino population are much more likely to increase support for restrictive immigration bills among White Democratic legislators than among Black Democratic legislators. Some district characteristics even evoke a commonalities response from African American Democratic legislators, reducing their support for restrictive legislation. Our research thus expands the competition/commonalities debate and highlights the complexities and contingencies of race relations and state policymaking in the twenty-first century. Introduction Rapid increases in immigration to new US destinations are transforming local demographic landscapes and challenging state politics traditionally based on a Black-White racial divide (Marrow 2011). The growing number of immigrants could spur African American political leaders to ally with conservative Whites to push anti-immigrant legislation, particularly if African Americans perceive immigrants as competitors for jobs and resources. Alternatively, African American leaders may unite with immigrants and their supporters if they see immigrants, like themselves, as disadvantaged minorities just a few generations away from the South’s Jim Crow system. Indeed, McKanders (2010) warns that Southern states in particular are creating a new system of “Juan Crow” that directly parallels Jim Crow, with laws preventing undocumented immigrants from attaining access to housing, jobs, education, and healthcare (Lovato 2008). Other anti-immigrant legislation tightens access to the ballot box and extends the reach of law enforcement—issues at the heart of the Black Civil Rights movement. With jobs at stake on the one hand and civil rights issues at stake on the other, do African American lawmakers support or oppose “Juan Crow” laws? We address this question by comparing African American and White Democratic legislators’ roll call votes on restrictive immigration bills proposed between 2005 and 2012 in the “Nuevo South”—the states of the former Confederacy identified as new immigrant destinations (AL, AR, GA, LA, MS, NC, SC, TN, and VA) (Marrow 2011). The unique racial history of these “Nuevo South” states has produced especially fertile conditions for testing the theories of Black/immigrant competition and commonalities that undergird our study. More than other regions in the United States, the social and political institutions in the “Nuevo South” are centered on the Black/White binary, and Black representatives have gained a large presence in state politics (Bositis 2011; King-Meadows and Schaller 2007). Approximately 44 percent of all Black state legislators serve in the nine “Nuevo South” states, and these Black legislators wield more political clout within the Democratic party than do their counterparts outside the region (King-Meadows and Schaller 2007). Black Democratic legislators in the “Nuevo South” do not share political power with Latino representatives as they do in traditional immigrant gateway states.1 While Black/White relations define state politics and virtually every important social and political outcome in the “Nuevo South” (McClain et al. 2009, 1), this region has experienced the most rapid growth in immigration of any region of the country (Pew 2015). The rapidly rising Latino immigrant population in states where the vast majority of representatives and their constituencies are Black or White could be perceived by Blacks as a threat to their hard-won political and economic gains in the region, prompting Black representatives to take an anti-immigrant stance. Yet, the history of race relations in the “Nuevo South” also could generate sympathy toward immigrants among Black representatives and prompt a pro-immigrant response. Unlike other regions of the United States, the “Nuevo South” carries the enduring legacy of Jim Crow racism, with a proud identity as the “cradle of the civil rights movement” (Black and Black 1989; Gaillard 2004; McAdam 1990). The few existing studies of African American lawmakers’ stances toward immigration policy are limited to traditional immigration gateways or national politics and primarily focus on conditions that promote alliances between Black Democratic legislators and Latino political caucuses (e.g., Diamond 1998; Hero and Preuhs 2013). These studies yield mixed hypotheses and findings with respect to African American Democratic legislators and immigration legislation, with some studies finding African American support for restrictive immigration legislation and other studies finding opposition to these bills (see Hero and Preuhs [2013] for a review). Our theoretical framework draws on scholarship that identifies the US racial hierarchy as a driving force behind immigration politics. Theories of group threat posit that lawmakers’ perception of immigrant competition arising from the relative position of Blacks, Latinos, and Whites in the racial hierarchy will stimulate support for restrictive immigration legislation among both Black and White Democratic legislators. Theories of symbolic politics suggest that beyond material group interests, the values and beliefs arising from a history of race relations and the continued struggle for racial justice will also influence African American legislative behavior toward immigration policy. In particular, Black support for restrictive immigration policy may drop if Black lawmakers perceive a common goal of racial justice and a shared social position with immigrants as disadvantaged minorities in a White-dominated racial hierarchy. White Democratic legislators are much less likely to perceive systematic racial inequality and to place racial justice among their legislative priorities when dealing with immigration (Masuoka and Junn 2013). We develop a research design that allows us to uncover the role of group interests as well as values and beliefs in shaping African American legislative responses to restrictive immigration policy, respectively. In particular, we divide bills by topic, and compare votes on topics related to economic group interest with votes on civil rights issues. By analyzing the roll call votes of Southern Democratic state legislators, we effectively control for the influence of other factors such as partisanship and political ideology. We exclude Republicans not only because there is virtually no variation in Republican votes on restrictive immigration bills, but also to ensure that the racial differences that we do observe cannot be attributed to the most prominent alternative explanation—partisanship. Given the uniformly higher rates of support among Republicans (99 percent of whom are White), the Black-White differences that we report in the paper would be even stronger if we included Republicans.2 Our findings suggest that rather than reflecting unilateral competition or commonalities responses, immigration politics in the “Nuevo South” is multidimensional and context specific. First, although the majority of African American Democratic legislators support restrictive immigration legislation, they are voting “yea” at lower rates than their White counterparts. Second, African American legislative behavior varies depending on the topics associated with specific bills. When bills concern immigrant employment, African American Democratic legislators join their White colleagues and vote to reduce economic threat. When bills concern civil rights, however, African American Democratic legislators diverge from their White colleagues and dampen their support for restrictive immigration legislation. Consistent with the theory of symbolic politics, these commonalities responses of African American Democratic lawmakers do not simply reflect African American group interest, but are tied to the values and ideologies of racial justice—that is, the symbolic significance—associated with specific bills. We also find that whether district demographics prompt voting patterns reflecting an immigration threat or an opportunity to build interracial bridges depends on both the bill topic and the race of the legislator (Jones-Correa 2011). We elucidate how these factors are interrelated below, first describing the context of our study and then discussing theories of group threat and symbolic politics. We state our hypotheses derived from these theories, describe the methods we use to test our hypotheses, and present our results. In the conclusion, we move beyond a consideration of legislative behavior as a response to racial inequality, and discuss the broader relevance of our study to theories that situate the state at the center of producing racial categories and enforcing the boundaries between these categories. The Rise of the Nuevo South and the Latino Backlash With local immigrant populations rising 200–800 percent in traditionally Black-White communities throughout the Southeast, immigrants have encountered a social and political backlash (Odem and Lacy 2009; Passel and Cohn 2011). Amid public outcry about “illegals” stealing jobs, burdening taxpayers, and increasing crime rates, state and local officials across the Southeast have passed a range of restrictive laws targeting unauthorized immigrants. These laws aim to limit access of unauthorized immigrants to resources such as jobs, public benefits, and educational opportunities, as well as expand the powers of local law enforcement to identify and detain unauthorized immigrants (Monogan 2013). Between 2005 and 2012, the nine states of the “Nuevo South” proposed 196 restrictive immigration bills that received roll call votes in at least one chamber. Bills covering employment and law enforcement were most common, with every state voting on at least one of each type of bill (table 1). Table 1. Restrictive Immigration Legislation, 2005–2012: Bill Topic by State AL AR GA LA MS NC SC TN VA Total Labor market restrictions (n = 70)  Employment 16 3 10 5 4 1 1 11 7 58  IDs 5 3 2 1 1 0 0 0 0 12 Civil rights restrictions (N = 78)  Police 2 2 3 5 6 3 2 10 9 42  Voting 1 0 4 1 1 1 1 1 2 12  Education 1 3 4 1 1 1 1 0 5 17  Omnibus 1 0 1 0 2 0 1 2 0 7 Other (n = 48)  Benefits 0 1 1 2 2 0 0 4 1 11  Budget 0 0 0 0 1 1 2 0 1 5  Driver’s licenses (nonwork) 2 2 0 2 2 1 1 1 2 13  Safety 0 1 4 0 0 0 0 0 3 8  Misc 0 0 2 0 1 1 0 4 3 11 Total 28 15 31 17 21 9 9 33 33 196 AL AR GA LA MS NC SC TN VA Total Labor market restrictions (n = 70)  Employment 16 3 10 5 4 1 1 11 7 58  IDs 5 3 2 1 1 0 0 0 0 12 Civil rights restrictions (N = 78)  Police 2 2 3 5 6 3 2 10 9 42  Voting 1 0 4 1 1 1 1 1 2 12  Education 1 3 4 1 1 1 1 0 5 17  Omnibus 1 0 1 0 2 0 1 2 0 7 Other (n = 48)  Benefits 0 1 1 2 2 0 0 4 1 11  Budget 0 0 0 0 1 1 2 0 1 5  Driver’s licenses (nonwork) 2 2 0 2 2 1 1 1 2 13  Safety 0 1 4 0 0 0 0 0 3 8  Misc 0 0 2 0 1 1 0 4 3 11 Total 28 15 31 17 21 9 9 33 33 196 Table 1. Restrictive Immigration Legislation, 2005–2012: Bill Topic by State AL AR GA LA MS NC SC TN VA Total Labor market restrictions (n = 70)  Employment 16 3 10 5 4 1 1 11 7 58  IDs 5 3 2 1 1 0 0 0 0 12 Civil rights restrictions (N = 78)  Police 2 2 3 5 6 3 2 10 9 42  Voting 1 0 4 1 1 1 1 1 2 12  Education 1 3 4 1 1 1 1 0 5 17  Omnibus 1 0 1 0 2 0 1 2 0 7 Other (n = 48)  Benefits 0 1 1 2 2 0 0 4 1 11  Budget 0 0 0 0 1 1 2 0 1 5  Driver’s licenses (nonwork) 2 2 0 2 2 1 1 1 2 13  Safety 0 1 4 0 0 0 0 0 3 8  Misc 0 0 2 0 1 1 0 4 3 11 Total 28 15 31 17 21 9 9 33 33 196 AL AR GA LA MS NC SC TN VA Total Labor market restrictions (n = 70)  Employment 16 3 10 5 4 1 1 11 7 58  IDs 5 3 2 1 1 0 0 0 0 12 Civil rights restrictions (N = 78)  Police 2 2 3 5 6 3 2 10 9 42  Voting 1 0 4 1 1 1 1 1 2 12  Education 1 3 4 1 1 1 1 0 5 17  Omnibus 1 0 1 0 2 0 1 2 0 7 Other (n = 48)  Benefits 0 1 1 2 2 0 0 4 1 11  Budget 0 0 0 0 1 1 2 0 1 5  Driver’s licenses (nonwork) 2 2 0 2 2 1 1 1 2 13  Safety 0 1 4 0 0 0 0 0 3 8  Misc 0 0 2 0 1 1 0 4 3 11 Total 28 15 31 17 21 9 9 33 33 196 While states outside the South have enacted similar measures, the historical and political context of racial politics in the South is both similar to and different from other regions of the country (Monogan 2013; Wilkinson 2015). In the new immigrant destination states of the South, African Americans hold a large number of seats in state legislatures among Democrats, while only four of the 503 Democratic legislators are Latino. And the Latino population, while growing, remains small, ranging from 3 percent (MS) to 9 percent (GA, NC) (Pew 2013). Across the social science literature, support for immigration in the United States is read as support for “Latino issues” (Hero and Preuhs 2013). Conversely, support for restricting immigration is interpreted as opposition to Latino issues and interests (Ramakrishnan and Wong 2010). The literature on public and elite positions toward immigration legislation thus encompasses attitudes toward Latinos as well as toward immigrants and immigration policy (Hainmueller and Hopkins 2014). Empirical and Theoretical Background The few published studies on the positions that African American legislators take on restrictive immigration legislation report mixed results. In his historical overview of Black elite stances toward immigration policy, Diamond observes, “[f]or almost two centuries, blacks have been torn on the issue of immigration” (1998, 247). Unlike Diamond (1998), Hero and Preuhs (2013) find consistent “commonalities” or “independent” responses to immigration policy at the national level with no evidence of competition stances.3 The authors emphasize that level of government matters; unlike national policies, restrictive immigration legislation might be perceived as zero-sum by Black politicians at the state or local level. Indeed, some research suggests that anti-immigrant responses do appear to be common at the local level. In research on politics in school boards and local jurisdictions, African American leaders support restrictive immigration policy (see Hero and Preuhs [2013] for a review). In one of the few studies of Black elite positions on state immigration policy, Williams and Hannon (2016) found that in 2007, Alabama’s African American politicians were neutral on immigration policy, claiming that immigration issues did not affect their constituency. By 2013, however, these Black lawmakers overwhelmingly opposed Alabama’s omnibus immigration bill, HB6. The authors argue that the shift occurred because Black elites framed the punitive omnibus bill as a civil rights issue. However, we do not know whether this pattern holds for state immigration policies across the full spectrum of issues or topics. Public opinion surveys also indicate that local demographics and the specific policy influence Black attitudes toward immigration—and their differences with Whites (Masuoka and Junn 2013, 171). The clearest pattern across surveys emerges when jobs are at stake. The majority of Blacks support restrictive employment-related policies such as sanctioning employers who hire undocumented immigrants or requiring IDs proving legal status for employment (Masuoka and Junn 2013, 172) or when their relative economic position is lower than their Latino neighbors (Gay 2006). There is some evidence that when the policy involves the civil rights issue of education, Black opinion is more sympathetic to immigrants (Masuoka and Junn 2013). For both labor market and civil rights issues, Black support for restrictive immigration policy is either equivalent to or more liberal (less restrictionist) than White opinion (Masuoka and Junn 2013, 172). Given the mixed findings in the literature, it is not clear whether African American Democratic legislators in the “Nuevo South” would support or oppose restrictive immigration legislation more frequently than White Democratic legislators. We address this question, and also draw upon group threat theories and theories of symbolic politics to identify conditions under which legislative behavior toward restrictive immigration legislation will differ for African American and White Democrats. Group Threat Group threat theory represents the most thoroughly tested explanation of attitudes toward restrictive immigration legislation (Ceobanu and Escandell 2010). According to Blumer (1958), perceptions of race-based threat are activated when members of one group on a race/ethnic hierarchy perceive that a group below them on the hierarchy is gaining resources in a zero-sum situation. Thus, if Blacks see Latinos—and by extension immigrants—as a growing population that is overtaking the political and social position of Blacks as the predominant racial/ethnic minority, African American Democratic legislators (and their constituents) may perceive immigrants as more of a threat than White legislators do. Other scholars of group threat theory assume that the main source of threat derives from labor market competition, and that the perception of threat reflects a “realistic” appraisal of that competition (Bobo and Hutchings 1996; Hopkins 2010). Studies have found that different sectors of the African American population hold different views of Latinos and/or immigrants, depending upon whether they are more or less likely to compete with Latinos or immigrants. Working-class Blacks embedded in labor market sectors where they are most likely to compete with Latino immigrants express views consistent with group threat theory (Gay 2006; McDermott 2011). Poll data from 2006 indicate that working-class Blacks were more likely to hold anti-immigrant attitudes compared to middle-class Blacks; however, the Black class differences appear to have disappeared by 2013 (Nteta 2013). In addition to economic threat, public concern over a growing cultural threat would arise if the increasing number of immigrants were perceived as transforming “the American way of life” through language and customs (Brown 2013; Hainmueller and Hopkins 2014). Usually, it is Latinos who are depicted as “the immigrant cultural threat” (Chavez 2013). This could trigger perceptions of threat among both Blacks and Whites equally, especially if both groups experience a “cultural threat” from news reports of a rapidly growing Latino population or from the increased visibility of Spanish-language establishments and services (Abascal 2015; Hainmueller and Hopkins 2014). Several scholars find that it is Whites who are more likely to support restrictive immigration legislation compared to Blacks (Citrin et al. 1997; Hutchings and Wong 2014; Masuoka and Junn 2013). The authors suggest that although both Whites and Blacks perceive that the growing number of Latino immigrants are posing a political and cultural threat, Whites are also driven by racial prejudice to a greater extent than Blacks are (Hutchings and Wong 2014). Tests of group threat theory often measure threat through variation in the number of Latinos or immigrants within a local community (Blalock 1967; Blumer 1958). Refinements of this theory suggest that more than the “raw percent” of Latinos or immigrants, perceptions of threat will be most acute in places where the Latino or immigrant population grows rapidly. For instance, Hopkins (2010, 41) finds that a rapid increase in the number of immigrants represents the single strongest predictor of whether a county considers an anti-immigrant ordinance. At the state level, Monogan (2013) also finds that states with faster-growing foreign-born populations are more likely to adopt restrictive immigration policies. Thus, even though Blacks and Whites far outnumber immigrants in the legislative districts of the “Nuevo South” that we analyze, the remarkable pace of immigration to the South could generate perceptions of competition or threat. Further, some scholars contend that perceived cultural threat is the strongest determinant of negative attitudes toward immigration policy, suggesting that rapid rises in the local immigration population are particularly strong triggers for threat. Indeed, Hainmueller and Hopkins (2014) argue that rather than economic self-interest, anti-immigrant attitudes are based on the framing of immigration issues and the perceived symbolic threat that immigrants pose to “the nation.” Race and Symbolic Politics Theories of symbolic politics extend beyond material self-interest to explain political stances toward immigration. These theories posit that legislators and their constituents will support or oppose policy depending on how the policy symbolically aligns with their core values and ideologies (Citrin et al. 1990). For African Americans, racial identity and national identity represent two of the most salient “symbolic predispositions” that influence attitudes toward immigration policy (Abascal 2015; Citrin et al. 1990; Masuoka and Junn 2013). Masuoka and Junn (2013, 2) argue that their lower position in the US racial hierarchy will lead Blacks to develop a distinct group identity and sense of belonging, which in turn influences their attitudes toward immigration. The authors find that African Americans oppose restrictive immigration policy when they have a strong Black identity and they believe that Blacks and Latinos share a common goal of ending discrimination and racial oppression (see also Wilkinson [2015]; Wilkinson and Bingham [2016]). Kaufmann describes this perception of “shared symbolic attachments” as the “symbolic glue” that is necessary for Blacks to perceive commonalities with Latinos (2003, 199–210). However, the theory of symbolic politics suggests that it is not simply a general perception of common goals and values between Latinos and Blacks that will generate Black alliance with Latinos in their opposition to restrictive immigration legislation. Rather, Black legislative responses will vary “depending upon the particular symbols that become associated with a proposal—that is, on how the issue is symbolically framed” (Citrin et al. 1990, 1126). We propose that bill topic captures this “symbolic framing” of immigration issues. Because African Americans hold a lower place on the racial hierarchy than Whites and a shared history of struggle against racial oppression, the symbolic significance of certain topics will differ for Black Democratic legislators than for White Democratic legislators. In particular, bill topics involving civil rights issues may be particularly likely to elicit Black commonality votes in solidarity with immigrants. Representation in Institutional Context If racial group identity is central in shaping the African American values and goals that can produce the necessary “symbolic glue” for commonalities responses, then both legislator race and the racial composition of their constituency may lead Black Democratic legislators to vote differently on immigration legislation compared to White Democratic legislators. This is particularly important in our nine “Nuevo South” states, where White Democratic legislators may have a fairly substantial number of Black constituents in their district.4 However, it is not clear that these attitudes among the Black population would translate directly onto Black legislator roll call votes on immigration bills. First, audit studies show that attitudes can differ from their behavior (Pager and Quillian 2005). Second, as we discuss below, Black legislators face a different set of institutional conditions compared to their Black constituencies. Third, there is evidence that race influences the political behavior of Black representatives independent of the preferences of their Black constituencies, and that Black elite views on immigration diverge from the attitudes of their constituencies (McClain et al. 2008; Tate 2003). African American Democratic legislators appear to be more liberal than their Black constituents, particularly on issues such as welfare reform, crime control, and illegal immigration (Griffin and Newman 2008; Tate 2003). Numerous studies of legislative roll call voting also show that Black representatives are significantly more liberal than their White Democratic counterparts, even controlling for constituency characteristics (Canon 1999; Grose 2011; Juenke and Preuhs 2012). Our analyses allow us to consider whether and how the race of African American Democratic legislators produces an independent effect on roll call voting above and beyond the influence of their Black constituencies and in comparison to White Democrats who also represent Black constituencies. State legislators not only respond to their constituency and their individual political preferences; they also face pressure from other stakeholders in the immigration debate (Gonzalez and Kamdar 2000; Monogan 2013). Historically, White Democratic legislators in the South (“Dixiecrats”) and their White constituencies were more conservative than White Democrats in states outside the South (Shafer and Johnston 2009). However, this ideological difference has narrowed since the 1960s, as conservative Southern Whites have moved to the Republican Party (McCarty, Poole, and Rosenthal 2016). Currently, party polarization defines the ideological split throughout the United States—including the South (Thurber and Yoshinaka 2015). Nonetheless, these Southern Democrats are relatively conservative overall (Shor and McCarty 2011). Black Democratic state legislators are embedded within a complex institutional field that includes the Democratic Party as well as organized groups with interests in their district and state; the context of office-holding in their state (such as term limits and party control); and their financial backers (Clawson 1989; Gonzalez and Kamdar 2000; Monogan 2013; Nicholson-Crotty and Nicholson-Crotty 2011). Studies show that political factors such as the existence of a Republican majority in a city’s electorate (Ramakrishnan and Wong 2010) or conservative political ideology among citizens (Chavez and Provine 2009; Nicholson-Crotty and Nicholson-Crotty 2011) appear to better predict restrictive immigration ordinances at the state and local level compared to district demographics. Unlike White Democratic state legislators, Black state legislators may also attend to African American advocacy groups such as the National Association for the Advancement of Colored People (NAACP) and the National Black Caucus of State Legislators (NBCSL). Both of these organizations have publicly opposed state and federal immigration policy that expands police powers, including restrictive “omnibus” legislation in such states as Arizona and Alabama, with “commonalities” claims that these policies encourage police to engage in racial profiling (NAACP 2011; NHCSL 2010). Finally, Black Democratic legislators also must consider the context in which a bill is proposed. They might support an immigration bill in exchange for receiving support on a bill that they are sponsoring, for instance. The presence of a Latino caucus also influences Black elite stances toward immigration policy (Hero and Preuhs 2013). Importantly, our “Nuevo South” study looks at how Black Democratic legislators vote when restrictive immigration bills are on the table in uniformly conservative states when they are free from pressure from a Latino caucus or a relatively large Latino constituency. This allows us to more clearly identify the threat and symbolic factors driving Black competition and commonalities responses. We argue that in the absence of elected and organized Latinos, the most proximate influence shaping a “competition” or a “commonalities” stance among African American Democratic legislators is the topic of the bill itself—once political institutional factors are controlled. Hypotheses Theories of group threat and symbolic politics lead to the following predictions about the relation between bill topics and legislative responses: H1: When the restrictive immigration bill topic is likely to evoke considerations of interracial group threat and competition, African American Democratic legislators will be equally likely as or more likely than White Democratic legislators to support the measure. Specifically, bill topics related to employment will likely evoke considerations of group threat and competition among both White Democratic and Black Democratic legislators H2: When the bill topic is likely to evoke considerations of Black-immigrant commonalities, African American Democratic legislators will be less likely than White Democratic legislators to support the measure. Specifically, bills on civil rights–related topics will elicit a “commonalities” response either by directly restricting African American access to social or political resources (e.g., voter ID laws) and/or by symbolically reflecting shared goals and values between Blacks and Latinos. We also expect that the underlying social and political mechanisms of group threat depend on (or are conditioned by) the bill topic and race of legislator combined: H3: When the bill topic concerns labor market competition, the conditions that prompt threat (i.e., pace of change in the district Latino population and district labor market characteristics) will have an equal or stronger effect on increasing support among African American Democratic legislators compared to White Democratic legislators. H4: When the bill topic concerns civil rights issues, then the conditions that prompt threat will have stronger effects on increasing support among White Democratic legislators compared to African American Democratic legislators. This is because African American Democratic legislators have the countervailing considerations of common interests and civil rights legacies that White legislators are not likely to share. Data and Methods Our dataset includes final votes on all proposed restrictive immigration legislation that received a roll call vote in at least one full chamber for 2005–2012, combining 182 bills that were enacted with 14 bills that were vetoed by the governor, died in committee, or failed to pass in at least one chamber. Hopkins (2010) identifies 2005 as marking the beginning of widespread state and local anti-immigration policymaking. The political context of immigration reform arguably changed after 2012, when the re-election of Barack Obama spurred a renewed effort toward federal immigration reform. We code bills as “restrictive” if they aim to exclude or restrict immigrants from access to resources (jobs, education, housing, etc.), increase surveillance of immigrants and enforcement of immigration laws (expanding police authority), or reinforce the exclusion of immigrants from the polity (voter ID). Bills that refine, clarify, or provide funds for existing restrictive laws are also included in our sample (e.g., participating in the e-verify program or specifying the type of ID that is valid for voting). We exclude resolutions from our sample, as they are not laws and often are not subject to recorded votes. We construct a comprehensive dataset to assess the effects of Democratic legislators’ race, the topic of each bill, and the legislator’s district characteristics on support for restrictive immigration legislation. Our final dataset excludes cases with missing values on our analysis variables, and consists of 998 Democratic legislators casting 12,920 votes on 196 bills. Race of Legislator We compile information on all state legislators serving from 2005 to 2012 in our nine states, coding each legislator’s name, party, chamber, and race. Race is coded through a series of steps, which include referring to the list of African American state legislators published by the National Conference of State Legislators (NCSL) in 2009, examining the organizations and educational institutions with which the legislator was affiliated, and visual inspection of legislator photos. We cross-checked the legislator race codes with a database of Black Democratic state representatives provided by the Joint Center for Political and Economic Research. The five Democratic legislators who are neither Black nor non-Hispanic White or whose race could not be determined are excluded from our analyses. Votes Following Facchini and Steinhardt (2011), we measure support for restrictive immigration legislation by coding final roll call votes on a bill. “Yea” votes were coded 1. “Nay” votes or votes that were recorded as “present,” “abstain,” or “pass” were coded as 0. We coded votes as “missing” for legislators who were recorded as absent when the final vote was taken for a bill. Bill Topic We used the annual reports published by the NCSL to develop a list of all of the restrictive immigration-related legislation enacted from 2005 to 2012 in our nine states. To find the bills that were not enacted, we conducted a search in Lexis/Nexis Statenet, using the same terms used by the NCSL. Only bills that met our criteria for “restrictive” (see above) are included in our final dataset. We code bill topics using the NCSL’s 11 topic codes (see table 1), with further recoding to test our hypotheses. Specifically, we subdivide the “ID/licenses” bills into three categories: 1) licenses and IDs related to employment (driver’s licenses, professional licenses, and business licenses); 2) driver’s licenses not specifically related to employment; and 3) firearms licenses (public safety). The few licenses that do not fall into these categories (e.g., fishing licenses) are included among the “miscellaneous” topics. Bills pertaining to budget items are coded according to the activity targeted by the budget bill. For instance, a bill allocating funds to support the e-verify program is coded under the employment topic. For each restrictive immigration bill in our dataset, we record the roll call vote of each state legislator. We differentiate the bill topics into four groups with respect to their relevance for competition or commonalities issues: the first group of topics (“competition”) involves labor market bills that restrict employment of undocumented immigrants. These include bills that both explicitly target labor market practices of employers or workers and bills that increase the difficulty of working through prohibiting work-related driver’s licenses, professional licenses, and business licenses.5 We categorize proposals regarding voter-identification, expanding police powers, education, and omnibus restrictions as “civil rights” bills likely to evoke a “commonalities” response among African American Democratic legislators. In addition to the continued struggles of civil rights groups to renew and implement the Voting Rights Act of 1965, poor African American voters are over-represented among those without identification to vote. Similarly, concerns about police brutality and racial profiling remain central for African American civil rights organizations (NAACP 2014). Education emerges as the most salient concern for African American Congressional legislators across all issues listed on the NAACP scorecard (Hero and Preuhs 2013, 105, figure 4.I). The final set of topics do not relate to either competition or commonalities frameworks. These include bills on the topics of welfare benefits, health benefits, budget bills, and miscellaneous bills (e.g., fishing licenses, refugee settlement). We also classified “public safety” bills under “other.” This topic operates differently than the labor market threat bills; immigrants are depicted as a “threat to public safety” among Whites but not among Blacks. Further, bills classified under “public safety” appeared in only three states (table 1), so we could not conduct analyses with these bills as a separate category. District Characteristics To test our hypotheses that bill topic conditions the effect of threat-related district characteristics on support for restrictive immigration legislation, we construct multiple measures at the state legislative district level. We calculate the rate of change in the Latino population ([% Latino 2010 – % Latino 2000]/% Latino 2000) within each legislative district using data from the 2000 and 2010 decennial censuses.6 We use change in % Latino rather than % immigrant because the latter is not available in the block-level summary data that we use to construct the district-level change measure. However, we are confident that the change in % Latino captures the dynamics described in the group threat literature. In addition to the high correlation between the percent immigrant and percent Latino variables in our model, there is much evidence to suggest that Latinos are commonly stereotyped as “immigrants”—often “illegal immigrants” (Abascal 2015; Brown 2013). Thus, a rapid increase in the Latino population would be perceived as a rapid increase in the % immigrant. We construct additional measures of the demographic and economic characteristics of the legislator’s district from the American Community Survey aggregated summary data for 2005–2009, 2006–2010, 2007–2011, 2008–2012, and 2009–2013. (Only the aggregated five-year ACS provides data at the state legislative district level.) Where possible, we match the year of the roll call vote with the mid-year of the five-year ACS data. That is, we merge the 2007 legislator data with the 2005–2009 ACS, the 2008 legislator data with the 2006–2010 ACS, the 2009 legislator data with the 2007–2011 ACS, and so on. Due to the availability of the five-year ACS, we merge the 2005 and 2006 legislation data with the 2005–2009 ACS and we merge the 2012 legislator data with the 2009–2013 ACS. These ACS variables are treated as time variant in the models, given that the district demographics change for legislators who changed districts between 2005 and 2012 and all legislators who held office long enough. We construct three district labor market measures from the five-year ACS. The district unemployment rate represents the general economic instability among constituents in our model. To estimate the proportion of the district constituency who may perceive direct labor market competition from immigrants, we calculate the percent of the district workforce employed in the five most common occupations for immigrants: construction, building/grounds cleaning and maintenance, food preparation, production, and transportation (Motel and Patten 2013). We exclude the percent agricultural workers from the immigrant occupation measure, as it is highly correlated with % Latino and % immigrant. Instead, we include a separate measure for the % working in the agriculture industry. This captures both the farmworker presence as well as the potential pressure on legislators from local farmers and agriculture industry lobbies (Nicholson-Crotty and Nicholson-Crotty 2011). Controls To differentiate the pace of change in the Latino population from overall increases in the district’s non-Latino populations, we use the 1990 and 2000 decennial census to construct controls for the change in the Black population and the change in the logarithm of the population. We also include the percent Latino from the ACS as a baseline for the change in the % Latino variable as well as a proxy of the % immigrant. Models using the percent foreign born rather than % Latino yield similar results.7 To gauge the influence of legislator race above and beyond the racial composition of their constituency, we construct an ACS measure of the % Black voting age—Black citizens age 18 and older. We also use the ACS five-year files to create controls for logged median family income. We control for effects of state-level partisan politics by including three measures: a dummy variable indicating whether the Republican Party held a majority in both chambers and the governorship (unified) using data from Monogan (2013) (https://dataverse.harvard.edu/dataverse/monogan); a measure of party polarization within the legislator’s chamber; and a measure of Democratic party cohesion within the chamber. The polarization and cohesion measures are provided by Shor and McCarty (2011) (https://americanlegislatures.com/data/). The Shor and McCarty dataset contains missing values for 2009 and 2010 in Arkansas and Tennessee, and for 2009 in Virginia. We use the state data for 2008 to impute these missing values.8 These party controls account for the possibility that restrictive immigration measures that come to a vote in Republican-controlled states are less likely to gain (or need) Democratic support than those voted on in Democratic-controlled or split states (Chavez and Provine 2009; Monogan 2013; Ramakrishnan and Wong 2010). In addition, legislators may feel more political pressure from their party in states in which there is relatively strong party polarization or cohesion (Pruehs 2006).9 We include fixed effects for each state to account for other state institutional factors that could affect support for restrictive immigration legislation (Monogan 2013; Ramakrishnan and Wong 2010). Estimation To examine how legislators’ race affects support for restrictive immigration bills, our units of analysis are legislator votes on restrictive immigration bills. As legislators often vote on multiple bills during their term, we use a multilevel model for binary responses that accounts for the clustered nature of our data (Guo and Zhao 2000; Raudenbush, Fotiu, and Cheong 1999). Our model consists of two levels: the vote-level model (level 1) that includes all time-variant factors such as district demographics, and a legislator-level model (level 2) that accounts for time-invariant characteristics such as legislators’ race. To assess whether legislator race affects voting depending on the bill topic, we model cross-level interaction effects between legislator race and bill topic. A multilevel approach enables us to distinguish between within- and between-legislator errors, therefore estimates effects more efficiently than an OLS regression would. Moreover, HLMs are better able to deal with unbalanced panels and situations where units have only few observations (i.e., legislators for whom we only observe votes on few bills) (Raudenbush and Bryk 2002). We pool all votes cast between 2005 and 2012 and include state and year fixed effects, to control for geographic and temporal variation. We provide a more detailed explanation of our HLM estimation in the Supplemental Methods Appendix. Hypotheses Testing To test hypotheses 1 and 2, we estimate the effect of legislator race, bill topic, district demographic characteristics, and our controls on support for restrictive immigration bills. To estimate whether the effect of legislator race on support for restrictive immigration legislation depends on whether the bill covers employment issues or civil rights–related issues, we run the model interacting legislator race with bill topic. To test hypotheses 3 and 4, we estimate separate models for the employment bills and the civil rights bills. As we explain above, employment topics include bills focused on restricting immigrant labor market access; civil rights topics include legislation increasing surveillance of immigrants, or restricting immigrant movement or access to resources in relation to the topics of law enforcement, voting, educational opportunities, as well as the omnibus bills. For each bill-topic subsample, hypotheses 3 and 4 are tested with interactions between legislator race and district characteristics. In addition to the potential for confounding coefficients with residual variation in logistic regression, the nonlinear functional form of the logistic regression model creates a situation in which interactions can be significant within some ranges of the independent variable but not in others. To address these issues, we follow the procedures developed by Long (Long and Freese 2014). For each bill-topic subsample, we estimate (AME) marginal probabilities with the logits generated in a model with interactions of legislator race with the district demographic variables (Supplemental Table iv). We fix the independent variable of interest at the mean and at additional values above and below the mean, holding the other independent variables at their observed values for each observation (Long and Freese 2014). We use the delta method to test for whether the probabilities for these two “conditions” are statistically significant. A full explanation of the estimation procedure is available in Long and Freese (2014). It is not possible to conduct statistical tests for differences in the coefficients across models, as the coefficient can be confounded by residual variation in logistic regression (Long and Freese 2014) and because our samples are not independent. In additional analyses, we therefore combine the employment and civil rights subsamples and test for significant three-way interactions between bill topic (employment or civil rights), legislator race (Black or White), and each of the independent variables. These analyses confirm the findings we report below (results available upon request). Results Descriptive Analyses Although the majority of Blacks favor restrictive immigration legislation (71 percent of Black votes are “yea”), this is significantly lower than the 84 percent of “yea” votes among their White Democratic counterparts (table 2). The literature suggests that the majority of White and Black representatives support restrictive immigration legislation for several reasons. First, only bills that will be enacted are typically brought up for a vote. Second, in the post-9/11 era, restrictive immigration legislation is popular among Democratic voters (Masuoka and Junn 2013). Third, in our nine Southern states, there are no organized groups of Latino legislators or sizeable concentrations of Latino constituents motivating Democratic representatives to oppose restrictive immigration legislation. Indeed, it is precisely this freedom from Latino influence that provides an important aspect of our paper. Table 2. Percent “Yea” Votes on Restrictive Immigration Legislation by Legislator Race and Bill Topic Black Dems White Dems GOP Bl-Wh Dems: Chi-sq, df Wh Dems-GOP: Chi-sq, df Bill topic Labor market  Employment 77% 87% 93% 60.5, 1*** 68.82,1***  Licenses (work-related) 84% 90% 90% 8.01, 1** 0.002,1 Civil rights  Police 76% 89% 93% 83.5, 1*** 21.5,1***  Voting 32% 49% 97% 19.2, 1*** 333.6,1***  Education 71% 85% 96% 36.3, 1*** 42.7,1***  Omnibus 34% 58% 89% 27.5, 1*** 102.3,1*** Other topics  Benefits 72% 77% 92% 2.3, 1 76.7,1***  Budget 63% 77% 86% 9.3, 1** 6.3,1*  DL’s 81% 87% 92% 5.9, 1* 8.9,1**  Safety 67% 89% 94% 33.9, 1*** 10.2,1**  Misc 65% 76% 90% 11.3, 1 43.4,1*** Total 71% 84% 93% 307.5, 1*** 401.4,1***  N (total) 5,303 7,617 12,884 Black Dems White Dems GOP Bl-Wh Dems: Chi-sq, df Wh Dems-GOP: Chi-sq, df Bill topic Labor market  Employment 77% 87% 93% 60.5, 1*** 68.82,1***  Licenses (work-related) 84% 90% 90% 8.01, 1** 0.002,1 Civil rights  Police 76% 89% 93% 83.5, 1*** 21.5,1***  Voting 32% 49% 97% 19.2, 1*** 333.6,1***  Education 71% 85% 96% 36.3, 1*** 42.7,1***  Omnibus 34% 58% 89% 27.5, 1*** 102.3,1*** Other topics  Benefits 72% 77% 92% 2.3, 1 76.7,1***  Budget 63% 77% 86% 9.3, 1** 6.3,1*  DL’s 81% 87% 92% 5.9, 1* 8.9,1**  Safety 67% 89% 94% 33.9, 1*** 10.2,1**  Misc 65% 76% 90% 11.3, 1 43.4,1*** Total 71% 84% 93% 307.5, 1*** 401.4,1***  N (total) 5,303 7,617 12,884 Note: Unit of analysis is the individual vote cast. Because the number of bills and legislators vary by state, the n’s vary by topic. The number of votes cast by Black Democratic legislators range from 163 (budget) to 1,544 (employment). The number of votes cast by White legislators range from 198 (budget) to 2,164 (employment). The range for Republican votes is 282 (budget) to 5,343 (employment). Of the 1,360 Republican legislators casting votes in our sample, all but eight are White. ***p < 0.001 **p < 0.01 *p < 0.05 Table 2. Percent “Yea” Votes on Restrictive Immigration Legislation by Legislator Race and Bill Topic Black Dems White Dems GOP Bl-Wh Dems: Chi-sq, df Wh Dems-GOP: Chi-sq, df Bill topic Labor market  Employment 77% 87% 93% 60.5, 1*** 68.82,1***  Licenses (work-related) 84% 90% 90% 8.01, 1** 0.002,1 Civil rights  Police 76% 89% 93% 83.5, 1*** 21.5,1***  Voting 32% 49% 97% 19.2, 1*** 333.6,1***  Education 71% 85% 96% 36.3, 1*** 42.7,1***  Omnibus 34% 58% 89% 27.5, 1*** 102.3,1*** Other topics  Benefits 72% 77% 92% 2.3, 1 76.7,1***  Budget 63% 77% 86% 9.3, 1** 6.3,1*  DL’s 81% 87% 92% 5.9, 1* 8.9,1**  Safety 67% 89% 94% 33.9, 1*** 10.2,1**  Misc 65% 76% 90% 11.3, 1 43.4,1*** Total 71% 84% 93% 307.5, 1*** 401.4,1***  N (total) 5,303 7,617 12,884 Black Dems White Dems GOP Bl-Wh Dems: Chi-sq, df Wh Dems-GOP: Chi-sq, df Bill topic Labor market  Employment 77% 87% 93% 60.5, 1*** 68.82,1***  Licenses (work-related) 84% 90% 90% 8.01, 1** 0.002,1 Civil rights  Police 76% 89% 93% 83.5, 1*** 21.5,1***  Voting 32% 49% 97% 19.2, 1*** 333.6,1***  Education 71% 85% 96% 36.3, 1*** 42.7,1***  Omnibus 34% 58% 89% 27.5, 1*** 102.3,1*** Other topics  Benefits 72% 77% 92% 2.3, 1 76.7,1***  Budget 63% 77% 86% 9.3, 1** 6.3,1*  DL’s 81% 87% 92% 5.9, 1* 8.9,1**  Safety 67% 89% 94% 33.9, 1*** 10.2,1**  Misc 65% 76% 90% 11.3, 1 43.4,1*** Total 71% 84% 93% 307.5, 1*** 401.4,1***  N (total) 5,303 7,617 12,884 Note: Unit of analysis is the individual vote cast. Because the number of bills and legislators vary by state, the n’s vary by topic. The number of votes cast by Black Democratic legislators range from 163 (budget) to 1,544 (employment). The number of votes cast by White legislators range from 198 (budget) to 2,164 (employment). The range for Republican votes is 282 (budget) to 5,343 (employment). Of the 1,360 Republican legislators casting votes in our sample, all but eight are White. ***p < 0.001 **p < 0.01 *p < 0.05 However, as our separate analyses by topic show, these aggregate numbers gloss over a more complex picture of Black-White differences in support for restrictive immigration legislation among Democrats (table 2). African American Democratic legislators are most likely to support “competition” bills restricting immigrant labor market access and are least likely to support voter ID, a civil rights topic. Republican support for restrictive immigration is significantly higher than White Democratic support on almost all bill topics; overall, 93 percent of Republicans support restrictive immigration bills (table 2). Importantly, while omnibus bills garner high support from Republicans (89 percent), these bills receive the second lowest level of support from Democrats (34 percent of Blacks, 58 percent of Whites). Thus, restricting our multivariate analyses to Democrats provides a conservative estimate of Black-White legislative differences in responses to state immigration legislation (1,435 Republicans in our sample years are White; only four Republicans are African American, one Republican is Latino, and three Republicans are Asian). To assess the underlying assumption in theories of symbolic politics that African American Democratic opposition to restrictive immigration legislation will be based on common values and goals beyond material group interest, we coded the civil rights bills according to whether opposing the bill would be in the direct group interest of African Americans. Bills negatively affecting African American group interest such as requiring voter ID were coded as 1 for self-interest; bills that would not negatively affect African Americans directly, such as requiring that noncitizens pay out-of-state tuition, were coded as 0 (see Supplemental Table i for examples). Approximately 80 percent of the restrictive bills covering civil rights–related topics did not negatively affect African Americans. Thus, we interpret the African American “commonalities” response—that is, their lack of support for these bills on civil rights topics—as motivated by the symbolic importance of civil rights–related issues, as the theory of symbolic politics would predict. Descriptive analyses show that district characteristics vary among Democratic legislators along dimensions described in the literature (Hopkins 2010) (we report these results in Supplemental Table ii). Black Democrats represent majority-Black districts, with the average black voting age population comprising 56 percent of their district population. Importantly for our study, many White Democrats also represent districts with Black constituents. On average, voting-age Blacks comprised 23 percent of the population in districts represented by White Democrats. In addition, both the Black and White Democratic legislators saw a doubling of % Latino in the district between 2000 and 2010. We turn to the multivariate analyses to understand how these district characteristics along with bill topic affect Black-White differences in Democratic support for restrictive immigration legislation. Multivariate Analyses Our multivariate model in table 3 shows that Black Democratic legislators are significantly less likely than their White Democratic peers to support restrictive immigration legislation regardless of topic, even with controls for relevant district characteristics and political context. When we consider Black-White differences by bill topic, however, a more complex picture emerges. Table 3. Effect of Legislator Race, Bill Topic, and District Characteristics on the Log Odds of Supporting Restrictive Immigration Legislation logit s.e. Black legislator −0.341 0.90*** Bill topics (Employment = omitted category) Labor market topics  Licenses/IDs 0.665 0.126*** Civil rights topics  Police −0.260 0.085**  Voting −1.512 0.118***  Education 0.167 0.101  Omnibus −1.871 0.130*** Other topics  Public benefits −1.225 0.128***  Budget −0.046 0.150  Public safety 0.517 0.142***  Misc −0.906 0.114*** District characteristics  % Chg Latino pop 0.088 0.044*  % Latino −1.161 0.534*  % Black −0.499 0.243*  % Chg Black pop −0.001 0.001  % Immigrant occ 2.427 0.940**  % Agr industry 3.958 1.935*  % Unemployed −2.435 1.467  Ln Faminc −0.016 0.209  Chg ln pop −0.191 0.360 Political context  Unified Republican govt. −1.664 0.134***  Party polarization (chamber) 0.016 0.316  Party cohesion (chamber) 0.593 0.456 State fixed effects Yes Year fixed effects Yes N (votes) 12,920 N (legislators) 999 BIC 10,363.6 logit s.e. Black legislator −0.341 0.90*** Bill topics (Employment = omitted category) Labor market topics  Licenses/IDs 0.665 0.126*** Civil rights topics  Police −0.260 0.085**  Voting −1.512 0.118***  Education 0.167 0.101  Omnibus −1.871 0.130*** Other topics  Public benefits −1.225 0.128***  Budget −0.046 0.150  Public safety 0.517 0.142***  Misc −0.906 0.114*** District characteristics  % Chg Latino pop 0.088 0.044*  % Latino −1.161 0.534*  % Black −0.499 0.243*  % Chg Black pop −0.001 0.001  % Immigrant occ 2.427 0.940**  % Agr industry 3.958 1.935*  % Unemployed −2.435 1.467  Ln Faminc −0.016 0.209  Chg ln pop −0.191 0.360 Political context  Unified Republican govt. −1.664 0.134***  Party polarization (chamber) 0.016 0.316  Party cohesion (chamber) 0.593 0.456 State fixed effects Yes Year fixed effects Yes N (votes) 12,920 N (legislators) 999 BIC 10,363.6 ***p < 0.001 **p < 0.01 *p < 0.05 Table 3. Effect of Legislator Race, Bill Topic, and District Characteristics on the Log Odds of Supporting Restrictive Immigration Legislation logit s.e. Black legislator −0.341 0.90*** Bill topics (Employment = omitted category) Labor market topics  Licenses/IDs 0.665 0.126*** Civil rights topics  Police −0.260 0.085**  Voting −1.512 0.118***  Education 0.167 0.101  Omnibus −1.871 0.130*** Other topics  Public benefits −1.225 0.128***  Budget −0.046 0.150  Public safety 0.517 0.142***  Misc −0.906 0.114*** District characteristics  % Chg Latino pop 0.088 0.044*  % Latino −1.161 0.534*  % Black −0.499 0.243*  % Chg Black pop −0.001 0.001  % Immigrant occ 2.427 0.940**  % Agr industry 3.958 1.935*  % Unemployed −2.435 1.467  Ln Faminc −0.016 0.209  Chg ln pop −0.191 0.360 Political context  Unified Republican govt. −1.664 0.134***  Party polarization (chamber) 0.016 0.316  Party cohesion (chamber) 0.593 0.456 State fixed effects Yes Year fixed effects Yes N (votes) 12,920 N (legislators) 999 BIC 10,363.6 logit s.e. Black legislator −0.341 0.90*** Bill topics (Employment = omitted category) Labor market topics  Licenses/IDs 0.665 0.126*** Civil rights topics  Police −0.260 0.085**  Voting −1.512 0.118***  Education 0.167 0.101  Omnibus −1.871 0.130*** Other topics  Public benefits −1.225 0.128***  Budget −0.046 0.150  Public safety 0.517 0.142***  Misc −0.906 0.114*** District characteristics  % Chg Latino pop 0.088 0.044*  % Latino −1.161 0.534*  % Black −0.499 0.243*  % Chg Black pop −0.001 0.001  % Immigrant occ 2.427 0.940**  % Agr industry 3.958 1.935*  % Unemployed −2.435 1.467  Ln Faminc −0.016 0.209  Chg ln pop −0.191 0.360 Political context  Unified Republican govt. −1.664 0.134***  Party polarization (chamber) 0.016 0.316  Party cohesion (chamber) 0.593 0.456 State fixed effects Yes Year fixed effects Yes N (votes) 12,920 N (legislators) 999 BIC 10,363.6 ***p < 0.001 **p < 0.01 *p < 0.05 We test hypotheses 1 and 2 that the Black-White difference in support for restrictive immigration legislation depends on the bill topic by estimating interactions of legislator race with bill topic in table 4. For ease of interpretation, we present the results as predicted probabilities. (See Supplemental Table ii for coefficients used to estimate probabilities.) In support of hypothesis 1, Black and White Democrats are equally likely to vote in favor of immigration bills restricting labor market competition (employment and employment-related ID/licensing topics). In support of hypothesis 2, African American Democratic legislators are significantly less likely than White Democratic legislators to cast “yea” votes for legislation covering the civil rights topics of voting, law enforcement, education, and omnibus restrictions.10 Table 4. Probability of Supporting Restrictive Immigration Legislation by Bill Topic and Legislator Race Black White Bl-Wh diff Bill topics Labor market  Employment 0.793 0.819 −0.026  Licenses (work related) 0.856 0.857 0.001 Civil rights  Police 0.707 0.810 −0.103***  Voting 0.489 0.640 −0.151***  Education 0.767 0.837 −0.070**  Omnibus 0.448 0.565 −0.118*** Other topics  Public benefits 0.620 0.629 −0.008  Budget 0.748 0.810 −0.062  Driver’s license 0.856 0.871 −0.015  Public safety 0.785 0.889 −0.104***  Misc 0.652 0.697 −0.045 N: 12,920 BIC: 10,419.8 Black White Bl-Wh diff Bill topics Labor market  Employment 0.793 0.819 −0.026  Licenses (work related) 0.856 0.857 0.001 Civil rights  Police 0.707 0.810 −0.103***  Voting 0.489 0.640 −0.151***  Education 0.767 0.837 −0.070**  Omnibus 0.448 0.565 −0.118*** Other topics  Public benefits 0.620 0.629 −0.008  Budget 0.748 0.810 −0.062  Driver’s license 0.856 0.871 −0.015  Public safety 0.785 0.889 −0.104***  Misc 0.652 0.697 −0.045 N: 12,920 BIC: 10,419.8 Note: Probabilities estimated from two-level model with interactions of legislator race with bill topic, main effects for district characteristics and unified Republican government, party polarization and party cohesion, and state fixed effects and year fixed effects. (Coefficients reported in Supplemental Table iii.) See text for explanation of how marginal probabilities are calculated. ***p < 0.001 **p < 0.01 *p < 0.05 Table 4. Probability of Supporting Restrictive Immigration Legislation by Bill Topic and Legislator Race Black White Bl-Wh diff Bill topics Labor market  Employment 0.793 0.819 −0.026  Licenses (work related) 0.856 0.857 0.001 Civil rights  Police 0.707 0.810 −0.103***  Voting 0.489 0.640 −0.151***  Education 0.767 0.837 −0.070**  Omnibus 0.448 0.565 −0.118*** Other topics  Public benefits 0.620 0.629 −0.008  Budget 0.748 0.810 −0.062  Driver’s license 0.856 0.871 −0.015  Public safety 0.785 0.889 −0.104***  Misc 0.652 0.697 −0.045 N: 12,920 BIC: 10,419.8 Black White Bl-Wh diff Bill topics Labor market  Employment 0.793 0.819 −0.026  Licenses (work related) 0.856 0.857 0.001 Civil rights  Police 0.707 0.810 −0.103***  Voting 0.489 0.640 −0.151***  Education 0.767 0.837 −0.070**  Omnibus 0.448 0.565 −0.118*** Other topics  Public benefits 0.620 0.629 −0.008  Budget 0.748 0.810 −0.062  Driver’s license 0.856 0.871 −0.015  Public safety 0.785 0.889 −0.104***  Misc 0.652 0.697 −0.045 N: 12,920 BIC: 10,419.8 Note: Probabilities estimated from two-level model with interactions of legislator race with bill topic, main effects for district characteristics and unified Republican government, party polarization and party cohesion, and state fixed effects and year fixed effects. (Coefficients reported in Supplemental Table iii.) See text for explanation of how marginal probabilities are calculated. ***p < 0.001 **p < 0.01 *p < 0.05 To test hypotheses 3 and 4 that district characteristics further condition the relationship between bill topic and legislator race previously examined in table 4, we estimate separate models for the labor market bills and civil rights bills and include an interaction between legislator race and district characteristics for each model. We report the results of the tests for Black-White differences in probabilities in table 5 and figures 1–6. (See Supplemental Table iii for coefficients used to estimate probabilities.) In the graphs, the effects of a selected independent variable differ for Black and White legislators when the confidence intervals (represented by shaded bands) for each group do not overlap. Table 5. Probability of Supporting Restrictive Immigration Legislation, by Demographic Characteristics and Legislator Race: Labor Market Topics and Civil Rights Topics Labor market topics Civil rights topics Black White Bl-Wh diff Black White Bl-Wh diff District Characteristics  % Change Latino pop   0% 0.836 0.866 −0.030 0.698 0.737 −0.039   50% 0.867 0.845 −0.022 0.696 0.753 −0.058**   100% 0.854 0.869 −0.015 0.693 0.770 −0.076***   150% 0.863 0.870 −0.007 0.691 0.785 −0.095***  % Latino   0% 0.861 0.871 −0.010 0.692 0.770 −0.078***   5% 0.856 0.865 −0.009 0.693 0.767 −0.073***   10% 0.851 0.859 −0.008 0.694 0.763 −0.069**  % Black   0% 0.924 0.886 0.048 0.704 0.807 −0.103*   30% 0.880 0.874 0.011 0.696 0.776 −0.080***   75% 0.781 0.854 −0.043* 0.687 0.742 −0.055*  % Immigrant occ   0% 0.733 0.817 −0.086 0.599 0.620 −0.006   20% 0.852 0.860 −0.013 0.689 0.762 −0.070***   40% 0.909 0.928 0.031 0.769 0.869 −0.103**  % Agriculture industry   0% 0.858 0.866 −0.009 0.685 0.754 −0.067**   10% 0.809 0.851 −0.042 0.747 0.837 −0.090   20% 0.751 0.835 −0.083 0.801 0.899 −0.097  % Unemployed   0% 0.807 0.854 −0.047 0.784 0.771 0.136   10% 0.845 0.867 −0.022 0.702 0.766 −0.064**   25% 0.893 0.884 0.009 0.559 0.759 −0.200*** N (votes) 4,618 5,187 N (legislators) 882 888 BIC 3989.2 4458.6 Labor market topics Civil rights topics Black White Bl-Wh diff Black White Bl-Wh diff District Characteristics  % Change Latino pop   0% 0.836 0.866 −0.030 0.698 0.737 −0.039   50% 0.867 0.845 −0.022 0.696 0.753 −0.058**   100% 0.854 0.869 −0.015 0.693 0.770 −0.076***   150% 0.863 0.870 −0.007 0.691 0.785 −0.095***  % Latino   0% 0.861 0.871 −0.010 0.692 0.770 −0.078***   5% 0.856 0.865 −0.009 0.693 0.767 −0.073***   10% 0.851 0.859 −0.008 0.694 0.763 −0.069**  % Black   0% 0.924 0.886 0.048 0.704 0.807 −0.103*   30% 0.880 0.874 0.011 0.696 0.776 −0.080***   75% 0.781 0.854 −0.043* 0.687 0.742 −0.055*  % Immigrant occ   0% 0.733 0.817 −0.086 0.599 0.620 −0.006   20% 0.852 0.860 −0.013 0.689 0.762 −0.070***   40% 0.909 0.928 0.031 0.769 0.869 −0.103**  % Agriculture industry   0% 0.858 0.866 −0.009 0.685 0.754 −0.067**   10% 0.809 0.851 −0.042 0.747 0.837 −0.090   20% 0.751 0.835 −0.083 0.801 0.899 −0.097  % Unemployed   0% 0.807 0.854 −0.047 0.784 0.771 0.136   10% 0.845 0.867 −0.022 0.702 0.766 −0.064**   25% 0.893 0.884 0.009 0.559 0.759 −0.200*** N (votes) 4,618 5,187 N (legislators) 882 888 BIC 3989.2 4458.6 Note: Probabilities estimated from two-level logit models run separately for labor market and civil rights topics, and controlling for unified Republican state govt. and party polarization and party cohesion within the legislator’s chamber, with state and year fixed effects. (Coefficients reported in Supplementary Table iv.) See text for explanation of how marginal probabilities are calculated. ***p < 0.001 **p < 0.01 *p < 0.05. Table 5. Probability of Supporting Restrictive Immigration Legislation, by Demographic Characteristics and Legislator Race: Labor Market Topics and Civil Rights Topics Labor market topics Civil rights topics Black White Bl-Wh diff Black White Bl-Wh diff District Characteristics  % Change Latino pop   0% 0.836 0.866 −0.030 0.698 0.737 −0.039   50% 0.867 0.845 −0.022 0.696 0.753 −0.058**   100% 0.854 0.869 −0.015 0.693 0.770 −0.076***   150% 0.863 0.870 −0.007 0.691 0.785 −0.095***  % Latino   0% 0.861 0.871 −0.010 0.692 0.770 −0.078***   5% 0.856 0.865 −0.009 0.693 0.767 −0.073***   10% 0.851 0.859 −0.008 0.694 0.763 −0.069**  % Black   0% 0.924 0.886 0.048 0.704 0.807 −0.103*   30% 0.880 0.874 0.011 0.696 0.776 −0.080***   75% 0.781 0.854 −0.043* 0.687 0.742 −0.055*  % Immigrant occ   0% 0.733 0.817 −0.086 0.599 0.620 −0.006   20% 0.852 0.860 −0.013 0.689 0.762 −0.070***   40% 0.909 0.928 0.031 0.769 0.869 −0.103**  % Agriculture industry   0% 0.858 0.866 −0.009 0.685 0.754 −0.067**   10% 0.809 0.851 −0.042 0.747 0.837 −0.090   20% 0.751 0.835 −0.083 0.801 0.899 −0.097  % Unemployed   0% 0.807 0.854 −0.047 0.784 0.771 0.136   10% 0.845 0.867 −0.022 0.702 0.766 −0.064**   25% 0.893 0.884 0.009 0.559 0.759 −0.200*** N (votes) 4,618 5,187 N (legislators) 882 888 BIC 3989.2 4458.6 Labor market topics Civil rights topics Black White Bl-Wh diff Black White Bl-Wh diff District Characteristics  % Change Latino pop   0% 0.836 0.866 −0.030 0.698 0.737 −0.039   50% 0.867 0.845 −0.022 0.696 0.753 −0.058**   100% 0.854 0.869 −0.015 0.693 0.770 −0.076***   150% 0.863 0.870 −0.007 0.691 0.785 −0.095***  % Latino   0% 0.861 0.871 −0.010 0.692 0.770 −0.078***   5% 0.856 0.865 −0.009 0.693 0.767 −0.073***   10% 0.851 0.859 −0.008 0.694 0.763 −0.069**  % Black   0% 0.924 0.886 0.048 0.704 0.807 −0.103*   30% 0.880 0.874 0.011 0.696 0.776 −0.080***   75% 0.781 0.854 −0.043* 0.687 0.742 −0.055*  % Immigrant occ   0% 0.733 0.817 −0.086 0.599 0.620 −0.006   20% 0.852 0.860 −0.013 0.689 0.762 −0.070***   40% 0.909 0.928 0.031 0.769 0.869 −0.103**  % Agriculture industry   0% 0.858 0.866 −0.009 0.685 0.754 −0.067**   10% 0.809 0.851 −0.042 0.747 0.837 −0.090   20% 0.751 0.835 −0.083 0.801 0.899 −0.097  % Unemployed   0% 0.807 0.854 −0.047 0.784 0.771 0.136   10% 0.845 0.867 −0.022 0.702 0.766 −0.064**   25% 0.893 0.884 0.009 0.559 0.759 −0.200*** N (votes) 4,618 5,187 N (legislators) 882 888 BIC 3989.2 4458.6 Note: Probabilities estimated from two-level logit models run separately for labor market and civil rights topics, and controlling for unified Republican state govt. and party polarization and party cohesion within the legislator’s chamber, with state and year fixed effects. (Coefficients reported in Supplementary Table iv.) See text for explanation of how marginal probabilities are calculated. ***p < 0.001 **p < 0.01 *p < 0.05. Figure 1. View largeDownload slide Probability of supporting bills with a labor market topic by legislator race and % change in the Latino population Figure 1. View largeDownload slide Probability of supporting bills with a labor market topic by legislator race and % change in the Latino population Figure 2. View largeDownload slide Probability of supporting bills with a labor market topic by legislator race and % district employed in immigrant occupations Figure 2. View largeDownload slide Probability of supporting bills with a labor market topic by legislator race and % district employed in immigrant occupations Figure 3. View largeDownload slide Probability of supporting bills with a labor market topic by legislator race and % unemployed in district Figure 3. View largeDownload slide Probability of supporting bills with a labor market topic by legislator race and % unemployed in district Figure 4. View largeDownload slide Probability of supporting bills with a civil rights topic by legislator race and % change in the Latino population Figure 4. View largeDownload slide Probability of supporting bills with a civil rights topic by legislator race and % change in the Latino population Figure 5. View largeDownload slide Probability of supporting bills with a civil rights topic by legislator race and % district employed in immigrant occupations Figure 5. View largeDownload slide Probability of supporting bills with a civil rights topic by legislator race and % district employed in immigrant occupations Figure 6. View largeDownload slide Probability of supporting bills with a civil rights topic by legislator race and % unemployed in district Figure 6. View largeDownload slide Probability of supporting bills with a civil rights topic by legislator race and % unemployed in district When we consider the labor market legislation, none of the district threat characteristics significantly affect the votes of either Black or White Democratic legislators (figures 1–3 and table 5). These results support hypothesis 3 that there will be no racial differences in how threat conditions affect legislators’ support for labor market bills. In contrast, when civil rights–related issues are on the floor, district threat characteristics evoke significantly stronger support among White Democratic legislators compared to their African American colleagues. Figure 4 shows that our key demographic indicator of “threat”—the % change in the district Latino population—increases the likelihood of supporting restrictive civil rights bills only among White legislators. A rapid rise in % Latino slightly decreases support for civil rights legislation among African American Democratic legislators. These Black-White differences in the effect of changes in the district Latino population are statistically significant, supporting hypothesis 4 (table 5). Consistent with group threat theories, larger percentages of constituents employed in immigrant occupations are associated with increased support for restrictive civil rights bills among both White and African American Democrats. However, this positive effect is significantly stronger among White Democratic legislators compared to their Black colleagues (table 5, figure 5). This too supports hypothesis 4. The effects of district unemployment also differ according to legislator race, but not in ways entirely anticipated by our hypotheses. Increases in district unemployment have no significant effect on the likelihood that White Democratic legislators will support restrictive civil rights bills. In contrast, higher rates of constituent unemployment decrease support for such civil rights measures among African American Democratic legislators (table 5, figure 6). This negative effect among Black Democratic legislators is the opposite of what we would expect in anticipation of a “threat” response. Given the possibility that districts with relatively high unemployment rates include even greater numbers of Black and/or poor residents vulnerable to voter ID laws and enhanced law enforcement, our results suggest that, in this context of proposed restrictions on civil rights, what is often considered an indicator of threat could actually be a source of commonality among many “Nuevo South” African American Democratic legislators. Conclusion Overall, the results offer considerable support for our hypotheses and their underlying logic. When the bill topic restricts immigrant employment, Black and White Democratic legislators maintain strong, equal support for these labor market “threat” measures (or “initiatives”). But when the bill topic signals that civil rights may be at stake, Black Democratic legislators are less likely to support the measure than White Democrats. Further, district-level “threat” indicators appear to affect African American and White legislators differently. More often than not, White Democratic legislators react strongly, as if threatened, by increasing their support for such restrictions. Black legislative responses are relatively weak and inconsistent, suggesting some uncertainty or ambivalence about what constitutes a threat or a commonality. Nevertheless, Black Democratic legislators consistently support restrictive immigration bills that touch on civil rights issues at lower rates than their White counterparts. Although the majority of African American Democratic legislators support restrictive immigration legislation on all but two topics (omnibus and voting), they do so at lower rates than White Democratic representatives. Indeed, if White Democrats voted like Black Democrats, the restrictive immigration legislation would still pass, given that Republicans were typically the majority party from 2005 to 2012 in every state but Arkansas and Louisiana. Despite the influence (or lack thereof) of Democratic votes on immigration bills, our study significantly advances our understanding of how the US racial hierarchy influences state policymaking. Our findings support theories of group threat and symbolic politics, suggesting that Black Democratic legislators’ perceptions of their group position in relation to Whites and Latinos, respectively, can engender legislative responses in line with group competition or with common goals grounded in the continued struggle for racial justice. We contend that the sociological significance of our study to issues of immigration, race, and politics extends beyond theories of competition and commonalities, however. Omi and Winant’s (1994) theory of racial formation suggests that immigration policy contributes to the construction of the US racial hierarchy (Ngai 2004). By legally defining who “belongs” and who does not, immigration policy creates legal and social boundaries between racialized categories and produces meanings behind those categories (Ngai 2004). For instance, the Trump administration’s proposal to build a wall at the US-Mexico border symbolically and legally designates the racialized group of “Latinos” as “not belonging” (Brown 2012; Masuoka and Junn 2013). To date, scholarship on immigration policy and race focuses on how White elites use immigration policy to maintain their dominant social position in the US racial hierarchy (Ngai 2004). Our analyses take this literature in a new direction, showing how African American Democratic legislators, representing a subordinate racial group, participate in this process of racialization through state policymaking. We show how race and symbolic politics are connected in shaping elite behavior toward immigration policy, and we identify an empirical tool (analyses by bill topic) to test hypotheses about these connections. Our focus on roll call votes to assess the role of legislator race in immigration policymaking does carry certain limitations. Only legislation that can garner sufficient support to pass is brought to the floor for a vote. Thus, the set of bills in our study is highly selective and constrained, representing just a small part of the legislative process. However, there are also advantages to analyzing roll call votes. In the public record, roll call votes are the most visible aspect of legislative behavior. Indeed, public interest groups maintain scorecards on legislator voting records on immigration legislation. Also, the outcome of roll call votes is arguably the most consequential. Enacted immigration laws institutionalize and codify symbolic boundaries, directly affecting the lives of immigrants (Ngai 2004). To the extent that these laws and their enforcement operate through targeting Latinos, restrictive immigration policies contribute to the positioning of Latinos in the US racial hierarchy (Masuoka and Junn 2013; Ngai 2004). Our analyses of roll call votes provide important direction for future research outside the “Nuevo South.” In particular, future research should address the questions of what circumstances are likely to push Black legislators toward a commonalities response for economic bills, and what circumstances are likely to push Black legislators toward a competition response to bills covering civil rights issues. One of the most important political conditions to test in future work involve the presence and power of Latino elected officials and the communities they represent. Notes 1 Given that almost half (44 percent) of all Black Democratic state legislators in the United States are located in our nine states, these states are of substantive importance as well, representing an epicenter of Black state immigration politics. 2 The two most politically impactful ways that White Democrats differ from White Republicans are the differences in political ideology and the much higher presence of Black constituents in the districts represented by White Democrats compared to White Republicans. 3 “Independence” occurs when Black Democratic legislators do not give Congressional testimony on issues promoted by the Latino Caucus. Our roll call data do not allow us to determine an “independence” effect. 4 A full 86 percent of Black representatives in our sample represented majority-Black districts; 11 percent of White representatives in our sample also represented majority-Black districts. 5 We also ran analyses including all driver’s license topics as “competition” measures. This changed the coefficients for licenses, but did not alter the main findings regarding the competition bills that we report in the paper. 6 Because the geographic boundaries for the state legislative districts and the census blocks changed between 2000 and 2010 in the census, we constructed our measure of % change in the Latino and total population by creating equivalent district boundaries. 7 The percent foreign born and percent Hispanic measures are correlated at 0.87 in our dataset. 8 The 2008 data most closely match the composition of the state assembly in these years for these states. 9 We thank an anonymous reviewer for pointing this out to us. 10 We report results from additional tests to check for the robustness of our findings in the Supplemental Methods Appendix. Supplementary Material Supplementary material is available at Social Forces online. About the Authors Irene Browne is Associate Professor of Sociology at Emory University. Her research areas include intersectionality (race, class, gender); immigration; and labor market inequality. She is editor of the book Latinas and African-American Women at Work: Race, Gender and Economic Inequality. She has also published articles in the American Sociological Review, Annual Review of Sociology, Social Forces, Sociological Quarterly, American Behavioral Scientist, and other journals. Beth Reingold is Associate Professor of Political Science and Women’s, Gender, and Sexuality Studies at Emory University. Her research on gender, race, and the politics of representation and identity in the United States has appeared most recently in the American Journal of Political Science, Political Research Quarterly, Politics & Gender, and Representation: The Case of Women (edited by Maria Escobar-Lemmon and Michelle Taylor-Robinson, 2014). Anne-Kathrin Kronberg is an Assistant Professor in the Sociology Department at the University of North Carolina–Charlotte. Her research focuses on organizations, work, and social inequality. Her work has appeared in Social Forces, Work and Occupations, and Mobilization and has been supported by the German Research Foundation. References Abascal , Maria . 2015 . “ Us and Them: Black-White Relations in the Wake of Hispanic Population Growth .” American Sociological Review 80 ( 4 ): 789 – 813 . Google Scholar CrossRef Search ADS Black , Earl , and Merle Black . 1989 . Politics and Society in the South . Cambridge, MA : Harvard University Press . Blalock , Hubert . 1967 . Toward a Theory of Minority-Group Relations . New York : Wiley . Blumer , Herbert . 1958 . “ Race Prejudice as a Sense of Group Position .” Pacific Sociological Review 1 : 3 – 7 . Google Scholar CrossRef Search ADS Bobo , Lawrence , and Vincent Hutchings . 1996 . “ Perceptions of Racial Group Competition: Extending Blumer’s Theory of Group Competition in a Multiracial Social Context .” American Sociological Review 61 ( 6 ): 951 – 72 . Google Scholar CrossRef Search ADS Bositis , David . 2011 . “Resegregation in Southern Politics?” Research Report, November. Washington, DC: Joint Center for Political and Economic Studies. Brown , Hana . 2013 . “ Race, Legality and the Social Policy Consequences of Anti-Immigration Mobilization .” American Sociological Review 78 ( 2 ): 290 – 314 . Google Scholar CrossRef Search ADS Canon , David . 1999 . Race, Redistricting, and Representation: The Unintended Consequences of Black Majority Districts . Chicago : University of Chicago Press . Ceobanu , Alin , and Xavier Escandell . 2010 . “ Comparative Analyses of Public Attitudes Toward Immigrants and Immigration Using Multinational Survey Data: A Review of Theories and Research .” Annual Review of Sociology 36 : 309 – 28 . Google Scholar CrossRef Search ADS Chavez , Jorge , and Doris Provine . 2009 . “ Race and the Response of State Legislatures to Unauthorized Immigrants .” Annals of the American Academy of Political and Social Science 623 ( 1 ): 78 – 92 . Google Scholar CrossRef Search ADS Chavez , Leo . 2013 . The Latino Threat Narrative , 2nd ed . Palo Alto, CA : Stanford University Press . Citrin , Jack , Donald Green , Christopher Muste , and Cara Wong . 1997 . “ Public Opinion toward Immigration Reform: The Role of Economic Motivations .” Journal of Politics 59 : 858 – 81 . Google Scholar CrossRef Search ADS Citrin , Jack , Beth Reingold , and Donald Green . 1990 . “ American Identity and the Politics of Ethnic Change .” The Journal of Politics 52 : 1124 – 1154 . Google Scholar CrossRef Search ADS Clawson , Daniel . 1989 . “ Interlocks, PACS, and Corporate Conservatism .” American Journal of Sociology 94 ( 4 ): 749 – 73 . Google Scholar CrossRef Search ADS Diamond , Jeff . 1998 . “ African-American Attitudes Towards United States Immigration Policy .” International Migration Review 32 : 51 – 70 . Google Scholar CrossRef Search ADS Facchini , Giovanni , and Max Friedrich Steinhardt . 2011 . “ What Drives U.S. Immigration Policy? Evidence from Congressional Roll Call Votes .” Journal of Public Economics 95 ( 7 ): 734 – 43 . Google Scholar CrossRef Search ADS Gaillard , Frye . 2004 . Cradle of Freedom: Alabama and the Movement That Changed America . Tuscaloosa : University of Alabama Press . Gay , Claudine . 2006 . “ Seeing Difference: The Effect of Economic Disparity on Black Attitudes towards Latinos .” American Sociological Review 50 ( 4 ): 982 – 97 . Gonzalez , Jorge , and Nipoli Kamdar . 2000 . “ Do Not Give Me Your Tired, Your Poor! Determinants of Legislator Voting on Immigration Issues .” Eastern Economic Journal 26 ( 2 ): 127 – 43 . Griffin , John , and Brian Newman . 2008 . Minority Report: Evaluating Political Equality in America . Chicago : University of Chicago Press . Google Scholar CrossRef Search ADS Grose , Christian . 2011 . Congress in Black and White . New York : Cambridge University Press . Google Scholar CrossRef Search ADS Guo , Guang , and H. X. Zhao . 2000 . “ Multilevel Modeling for Binary Data .” Annual Review of Sociology 26 : 441 – 62 . Google Scholar CrossRef Search ADS Hainmueller , Jens , and Daniel Hopkins . 2014 . “ Public Attitudes towards Immigration .” Annual Review of Political Science 17 : 225 – 49 . Google Scholar CrossRef Search ADS Hero , Rodney , and Robert Preuhs . 2013 . Black-Latino Relations in US National Politics: Beyond Conflict or Cooperation . New York : Cambridge University Press . Hopkins , Daniel . 2010 . “ Politicized Places: Explaining Where and When Immigrants Provoke Local Opposition .” American Political Science Review 104 ( 1 ): 40 – 60 . Google Scholar CrossRef Search ADS Hutchings , Vincent , and Cara Wong . 2014 . “ Racism, Group Position, and Attitudes about Immigration among Blacks and Whites .” Du Bois Review 11 ( 2 ): 419 – 42 . Google Scholar CrossRef Search ADS Jones-Correa , Michael. 2011 . “Commonalities, Competition, and Linked Fate.” In Just Neighbors? Research on African-American and Latino Relations in the United States , edited by Edward Telles , Mark Q. Sawyer , and Gaspar Rivera-Salgado , 63 – 95 . New York : Russell Sage Foundation . Juenke , Eric Gonzalez , and Robert Preuhs . 2012 . “ Irreplaceable Legislators? Rethinking Minority Representatives in the New Century .” American Journal of Political Science 56 ( 3 ): 705 – 15 . Google Scholar CrossRef Search ADS Kaufmann , Karen . 2003 . “ Cracks in the Rainbow: Group Commonality as a Basis for Latino and African-American Political Coalitions .” Political Research Quarterly 56 ( 2 ): 107 – 26 . Google Scholar CrossRef Search ADS King-Meadows , Tyson , and Thomas Schaller . 2007 . Devolution and Black State Legislators: Challenges and Choices in the Twenty-First Century . New York : SUNY Press . Long , J. Scott , and Jeremy Freese . 2014 . Regression Models for Categorical Dependent Variables Using Stata , 3rd ed . College Station, TX : Stata Press . Lovato , Roberto . 2008 . “Juan Crow in Georgia.” The Nation, May 26. http://www.thenation.com/article/juan-crow-georgia. Marrow , Helen . 2011 . “Intergroup Relations: Reconceptualizing Discrimination and Hierarchy.” In Being Brown in Dixie: Race, Ethnicity, and Latino Immigration in the New South , edited by Cameron Lippard and Charles Gallagher , 53 – 76 . Boulder : First Forum Press . Masuoka , Natalie , and Jane Junn . 2013 . The Politics of Belonging: Race, Public Opinion and Immigration . Chicago : University of Chicago Press . Google Scholar CrossRef Search ADS McAdam , Douglas . 1990 . Freedom Summer . New York : Oxford University Press . McCarty , Nolan , Keith Poole , and Howard Rosenthal . 2016 . Polarized America: The Dance of Ideology and Unequal Riches . Cambridge, MA : MIT Press . McClain , Paula , Victoria DeFrancesco Soto , Monique Lyle , Gerald Lackey , Jeffrey Grynaviski , Kendra Cotton , Shayla Nunnally , Thomas Scotto , and Allan Kendrick . 2008 . “Black Elites and Latino Immigrant Relations in a Southern City: Do Black Elites and the Black Masses Agree?” In New Race Politics in America , edited by Jane Junn and Kerry Haynie , 145 – 65 . Cambridge : Cambridge University Press . Google Scholar CrossRef Search ADS McClain , Paula , Monique Lyle , Efren Perez , Jessica Johnson Carew , Eugene Walton , Candis Watts , Gerald Lackey , Danielle Clealand , and Shayla Nunnally . 2009 . “Black and White Americans and Latino Immigrants: A Preliminary Look at Attitudes in Three Southern Cities.” Presented at the Annual Meeting of the American Political Science Association, Toronto, Canada. McDermott , Monica . 2011 . “Black Attitudes and Hispanic Immigrants in South Carolina.” In Just Neighbors? Research on African-American and Latino Relations in the United States , edited by Edward Telles , Mark Sawyer , and Gaspar Rivera-Salgado , 242 – 66 . New York : Russell Sage Foundation . McKanders , Karla . 2010 . “ Sustaining Tiered Personhood: Jim Crow and Anti-Immigrant Laws .” Harvard Journal of Racial and Ethnic Justice 26 : 163 . Monogan , James . 2013 . “ The Politics of Immigrant Policy in the 50 US States, 2005–2011 .” Journal of Public Policy 33 ( 1 ): 35 – 64 . Google Scholar CrossRef Search ADS Motel , Seth , and Eileen Patten . 2013 . Statistical Portrait of the Foreign-Born Population in the United States, 2011. PEW Hispanic Center. http://www.pewhispanic.org/files/2013/01/PHC-2011-FB-Stat-Profiles.pdf. NAACP . 2014 . “Born Suspect: Stop-and-Frisk Abuses and the Continued Struggle to End Racial Profiling in America.” Report. https://action.naacp.org/page/-/Criminal%20Justice/Born_Suspect_Report_final_web.pdf.. NAACP . 2011 . “NAACP Immigration Fact Sheet and Talking Points.” http://www.naacp.org/wp-content/uploads/2016/04/Immig%20Factsheet%20Tlkng%20Pts%20Final%20July%202011.pdf. Ngai , Mae . 2004 . Impossible Subjects: Illegal Aliens and the Making of Modern America . Princeton, NJ : Princeton University Press . NHCSL . 2010 . “In Response to Draconian Immigration Law, Black and Hispanic State Legislators Pull Joint Conference from Arizona Venue.” Press release. http://www.altoarizona.com/documents/Black_Hispanic_State_Legislators_Pull_Joint_Conference_From_Arizona_Venue.pdf. Nicholson-Crotty , Jill , and Sean Nicholson-Crotty . 2011 . “ Industry Strength and Immigrant Policy in the American States .” Political Research Quarterly 64 ( 3 ): 612 – 24 . Google Scholar CrossRef Search ADS Nteta , Tatishe . 2013 . “ United We Stand? African-Americans, Self-Interest and Immigration Reform .” American Politics Research 41 ( 1 ): 147 – 72 . Google Scholar CrossRef Search ADS Odem , Mary , and Elaine Lacy . 2009 . Latino Immigrants and the Transformation of the US South . Athens : University of Georgia Press . Omi , Michael , and Howard Winant . 1994 . Racial Formation in the United States: From the 1960s to the 1980s , 2nd ed . New York : Routledge and Kegan-Paul . Pager , Devah , and Lincoln Quillian . 2005 . “ Walking the Talk? What Employers Say Versus What They Do .” American Sociological Review 70 : 355 – 80 . Google Scholar CrossRef Search ADS Passel , Jeffrey , and D’Vera Cohn . 2011 . Unauthorized Immigrant Population: National and State Trends, 2010 . Washington, DC : Pew Hispanic Center . Pew Research Center . 2013 . “Mapping the Latino Population, by State, County and City.” http://www.pewhispanic.org/files/2013/08/latino_populations_in_the_states_counties_and_cities_FINAL.pdf. Pew Research Center . 2015 . “U.S. Foreign Born Population Trends.” http://www.pewhispanic.org/2015/09/28/chapter-5-u-s-foreign-born-population-trends/. Pruehs , Robert . 2006 . “ The Conditional Effects of Minority Descriptive Representation: Black Democratic Legislators and Policy Influence in the American States .” Journal of Politics 68 ( 3 ): 585 – 99 . Google Scholar CrossRef Search ADS Ramakrishnan , S. Karthick , and Tom Wong . 2010 . “Partisanship, Not Spanish: Explaining Municipal Ordinances Affecting Undocumented Immigrants. In Taking Local Control: Immigration Policy Activism in US Cities and States , edited by Monica W. Varsanyi , 73 – 93 . Palo Alto, CA : Stanford University Press . Raudenbush , Stephen , and Anthony Bryk . 2002 . Hierarchical Linear Models: Applications and Data Analysis. Thousand Oaks, CA : Sage Publications . Raudenbush , Stephen , Randall Fotiu , and Yuk Fai Cheong . 1999 . “ Synthesizing Results from the Trial State Assessment .” Journal of Educational and Behavioral Statistics . 24 ( 4 ): 413 – 38 . Google Scholar CrossRef Search ADS Shafer , Byron , and Richard Johnston . 2009 . The End of Southern Exceptionalism: Class, Race, and Partisan Change in the Postwar South . Cambridge, MA : Harvard University Press . Shor , Boris , and Nolan McCarty . 2011 . “ The Ideological Mapping of American Legislatures .” American Political Science Review 105 ( 3 ): 530 – 55 . Google Scholar CrossRef Search ADS Tate , Katherine . 2003 . Black Faces in the Mirror: African-Americans and their Representatives in Congress . Princeton, NJ : Princeton University Press . Thurber , James , and Antoine Yoshinaka , eds. 2015 . American Gridlock: The Sources, Character, and Impact of Political Polarization . New York : Cambridge University Press . Google Scholar CrossRef Search ADS Wilkinson , Betina Curaia . 2015 . Partners or Rivals? Charlottesville : University of Virginia Press . Williams , Kim , and Lonnie Hannon III . 2016 . “ Immigrant Rights in a Deep South City: The Effects of Anti-Immigrant Legislation on Black Elite Opinion in Birmingham, AL .” Du Bois Review 13 ( 1 ): 139 – 57 . Google Scholar CrossRef Search ADS Author notes We thank Tim Dowd, Helen Marrow, Joya Misra, Alex Hicks, and Kim Williams for their comments on the paper. We are grateful to Dr. Rob O’Reilly and Shannon McClintock for their expert statistics and data advice. We were fortunate to have an exceptionally skilled group of research assistants who assisted in the data collection and coding: Shilpi Agrawala, Alex Reibman, Joanna Chang, Tiffany Chen, In Young Park, Yordanos Agajyelleh, and Jeffeline Ermilus. We thank the Emory SIRE and RISE Undergraduate Research Programs for supporting these students. © The Author(s) 2018. Published by Oxford University Press on behalf of the University of North Carolina at Chapel Hill. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Social Forces Oxford University Press

Race Relations, Black Elites, and Immigration Politics: Conflict, Commonalities, and Context

Loading next page...
 
/lp/ou_press/race-relations-black-elites-and-immigration-politics-conflict-A0f7S0SZu4
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the University of North Carolina at Chapel Hill. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
0037-7732
eISSN
1534-7605
D.O.I.
10.1093/sf/sox102
Publisher site
See Article on Publisher Site

Abstract

Abstract This paper investigates the question of whether and how race and race relations affect state legislators’ support for restrictive immigration policy. Focusing on the nine new immigration destinations in the US Southeast, we compare the roll call votes cast by African American and White Democratic state legislators on 196 restrictive immigration measures proposed between 2005 and 2012. Overall, the majority of Black Democratic legislators support restrictive immigration legislation, although at a consistently lower rate than White Democratic legislators. We test hypotheses that Black-White differences in support for restrictive immigration legislation depend upon the material resources at stake and the symbolic significance associated with specific legislation. Using bill topic to measure these conditions, we find that Black and White Democrats equally support immigrant “competition” bills on employment topics. Black Democratic legislators are less likely than their White counterparts to support immigration bills on Black-immigrant “commonalities” topics related to civil rights issues: voter ID regulations, education, police, and omnibus restrictions. Further, for civil rights bills only, district threat indicators such as a rapidly growing district Latino population are much more likely to increase support for restrictive immigration bills among White Democratic legislators than among Black Democratic legislators. Some district characteristics even evoke a commonalities response from African American Democratic legislators, reducing their support for restrictive legislation. Our research thus expands the competition/commonalities debate and highlights the complexities and contingencies of race relations and state policymaking in the twenty-first century. Introduction Rapid increases in immigration to new US destinations are transforming local demographic landscapes and challenging state politics traditionally based on a Black-White racial divide (Marrow 2011). The growing number of immigrants could spur African American political leaders to ally with conservative Whites to push anti-immigrant legislation, particularly if African Americans perceive immigrants as competitors for jobs and resources. Alternatively, African American leaders may unite with immigrants and their supporters if they see immigrants, like themselves, as disadvantaged minorities just a few generations away from the South’s Jim Crow system. Indeed, McKanders (2010) warns that Southern states in particular are creating a new system of “Juan Crow” that directly parallels Jim Crow, with laws preventing undocumented immigrants from attaining access to housing, jobs, education, and healthcare (Lovato 2008). Other anti-immigrant legislation tightens access to the ballot box and extends the reach of law enforcement—issues at the heart of the Black Civil Rights movement. With jobs at stake on the one hand and civil rights issues at stake on the other, do African American lawmakers support or oppose “Juan Crow” laws? We address this question by comparing African American and White Democratic legislators’ roll call votes on restrictive immigration bills proposed between 2005 and 2012 in the “Nuevo South”—the states of the former Confederacy identified as new immigrant destinations (AL, AR, GA, LA, MS, NC, SC, TN, and VA) (Marrow 2011). The unique racial history of these “Nuevo South” states has produced especially fertile conditions for testing the theories of Black/immigrant competition and commonalities that undergird our study. More than other regions in the United States, the social and political institutions in the “Nuevo South” are centered on the Black/White binary, and Black representatives have gained a large presence in state politics (Bositis 2011; King-Meadows and Schaller 2007). Approximately 44 percent of all Black state legislators serve in the nine “Nuevo South” states, and these Black legislators wield more political clout within the Democratic party than do their counterparts outside the region (King-Meadows and Schaller 2007). Black Democratic legislators in the “Nuevo South” do not share political power with Latino representatives as they do in traditional immigrant gateway states.1 While Black/White relations define state politics and virtually every important social and political outcome in the “Nuevo South” (McClain et al. 2009, 1), this region has experienced the most rapid growth in immigration of any region of the country (Pew 2015). The rapidly rising Latino immigrant population in states where the vast majority of representatives and their constituencies are Black or White could be perceived by Blacks as a threat to their hard-won political and economic gains in the region, prompting Black representatives to take an anti-immigrant stance. Yet, the history of race relations in the “Nuevo South” also could generate sympathy toward immigrants among Black representatives and prompt a pro-immigrant response. Unlike other regions of the United States, the “Nuevo South” carries the enduring legacy of Jim Crow racism, with a proud identity as the “cradle of the civil rights movement” (Black and Black 1989; Gaillard 2004; McAdam 1990). The few existing studies of African American lawmakers’ stances toward immigration policy are limited to traditional immigration gateways or national politics and primarily focus on conditions that promote alliances between Black Democratic legislators and Latino political caucuses (e.g., Diamond 1998; Hero and Preuhs 2013). These studies yield mixed hypotheses and findings with respect to African American Democratic legislators and immigration legislation, with some studies finding African American support for restrictive immigration legislation and other studies finding opposition to these bills (see Hero and Preuhs [2013] for a review). Our theoretical framework draws on scholarship that identifies the US racial hierarchy as a driving force behind immigration politics. Theories of group threat posit that lawmakers’ perception of immigrant competition arising from the relative position of Blacks, Latinos, and Whites in the racial hierarchy will stimulate support for restrictive immigration legislation among both Black and White Democratic legislators. Theories of symbolic politics suggest that beyond material group interests, the values and beliefs arising from a history of race relations and the continued struggle for racial justice will also influence African American legislative behavior toward immigration policy. In particular, Black support for restrictive immigration policy may drop if Black lawmakers perceive a common goal of racial justice and a shared social position with immigrants as disadvantaged minorities in a White-dominated racial hierarchy. White Democratic legislators are much less likely to perceive systematic racial inequality and to place racial justice among their legislative priorities when dealing with immigration (Masuoka and Junn 2013). We develop a research design that allows us to uncover the role of group interests as well as values and beliefs in shaping African American legislative responses to restrictive immigration policy, respectively. In particular, we divide bills by topic, and compare votes on topics related to economic group interest with votes on civil rights issues. By analyzing the roll call votes of Southern Democratic state legislators, we effectively control for the influence of other factors such as partisanship and political ideology. We exclude Republicans not only because there is virtually no variation in Republican votes on restrictive immigration bills, but also to ensure that the racial differences that we do observe cannot be attributed to the most prominent alternative explanation—partisanship. Given the uniformly higher rates of support among Republicans (99 percent of whom are White), the Black-White differences that we report in the paper would be even stronger if we included Republicans.2 Our findings suggest that rather than reflecting unilateral competition or commonalities responses, immigration politics in the “Nuevo South” is multidimensional and context specific. First, although the majority of African American Democratic legislators support restrictive immigration legislation, they are voting “yea” at lower rates than their White counterparts. Second, African American legislative behavior varies depending on the topics associated with specific bills. When bills concern immigrant employment, African American Democratic legislators join their White colleagues and vote to reduce economic threat. When bills concern civil rights, however, African American Democratic legislators diverge from their White colleagues and dampen their support for restrictive immigration legislation. Consistent with the theory of symbolic politics, these commonalities responses of African American Democratic lawmakers do not simply reflect African American group interest, but are tied to the values and ideologies of racial justice—that is, the symbolic significance—associated with specific bills. We also find that whether district demographics prompt voting patterns reflecting an immigration threat or an opportunity to build interracial bridges depends on both the bill topic and the race of the legislator (Jones-Correa 2011). We elucidate how these factors are interrelated below, first describing the context of our study and then discussing theories of group threat and symbolic politics. We state our hypotheses derived from these theories, describe the methods we use to test our hypotheses, and present our results. In the conclusion, we move beyond a consideration of legislative behavior as a response to racial inequality, and discuss the broader relevance of our study to theories that situate the state at the center of producing racial categories and enforcing the boundaries between these categories. The Rise of the Nuevo South and the Latino Backlash With local immigrant populations rising 200–800 percent in traditionally Black-White communities throughout the Southeast, immigrants have encountered a social and political backlash (Odem and Lacy 2009; Passel and Cohn 2011). Amid public outcry about “illegals” stealing jobs, burdening taxpayers, and increasing crime rates, state and local officials across the Southeast have passed a range of restrictive laws targeting unauthorized immigrants. These laws aim to limit access of unauthorized immigrants to resources such as jobs, public benefits, and educational opportunities, as well as expand the powers of local law enforcement to identify and detain unauthorized immigrants (Monogan 2013). Between 2005 and 2012, the nine states of the “Nuevo South” proposed 196 restrictive immigration bills that received roll call votes in at least one chamber. Bills covering employment and law enforcement were most common, with every state voting on at least one of each type of bill (table 1). Table 1. Restrictive Immigration Legislation, 2005–2012: Bill Topic by State AL AR GA LA MS NC SC TN VA Total Labor market restrictions (n = 70)  Employment 16 3 10 5 4 1 1 11 7 58  IDs 5 3 2 1 1 0 0 0 0 12 Civil rights restrictions (N = 78)  Police 2 2 3 5 6 3 2 10 9 42  Voting 1 0 4 1 1 1 1 1 2 12  Education 1 3 4 1 1 1 1 0 5 17  Omnibus 1 0 1 0 2 0 1 2 0 7 Other (n = 48)  Benefits 0 1 1 2 2 0 0 4 1 11  Budget 0 0 0 0 1 1 2 0 1 5  Driver’s licenses (nonwork) 2 2 0 2 2 1 1 1 2 13  Safety 0 1 4 0 0 0 0 0 3 8  Misc 0 0 2 0 1 1 0 4 3 11 Total 28 15 31 17 21 9 9 33 33 196 AL AR GA LA MS NC SC TN VA Total Labor market restrictions (n = 70)  Employment 16 3 10 5 4 1 1 11 7 58  IDs 5 3 2 1 1 0 0 0 0 12 Civil rights restrictions (N = 78)  Police 2 2 3 5 6 3 2 10 9 42  Voting 1 0 4 1 1 1 1 1 2 12  Education 1 3 4 1 1 1 1 0 5 17  Omnibus 1 0 1 0 2 0 1 2 0 7 Other (n = 48)  Benefits 0 1 1 2 2 0 0 4 1 11  Budget 0 0 0 0 1 1 2 0 1 5  Driver’s licenses (nonwork) 2 2 0 2 2 1 1 1 2 13  Safety 0 1 4 0 0 0 0 0 3 8  Misc 0 0 2 0 1 1 0 4 3 11 Total 28 15 31 17 21 9 9 33 33 196 Table 1. Restrictive Immigration Legislation, 2005–2012: Bill Topic by State AL AR GA LA MS NC SC TN VA Total Labor market restrictions (n = 70)  Employment 16 3 10 5 4 1 1 11 7 58  IDs 5 3 2 1 1 0 0 0 0 12 Civil rights restrictions (N = 78)  Police 2 2 3 5 6 3 2 10 9 42  Voting 1 0 4 1 1 1 1 1 2 12  Education 1 3 4 1 1 1 1 0 5 17  Omnibus 1 0 1 0 2 0 1 2 0 7 Other (n = 48)  Benefits 0 1 1 2 2 0 0 4 1 11  Budget 0 0 0 0 1 1 2 0 1 5  Driver’s licenses (nonwork) 2 2 0 2 2 1 1 1 2 13  Safety 0 1 4 0 0 0 0 0 3 8  Misc 0 0 2 0 1 1 0 4 3 11 Total 28 15 31 17 21 9 9 33 33 196 AL AR GA LA MS NC SC TN VA Total Labor market restrictions (n = 70)  Employment 16 3 10 5 4 1 1 11 7 58  IDs 5 3 2 1 1 0 0 0 0 12 Civil rights restrictions (N = 78)  Police 2 2 3 5 6 3 2 10 9 42  Voting 1 0 4 1 1 1 1 1 2 12  Education 1 3 4 1 1 1 1 0 5 17  Omnibus 1 0 1 0 2 0 1 2 0 7 Other (n = 48)  Benefits 0 1 1 2 2 0 0 4 1 11  Budget 0 0 0 0 1 1 2 0 1 5  Driver’s licenses (nonwork) 2 2 0 2 2 1 1 1 2 13  Safety 0 1 4 0 0 0 0 0 3 8  Misc 0 0 2 0 1 1 0 4 3 11 Total 28 15 31 17 21 9 9 33 33 196 While states outside the South have enacted similar measures, the historical and political context of racial politics in the South is both similar to and different from other regions of the country (Monogan 2013; Wilkinson 2015). In the new immigrant destination states of the South, African Americans hold a large number of seats in state legislatures among Democrats, while only four of the 503 Democratic legislators are Latino. And the Latino population, while growing, remains small, ranging from 3 percent (MS) to 9 percent (GA, NC) (Pew 2013). Across the social science literature, support for immigration in the United States is read as support for “Latino issues” (Hero and Preuhs 2013). Conversely, support for restricting immigration is interpreted as opposition to Latino issues and interests (Ramakrishnan and Wong 2010). The literature on public and elite positions toward immigration legislation thus encompasses attitudes toward Latinos as well as toward immigrants and immigration policy (Hainmueller and Hopkins 2014). Empirical and Theoretical Background The few published studies on the positions that African American legislators take on restrictive immigration legislation report mixed results. In his historical overview of Black elite stances toward immigration policy, Diamond observes, “[f]or almost two centuries, blacks have been torn on the issue of immigration” (1998, 247). Unlike Diamond (1998), Hero and Preuhs (2013) find consistent “commonalities” or “independent” responses to immigration policy at the national level with no evidence of competition stances.3 The authors emphasize that level of government matters; unlike national policies, restrictive immigration legislation might be perceived as zero-sum by Black politicians at the state or local level. Indeed, some research suggests that anti-immigrant responses do appear to be common at the local level. In research on politics in school boards and local jurisdictions, African American leaders support restrictive immigration policy (see Hero and Preuhs [2013] for a review). In one of the few studies of Black elite positions on state immigration policy, Williams and Hannon (2016) found that in 2007, Alabama’s African American politicians were neutral on immigration policy, claiming that immigration issues did not affect their constituency. By 2013, however, these Black lawmakers overwhelmingly opposed Alabama’s omnibus immigration bill, HB6. The authors argue that the shift occurred because Black elites framed the punitive omnibus bill as a civil rights issue. However, we do not know whether this pattern holds for state immigration policies across the full spectrum of issues or topics. Public opinion surveys also indicate that local demographics and the specific policy influence Black attitudes toward immigration—and their differences with Whites (Masuoka and Junn 2013, 171). The clearest pattern across surveys emerges when jobs are at stake. The majority of Blacks support restrictive employment-related policies such as sanctioning employers who hire undocumented immigrants or requiring IDs proving legal status for employment (Masuoka and Junn 2013, 172) or when their relative economic position is lower than their Latino neighbors (Gay 2006). There is some evidence that when the policy involves the civil rights issue of education, Black opinion is more sympathetic to immigrants (Masuoka and Junn 2013). For both labor market and civil rights issues, Black support for restrictive immigration policy is either equivalent to or more liberal (less restrictionist) than White opinion (Masuoka and Junn 2013, 172). Given the mixed findings in the literature, it is not clear whether African American Democratic legislators in the “Nuevo South” would support or oppose restrictive immigration legislation more frequently than White Democratic legislators. We address this question, and also draw upon group threat theories and theories of symbolic politics to identify conditions under which legislative behavior toward restrictive immigration legislation will differ for African American and White Democrats. Group Threat Group threat theory represents the most thoroughly tested explanation of attitudes toward restrictive immigration legislation (Ceobanu and Escandell 2010). According to Blumer (1958), perceptions of race-based threat are activated when members of one group on a race/ethnic hierarchy perceive that a group below them on the hierarchy is gaining resources in a zero-sum situation. Thus, if Blacks see Latinos—and by extension immigrants—as a growing population that is overtaking the political and social position of Blacks as the predominant racial/ethnic minority, African American Democratic legislators (and their constituents) may perceive immigrants as more of a threat than White legislators do. Other scholars of group threat theory assume that the main source of threat derives from labor market competition, and that the perception of threat reflects a “realistic” appraisal of that competition (Bobo and Hutchings 1996; Hopkins 2010). Studies have found that different sectors of the African American population hold different views of Latinos and/or immigrants, depending upon whether they are more or less likely to compete with Latinos or immigrants. Working-class Blacks embedded in labor market sectors where they are most likely to compete with Latino immigrants express views consistent with group threat theory (Gay 2006; McDermott 2011). Poll data from 2006 indicate that working-class Blacks were more likely to hold anti-immigrant attitudes compared to middle-class Blacks; however, the Black class differences appear to have disappeared by 2013 (Nteta 2013). In addition to economic threat, public concern over a growing cultural threat would arise if the increasing number of immigrants were perceived as transforming “the American way of life” through language and customs (Brown 2013; Hainmueller and Hopkins 2014). Usually, it is Latinos who are depicted as “the immigrant cultural threat” (Chavez 2013). This could trigger perceptions of threat among both Blacks and Whites equally, especially if both groups experience a “cultural threat” from news reports of a rapidly growing Latino population or from the increased visibility of Spanish-language establishments and services (Abascal 2015; Hainmueller and Hopkins 2014). Several scholars find that it is Whites who are more likely to support restrictive immigration legislation compared to Blacks (Citrin et al. 1997; Hutchings and Wong 2014; Masuoka and Junn 2013). The authors suggest that although both Whites and Blacks perceive that the growing number of Latino immigrants are posing a political and cultural threat, Whites are also driven by racial prejudice to a greater extent than Blacks are (Hutchings and Wong 2014). Tests of group threat theory often measure threat through variation in the number of Latinos or immigrants within a local community (Blalock 1967; Blumer 1958). Refinements of this theory suggest that more than the “raw percent” of Latinos or immigrants, perceptions of threat will be most acute in places where the Latino or immigrant population grows rapidly. For instance, Hopkins (2010, 41) finds that a rapid increase in the number of immigrants represents the single strongest predictor of whether a county considers an anti-immigrant ordinance. At the state level, Monogan (2013) also finds that states with faster-growing foreign-born populations are more likely to adopt restrictive immigration policies. Thus, even though Blacks and Whites far outnumber immigrants in the legislative districts of the “Nuevo South” that we analyze, the remarkable pace of immigration to the South could generate perceptions of competition or threat. Further, some scholars contend that perceived cultural threat is the strongest determinant of negative attitudes toward immigration policy, suggesting that rapid rises in the local immigration population are particularly strong triggers for threat. Indeed, Hainmueller and Hopkins (2014) argue that rather than economic self-interest, anti-immigrant attitudes are based on the framing of immigration issues and the perceived symbolic threat that immigrants pose to “the nation.” Race and Symbolic Politics Theories of symbolic politics extend beyond material self-interest to explain political stances toward immigration. These theories posit that legislators and their constituents will support or oppose policy depending on how the policy symbolically aligns with their core values and ideologies (Citrin et al. 1990). For African Americans, racial identity and national identity represent two of the most salient “symbolic predispositions” that influence attitudes toward immigration policy (Abascal 2015; Citrin et al. 1990; Masuoka and Junn 2013). Masuoka and Junn (2013, 2) argue that their lower position in the US racial hierarchy will lead Blacks to develop a distinct group identity and sense of belonging, which in turn influences their attitudes toward immigration. The authors find that African Americans oppose restrictive immigration policy when they have a strong Black identity and they believe that Blacks and Latinos share a common goal of ending discrimination and racial oppression (see also Wilkinson [2015]; Wilkinson and Bingham [2016]). Kaufmann describes this perception of “shared symbolic attachments” as the “symbolic glue” that is necessary for Blacks to perceive commonalities with Latinos (2003, 199–210). However, the theory of symbolic politics suggests that it is not simply a general perception of common goals and values between Latinos and Blacks that will generate Black alliance with Latinos in their opposition to restrictive immigration legislation. Rather, Black legislative responses will vary “depending upon the particular symbols that become associated with a proposal—that is, on how the issue is symbolically framed” (Citrin et al. 1990, 1126). We propose that bill topic captures this “symbolic framing” of immigration issues. Because African Americans hold a lower place on the racial hierarchy than Whites and a shared history of struggle against racial oppression, the symbolic significance of certain topics will differ for Black Democratic legislators than for White Democratic legislators. In particular, bill topics involving civil rights issues may be particularly likely to elicit Black commonality votes in solidarity with immigrants. Representation in Institutional Context If racial group identity is central in shaping the African American values and goals that can produce the necessary “symbolic glue” for commonalities responses, then both legislator race and the racial composition of their constituency may lead Black Democratic legislators to vote differently on immigration legislation compared to White Democratic legislators. This is particularly important in our nine “Nuevo South” states, where White Democratic legislators may have a fairly substantial number of Black constituents in their district.4 However, it is not clear that these attitudes among the Black population would translate directly onto Black legislator roll call votes on immigration bills. First, audit studies show that attitudes can differ from their behavior (Pager and Quillian 2005). Second, as we discuss below, Black legislators face a different set of institutional conditions compared to their Black constituencies. Third, there is evidence that race influences the political behavior of Black representatives independent of the preferences of their Black constituencies, and that Black elite views on immigration diverge from the attitudes of their constituencies (McClain et al. 2008; Tate 2003). African American Democratic legislators appear to be more liberal than their Black constituents, particularly on issues such as welfare reform, crime control, and illegal immigration (Griffin and Newman 2008; Tate 2003). Numerous studies of legislative roll call voting also show that Black representatives are significantly more liberal than their White Democratic counterparts, even controlling for constituency characteristics (Canon 1999; Grose 2011; Juenke and Preuhs 2012). Our analyses allow us to consider whether and how the race of African American Democratic legislators produces an independent effect on roll call voting above and beyond the influence of their Black constituencies and in comparison to White Democrats who also represent Black constituencies. State legislators not only respond to their constituency and their individual political preferences; they also face pressure from other stakeholders in the immigration debate (Gonzalez and Kamdar 2000; Monogan 2013). Historically, White Democratic legislators in the South (“Dixiecrats”) and their White constituencies were more conservative than White Democrats in states outside the South (Shafer and Johnston 2009). However, this ideological difference has narrowed since the 1960s, as conservative Southern Whites have moved to the Republican Party (McCarty, Poole, and Rosenthal 2016). Currently, party polarization defines the ideological split throughout the United States—including the South (Thurber and Yoshinaka 2015). Nonetheless, these Southern Democrats are relatively conservative overall (Shor and McCarty 2011). Black Democratic state legislators are embedded within a complex institutional field that includes the Democratic Party as well as organized groups with interests in their district and state; the context of office-holding in their state (such as term limits and party control); and their financial backers (Clawson 1989; Gonzalez and Kamdar 2000; Monogan 2013; Nicholson-Crotty and Nicholson-Crotty 2011). Studies show that political factors such as the existence of a Republican majority in a city’s electorate (Ramakrishnan and Wong 2010) or conservative political ideology among citizens (Chavez and Provine 2009; Nicholson-Crotty and Nicholson-Crotty 2011) appear to better predict restrictive immigration ordinances at the state and local level compared to district demographics. Unlike White Democratic state legislators, Black state legislators may also attend to African American advocacy groups such as the National Association for the Advancement of Colored People (NAACP) and the National Black Caucus of State Legislators (NBCSL). Both of these organizations have publicly opposed state and federal immigration policy that expands police powers, including restrictive “omnibus” legislation in such states as Arizona and Alabama, with “commonalities” claims that these policies encourage police to engage in racial profiling (NAACP 2011; NHCSL 2010). Finally, Black Democratic legislators also must consider the context in which a bill is proposed. They might support an immigration bill in exchange for receiving support on a bill that they are sponsoring, for instance. The presence of a Latino caucus also influences Black elite stances toward immigration policy (Hero and Preuhs 2013). Importantly, our “Nuevo South” study looks at how Black Democratic legislators vote when restrictive immigration bills are on the table in uniformly conservative states when they are free from pressure from a Latino caucus or a relatively large Latino constituency. This allows us to more clearly identify the threat and symbolic factors driving Black competition and commonalities responses. We argue that in the absence of elected and organized Latinos, the most proximate influence shaping a “competition” or a “commonalities” stance among African American Democratic legislators is the topic of the bill itself—once political institutional factors are controlled. Hypotheses Theories of group threat and symbolic politics lead to the following predictions about the relation between bill topics and legislative responses: H1: When the restrictive immigration bill topic is likely to evoke considerations of interracial group threat and competition, African American Democratic legislators will be equally likely as or more likely than White Democratic legislators to support the measure. Specifically, bill topics related to employment will likely evoke considerations of group threat and competition among both White Democratic and Black Democratic legislators H2: When the bill topic is likely to evoke considerations of Black-immigrant commonalities, African American Democratic legislators will be less likely than White Democratic legislators to support the measure. Specifically, bills on civil rights–related topics will elicit a “commonalities” response either by directly restricting African American access to social or political resources (e.g., voter ID laws) and/or by symbolically reflecting shared goals and values between Blacks and Latinos. We also expect that the underlying social and political mechanisms of group threat depend on (or are conditioned by) the bill topic and race of legislator combined: H3: When the bill topic concerns labor market competition, the conditions that prompt threat (i.e., pace of change in the district Latino population and district labor market characteristics) will have an equal or stronger effect on increasing support among African American Democratic legislators compared to White Democratic legislators. H4: When the bill topic concerns civil rights issues, then the conditions that prompt threat will have stronger effects on increasing support among White Democratic legislators compared to African American Democratic legislators. This is because African American Democratic legislators have the countervailing considerations of common interests and civil rights legacies that White legislators are not likely to share. Data and Methods Our dataset includes final votes on all proposed restrictive immigration legislation that received a roll call vote in at least one full chamber for 2005–2012, combining 182 bills that were enacted with 14 bills that were vetoed by the governor, died in committee, or failed to pass in at least one chamber. Hopkins (2010) identifies 2005 as marking the beginning of widespread state and local anti-immigration policymaking. The political context of immigration reform arguably changed after 2012, when the re-election of Barack Obama spurred a renewed effort toward federal immigration reform. We code bills as “restrictive” if they aim to exclude or restrict immigrants from access to resources (jobs, education, housing, etc.), increase surveillance of immigrants and enforcement of immigration laws (expanding police authority), or reinforce the exclusion of immigrants from the polity (voter ID). Bills that refine, clarify, or provide funds for existing restrictive laws are also included in our sample (e.g., participating in the e-verify program or specifying the type of ID that is valid for voting). We exclude resolutions from our sample, as they are not laws and often are not subject to recorded votes. We construct a comprehensive dataset to assess the effects of Democratic legislators’ race, the topic of each bill, and the legislator’s district characteristics on support for restrictive immigration legislation. Our final dataset excludes cases with missing values on our analysis variables, and consists of 998 Democratic legislators casting 12,920 votes on 196 bills. Race of Legislator We compile information on all state legislators serving from 2005 to 2012 in our nine states, coding each legislator’s name, party, chamber, and race. Race is coded through a series of steps, which include referring to the list of African American state legislators published by the National Conference of State Legislators (NCSL) in 2009, examining the organizations and educational institutions with which the legislator was affiliated, and visual inspection of legislator photos. We cross-checked the legislator race codes with a database of Black Democratic state representatives provided by the Joint Center for Political and Economic Research. The five Democratic legislators who are neither Black nor non-Hispanic White or whose race could not be determined are excluded from our analyses. Votes Following Facchini and Steinhardt (2011), we measure support for restrictive immigration legislation by coding final roll call votes on a bill. “Yea” votes were coded 1. “Nay” votes or votes that were recorded as “present,” “abstain,” or “pass” were coded as 0. We coded votes as “missing” for legislators who were recorded as absent when the final vote was taken for a bill. Bill Topic We used the annual reports published by the NCSL to develop a list of all of the restrictive immigration-related legislation enacted from 2005 to 2012 in our nine states. To find the bills that were not enacted, we conducted a search in Lexis/Nexis Statenet, using the same terms used by the NCSL. Only bills that met our criteria for “restrictive” (see above) are included in our final dataset. We code bill topics using the NCSL’s 11 topic codes (see table 1), with further recoding to test our hypotheses. Specifically, we subdivide the “ID/licenses” bills into three categories: 1) licenses and IDs related to employment (driver’s licenses, professional licenses, and business licenses); 2) driver’s licenses not specifically related to employment; and 3) firearms licenses (public safety). The few licenses that do not fall into these categories (e.g., fishing licenses) are included among the “miscellaneous” topics. Bills pertaining to budget items are coded according to the activity targeted by the budget bill. For instance, a bill allocating funds to support the e-verify program is coded under the employment topic. For each restrictive immigration bill in our dataset, we record the roll call vote of each state legislator. We differentiate the bill topics into four groups with respect to their relevance for competition or commonalities issues: the first group of topics (“competition”) involves labor market bills that restrict employment of undocumented immigrants. These include bills that both explicitly target labor market practices of employers or workers and bills that increase the difficulty of working through prohibiting work-related driver’s licenses, professional licenses, and business licenses.5 We categorize proposals regarding voter-identification, expanding police powers, education, and omnibus restrictions as “civil rights” bills likely to evoke a “commonalities” response among African American Democratic legislators. In addition to the continued struggles of civil rights groups to renew and implement the Voting Rights Act of 1965, poor African American voters are over-represented among those without identification to vote. Similarly, concerns about police brutality and racial profiling remain central for African American civil rights organizations (NAACP 2014). Education emerges as the most salient concern for African American Congressional legislators across all issues listed on the NAACP scorecard (Hero and Preuhs 2013, 105, figure 4.I). The final set of topics do not relate to either competition or commonalities frameworks. These include bills on the topics of welfare benefits, health benefits, budget bills, and miscellaneous bills (e.g., fishing licenses, refugee settlement). We also classified “public safety” bills under “other.” This topic operates differently than the labor market threat bills; immigrants are depicted as a “threat to public safety” among Whites but not among Blacks. Further, bills classified under “public safety” appeared in only three states (table 1), so we could not conduct analyses with these bills as a separate category. District Characteristics To test our hypotheses that bill topic conditions the effect of threat-related district characteristics on support for restrictive immigration legislation, we construct multiple measures at the state legislative district level. We calculate the rate of change in the Latino population ([% Latino 2010 – % Latino 2000]/% Latino 2000) within each legislative district using data from the 2000 and 2010 decennial censuses.6 We use change in % Latino rather than % immigrant because the latter is not available in the block-level summary data that we use to construct the district-level change measure. However, we are confident that the change in % Latino captures the dynamics described in the group threat literature. In addition to the high correlation between the percent immigrant and percent Latino variables in our model, there is much evidence to suggest that Latinos are commonly stereotyped as “immigrants”—often “illegal immigrants” (Abascal 2015; Brown 2013). Thus, a rapid increase in the Latino population would be perceived as a rapid increase in the % immigrant. We construct additional measures of the demographic and economic characteristics of the legislator’s district from the American Community Survey aggregated summary data for 2005–2009, 2006–2010, 2007–2011, 2008–2012, and 2009–2013. (Only the aggregated five-year ACS provides data at the state legislative district level.) Where possible, we match the year of the roll call vote with the mid-year of the five-year ACS data. That is, we merge the 2007 legislator data with the 2005–2009 ACS, the 2008 legislator data with the 2006–2010 ACS, the 2009 legislator data with the 2007–2011 ACS, and so on. Due to the availability of the five-year ACS, we merge the 2005 and 2006 legislation data with the 2005–2009 ACS and we merge the 2012 legislator data with the 2009–2013 ACS. These ACS variables are treated as time variant in the models, given that the district demographics change for legislators who changed districts between 2005 and 2012 and all legislators who held office long enough. We construct three district labor market measures from the five-year ACS. The district unemployment rate represents the general economic instability among constituents in our model. To estimate the proportion of the district constituency who may perceive direct labor market competition from immigrants, we calculate the percent of the district workforce employed in the five most common occupations for immigrants: construction, building/grounds cleaning and maintenance, food preparation, production, and transportation (Motel and Patten 2013). We exclude the percent agricultural workers from the immigrant occupation measure, as it is highly correlated with % Latino and % immigrant. Instead, we include a separate measure for the % working in the agriculture industry. This captures both the farmworker presence as well as the potential pressure on legislators from local farmers and agriculture industry lobbies (Nicholson-Crotty and Nicholson-Crotty 2011). Controls To differentiate the pace of change in the Latino population from overall increases in the district’s non-Latino populations, we use the 1990 and 2000 decennial census to construct controls for the change in the Black population and the change in the logarithm of the population. We also include the percent Latino from the ACS as a baseline for the change in the % Latino variable as well as a proxy of the % immigrant. Models using the percent foreign born rather than % Latino yield similar results.7 To gauge the influence of legislator race above and beyond the racial composition of their constituency, we construct an ACS measure of the % Black voting age—Black citizens age 18 and older. We also use the ACS five-year files to create controls for logged median family income. We control for effects of state-level partisan politics by including three measures: a dummy variable indicating whether the Republican Party held a majority in both chambers and the governorship (unified) using data from Monogan (2013) (https://dataverse.harvard.edu/dataverse/monogan); a measure of party polarization within the legislator’s chamber; and a measure of Democratic party cohesion within the chamber. The polarization and cohesion measures are provided by Shor and McCarty (2011) (https://americanlegislatures.com/data/). The Shor and McCarty dataset contains missing values for 2009 and 2010 in Arkansas and Tennessee, and for 2009 in Virginia. We use the state data for 2008 to impute these missing values.8 These party controls account for the possibility that restrictive immigration measures that come to a vote in Republican-controlled states are less likely to gain (or need) Democratic support than those voted on in Democratic-controlled or split states (Chavez and Provine 2009; Monogan 2013; Ramakrishnan and Wong 2010). In addition, legislators may feel more political pressure from their party in states in which there is relatively strong party polarization or cohesion (Pruehs 2006).9 We include fixed effects for each state to account for other state institutional factors that could affect support for restrictive immigration legislation (Monogan 2013; Ramakrishnan and Wong 2010). Estimation To examine how legislators’ race affects support for restrictive immigration bills, our units of analysis are legislator votes on restrictive immigration bills. As legislators often vote on multiple bills during their term, we use a multilevel model for binary responses that accounts for the clustered nature of our data (Guo and Zhao 2000; Raudenbush, Fotiu, and Cheong 1999). Our model consists of two levels: the vote-level model (level 1) that includes all time-variant factors such as district demographics, and a legislator-level model (level 2) that accounts for time-invariant characteristics such as legislators’ race. To assess whether legislator race affects voting depending on the bill topic, we model cross-level interaction effects between legislator race and bill topic. A multilevel approach enables us to distinguish between within- and between-legislator errors, therefore estimates effects more efficiently than an OLS regression would. Moreover, HLMs are better able to deal with unbalanced panels and situations where units have only few observations (i.e., legislators for whom we only observe votes on few bills) (Raudenbush and Bryk 2002). We pool all votes cast between 2005 and 2012 and include state and year fixed effects, to control for geographic and temporal variation. We provide a more detailed explanation of our HLM estimation in the Supplemental Methods Appendix. Hypotheses Testing To test hypotheses 1 and 2, we estimate the effect of legislator race, bill topic, district demographic characteristics, and our controls on support for restrictive immigration bills. To estimate whether the effect of legislator race on support for restrictive immigration legislation depends on whether the bill covers employment issues or civil rights–related issues, we run the model interacting legislator race with bill topic. To test hypotheses 3 and 4, we estimate separate models for the employment bills and the civil rights bills. As we explain above, employment topics include bills focused on restricting immigrant labor market access; civil rights topics include legislation increasing surveillance of immigrants, or restricting immigrant movement or access to resources in relation to the topics of law enforcement, voting, educational opportunities, as well as the omnibus bills. For each bill-topic subsample, hypotheses 3 and 4 are tested with interactions between legislator race and district characteristics. In addition to the potential for confounding coefficients with residual variation in logistic regression, the nonlinear functional form of the logistic regression model creates a situation in which interactions can be significant within some ranges of the independent variable but not in others. To address these issues, we follow the procedures developed by Long (Long and Freese 2014). For each bill-topic subsample, we estimate (AME) marginal probabilities with the logits generated in a model with interactions of legislator race with the district demographic variables (Supplemental Table iv). We fix the independent variable of interest at the mean and at additional values above and below the mean, holding the other independent variables at their observed values for each observation (Long and Freese 2014). We use the delta method to test for whether the probabilities for these two “conditions” are statistically significant. A full explanation of the estimation procedure is available in Long and Freese (2014). It is not possible to conduct statistical tests for differences in the coefficients across models, as the coefficient can be confounded by residual variation in logistic regression (Long and Freese 2014) and because our samples are not independent. In additional analyses, we therefore combine the employment and civil rights subsamples and test for significant three-way interactions between bill topic (employment or civil rights), legislator race (Black or White), and each of the independent variables. These analyses confirm the findings we report below (results available upon request). Results Descriptive Analyses Although the majority of Blacks favor restrictive immigration legislation (71 percent of Black votes are “yea”), this is significantly lower than the 84 percent of “yea” votes among their White Democratic counterparts (table 2). The literature suggests that the majority of White and Black representatives support restrictive immigration legislation for several reasons. First, only bills that will be enacted are typically brought up for a vote. Second, in the post-9/11 era, restrictive immigration legislation is popular among Democratic voters (Masuoka and Junn 2013). Third, in our nine Southern states, there are no organized groups of Latino legislators or sizeable concentrations of Latino constituents motivating Democratic representatives to oppose restrictive immigration legislation. Indeed, it is precisely this freedom from Latino influence that provides an important aspect of our paper. Table 2. Percent “Yea” Votes on Restrictive Immigration Legislation by Legislator Race and Bill Topic Black Dems White Dems GOP Bl-Wh Dems: Chi-sq, df Wh Dems-GOP: Chi-sq, df Bill topic Labor market  Employment 77% 87% 93% 60.5, 1*** 68.82,1***  Licenses (work-related) 84% 90% 90% 8.01, 1** 0.002,1 Civil rights  Police 76% 89% 93% 83.5, 1*** 21.5,1***  Voting 32% 49% 97% 19.2, 1*** 333.6,1***  Education 71% 85% 96% 36.3, 1*** 42.7,1***  Omnibus 34% 58% 89% 27.5, 1*** 102.3,1*** Other topics  Benefits 72% 77% 92% 2.3, 1 76.7,1***  Budget 63% 77% 86% 9.3, 1** 6.3,1*  DL’s 81% 87% 92% 5.9, 1* 8.9,1**  Safety 67% 89% 94% 33.9, 1*** 10.2,1**  Misc 65% 76% 90% 11.3, 1 43.4,1*** Total 71% 84% 93% 307.5, 1*** 401.4,1***  N (total) 5,303 7,617 12,884 Black Dems White Dems GOP Bl-Wh Dems: Chi-sq, df Wh Dems-GOP: Chi-sq, df Bill topic Labor market  Employment 77% 87% 93% 60.5, 1*** 68.82,1***  Licenses (work-related) 84% 90% 90% 8.01, 1** 0.002,1 Civil rights  Police 76% 89% 93% 83.5, 1*** 21.5,1***  Voting 32% 49% 97% 19.2, 1*** 333.6,1***  Education 71% 85% 96% 36.3, 1*** 42.7,1***  Omnibus 34% 58% 89% 27.5, 1*** 102.3,1*** Other topics  Benefits 72% 77% 92% 2.3, 1 76.7,1***  Budget 63% 77% 86% 9.3, 1** 6.3,1*  DL’s 81% 87% 92% 5.9, 1* 8.9,1**  Safety 67% 89% 94% 33.9, 1*** 10.2,1**  Misc 65% 76% 90% 11.3, 1 43.4,1*** Total 71% 84% 93% 307.5, 1*** 401.4,1***  N (total) 5,303 7,617 12,884 Note: Unit of analysis is the individual vote cast. Because the number of bills and legislators vary by state, the n’s vary by topic. The number of votes cast by Black Democratic legislators range from 163 (budget) to 1,544 (employment). The number of votes cast by White legislators range from 198 (budget) to 2,164 (employment). The range for Republican votes is 282 (budget) to 5,343 (employment). Of the 1,360 Republican legislators casting votes in our sample, all but eight are White. ***p < 0.001 **p < 0.01 *p < 0.05 Table 2. Percent “Yea” Votes on Restrictive Immigration Legislation by Legislator Race and Bill Topic Black Dems White Dems GOP Bl-Wh Dems: Chi-sq, df Wh Dems-GOP: Chi-sq, df Bill topic Labor market  Employment 77% 87% 93% 60.5, 1*** 68.82,1***  Licenses (work-related) 84% 90% 90% 8.01, 1** 0.002,1 Civil rights  Police 76% 89% 93% 83.5, 1*** 21.5,1***  Voting 32% 49% 97% 19.2, 1*** 333.6,1***  Education 71% 85% 96% 36.3, 1*** 42.7,1***  Omnibus 34% 58% 89% 27.5, 1*** 102.3,1*** Other topics  Benefits 72% 77% 92% 2.3, 1 76.7,1***  Budget 63% 77% 86% 9.3, 1** 6.3,1*  DL’s 81% 87% 92% 5.9, 1* 8.9,1**  Safety 67% 89% 94% 33.9, 1*** 10.2,1**  Misc 65% 76% 90% 11.3, 1 43.4,1*** Total 71% 84% 93% 307.5, 1*** 401.4,1***  N (total) 5,303 7,617 12,884 Black Dems White Dems GOP Bl-Wh Dems: Chi-sq, df Wh Dems-GOP: Chi-sq, df Bill topic Labor market  Employment 77% 87% 93% 60.5, 1*** 68.82,1***  Licenses (work-related) 84% 90% 90% 8.01, 1** 0.002,1 Civil rights  Police 76% 89% 93% 83.5, 1*** 21.5,1***  Voting 32% 49% 97% 19.2, 1*** 333.6,1***  Education 71% 85% 96% 36.3, 1*** 42.7,1***  Omnibus 34% 58% 89% 27.5, 1*** 102.3,1*** Other topics  Benefits 72% 77% 92% 2.3, 1 76.7,1***  Budget 63% 77% 86% 9.3, 1** 6.3,1*  DL’s 81% 87% 92% 5.9, 1* 8.9,1**  Safety 67% 89% 94% 33.9, 1*** 10.2,1**  Misc 65% 76% 90% 11.3, 1 43.4,1*** Total 71% 84% 93% 307.5, 1*** 401.4,1***  N (total) 5,303 7,617 12,884 Note: Unit of analysis is the individual vote cast. Because the number of bills and legislators vary by state, the n’s vary by topic. The number of votes cast by Black Democratic legislators range from 163 (budget) to 1,544 (employment). The number of votes cast by White legislators range from 198 (budget) to 2,164 (employment). The range for Republican votes is 282 (budget) to 5,343 (employment). Of the 1,360 Republican legislators casting votes in our sample, all but eight are White. ***p < 0.001 **p < 0.01 *p < 0.05 However, as our separate analyses by topic show, these aggregate numbers gloss over a more complex picture of Black-White differences in support for restrictive immigration legislation among Democrats (table 2). African American Democratic legislators are most likely to support “competition” bills restricting immigrant labor market access and are least likely to support voter ID, a civil rights topic. Republican support for restrictive immigration is significantly higher than White Democratic support on almost all bill topics; overall, 93 percent of Republicans support restrictive immigration bills (table 2). Importantly, while omnibus bills garner high support from Republicans (89 percent), these bills receive the second lowest level of support from Democrats (34 percent of Blacks, 58 percent of Whites). Thus, restricting our multivariate analyses to Democrats provides a conservative estimate of Black-White legislative differences in responses to state immigration legislation (1,435 Republicans in our sample years are White; only four Republicans are African American, one Republican is Latino, and three Republicans are Asian). To assess the underlying assumption in theories of symbolic politics that African American Democratic opposition to restrictive immigration legislation will be based on common values and goals beyond material group interest, we coded the civil rights bills according to whether opposing the bill would be in the direct group interest of African Americans. Bills negatively affecting African American group interest such as requiring voter ID were coded as 1 for self-interest; bills that would not negatively affect African Americans directly, such as requiring that noncitizens pay out-of-state tuition, were coded as 0 (see Supplemental Table i for examples). Approximately 80 percent of the restrictive bills covering civil rights–related topics did not negatively affect African Americans. Thus, we interpret the African American “commonalities” response—that is, their lack of support for these bills on civil rights topics—as motivated by the symbolic importance of civil rights–related issues, as the theory of symbolic politics would predict. Descriptive analyses show that district characteristics vary among Democratic legislators along dimensions described in the literature (Hopkins 2010) (we report these results in Supplemental Table ii). Black Democrats represent majority-Black districts, with the average black voting age population comprising 56 percent of their district population. Importantly for our study, many White Democrats also represent districts with Black constituents. On average, voting-age Blacks comprised 23 percent of the population in districts represented by White Democrats. In addition, both the Black and White Democratic legislators saw a doubling of % Latino in the district between 2000 and 2010. We turn to the multivariate analyses to understand how these district characteristics along with bill topic affect Black-White differences in Democratic support for restrictive immigration legislation. Multivariate Analyses Our multivariate model in table 3 shows that Black Democratic legislators are significantly less likely than their White Democratic peers to support restrictive immigration legislation regardless of topic, even with controls for relevant district characteristics and political context. When we consider Black-White differences by bill topic, however, a more complex picture emerges. Table 3. Effect of Legislator Race, Bill Topic, and District Characteristics on the Log Odds of Supporting Restrictive Immigration Legislation logit s.e. Black legislator −0.341 0.90*** Bill topics (Employment = omitted category) Labor market topics  Licenses/IDs 0.665 0.126*** Civil rights topics  Police −0.260 0.085**  Voting −1.512 0.118***  Education 0.167 0.101  Omnibus −1.871 0.130*** Other topics  Public benefits −1.225 0.128***  Budget −0.046 0.150  Public safety 0.517 0.142***  Misc −0.906 0.114*** District characteristics  % Chg Latino pop 0.088 0.044*  % Latino −1.161 0.534*  % Black −0.499 0.243*  % Chg Black pop −0.001 0.001  % Immigrant occ 2.427 0.940**  % Agr industry 3.958 1.935*  % Unemployed −2.435 1.467  Ln Faminc −0.016 0.209  Chg ln pop −0.191 0.360 Political context  Unified Republican govt. −1.664 0.134***  Party polarization (chamber) 0.016 0.316  Party cohesion (chamber) 0.593 0.456 State fixed effects Yes Year fixed effects Yes N (votes) 12,920 N (legislators) 999 BIC 10,363.6 logit s.e. Black legislator −0.341 0.90*** Bill topics (Employment = omitted category) Labor market topics  Licenses/IDs 0.665 0.126*** Civil rights topics  Police −0.260 0.085**  Voting −1.512 0.118***  Education 0.167 0.101  Omnibus −1.871 0.130*** Other topics  Public benefits −1.225 0.128***  Budget −0.046 0.150  Public safety 0.517 0.142***  Misc −0.906 0.114*** District characteristics  % Chg Latino pop 0.088 0.044*  % Latino −1.161 0.534*  % Black −0.499 0.243*  % Chg Black pop −0.001 0.001  % Immigrant occ 2.427 0.940**  % Agr industry 3.958 1.935*  % Unemployed −2.435 1.467  Ln Faminc −0.016 0.209  Chg ln pop −0.191 0.360 Political context  Unified Republican govt. −1.664 0.134***  Party polarization (chamber) 0.016 0.316  Party cohesion (chamber) 0.593 0.456 State fixed effects Yes Year fixed effects Yes N (votes) 12,920 N (legislators) 999 BIC 10,363.6 ***p < 0.001 **p < 0.01 *p < 0.05 Table 3. Effect of Legislator Race, Bill Topic, and District Characteristics on the Log Odds of Supporting Restrictive Immigration Legislation logit s.e. Black legislator −0.341 0.90*** Bill topics (Employment = omitted category) Labor market topics  Licenses/IDs 0.665 0.126*** Civil rights topics  Police −0.260 0.085**  Voting −1.512 0.118***  Education 0.167 0.101  Omnibus −1.871 0.130*** Other topics  Public benefits −1.225 0.128***  Budget −0.046 0.150  Public safety 0.517 0.142***  Misc −0.906 0.114*** District characteristics  % Chg Latino pop 0.088 0.044*  % Latino −1.161 0.534*  % Black −0.499 0.243*  % Chg Black pop −0.001 0.001  % Immigrant occ 2.427 0.940**  % Agr industry 3.958 1.935*  % Unemployed −2.435 1.467  Ln Faminc −0.016 0.209  Chg ln pop −0.191 0.360 Political context  Unified Republican govt. −1.664 0.134***  Party polarization (chamber) 0.016 0.316  Party cohesion (chamber) 0.593 0.456 State fixed effects Yes Year fixed effects Yes N (votes) 12,920 N (legislators) 999 BIC 10,363.6 logit s.e. Black legislator −0.341 0.90*** Bill topics (Employment = omitted category) Labor market topics  Licenses/IDs 0.665 0.126*** Civil rights topics  Police −0.260 0.085**  Voting −1.512 0.118***  Education 0.167 0.101  Omnibus −1.871 0.130*** Other topics  Public benefits −1.225 0.128***  Budget −0.046 0.150  Public safety 0.517 0.142***  Misc −0.906 0.114*** District characteristics  % Chg Latino pop 0.088 0.044*  % Latino −1.161 0.534*  % Black −0.499 0.243*  % Chg Black pop −0.001 0.001  % Immigrant occ 2.427 0.940**  % Agr industry 3.958 1.935*  % Unemployed −2.435 1.467  Ln Faminc −0.016 0.209  Chg ln pop −0.191 0.360 Political context  Unified Republican govt. −1.664 0.134***  Party polarization (chamber) 0.016 0.316  Party cohesion (chamber) 0.593 0.456 State fixed effects Yes Year fixed effects Yes N (votes) 12,920 N (legislators) 999 BIC 10,363.6 ***p < 0.001 **p < 0.01 *p < 0.05 We test hypotheses 1 and 2 that the Black-White difference in support for restrictive immigration legislation depends on the bill topic by estimating interactions of legislator race with bill topic in table 4. For ease of interpretation, we present the results as predicted probabilities. (See Supplemental Table ii for coefficients used to estimate probabilities.) In support of hypothesis 1, Black and White Democrats are equally likely to vote in favor of immigration bills restricting labor market competition (employment and employment-related ID/licensing topics). In support of hypothesis 2, African American Democratic legislators are significantly less likely than White Democratic legislators to cast “yea” votes for legislation covering the civil rights topics of voting, law enforcement, education, and omnibus restrictions.10 Table 4. Probability of Supporting Restrictive Immigration Legislation by Bill Topic and Legislator Race Black White Bl-Wh diff Bill topics Labor market  Employment 0.793 0.819 −0.026  Licenses (work related) 0.856 0.857 0.001 Civil rights  Police 0.707 0.810 −0.103***  Voting 0.489 0.640 −0.151***  Education 0.767 0.837 −0.070**  Omnibus 0.448 0.565 −0.118*** Other topics  Public benefits 0.620 0.629 −0.008  Budget 0.748 0.810 −0.062  Driver’s license 0.856 0.871 −0.015  Public safety 0.785 0.889 −0.104***  Misc 0.652 0.697 −0.045 N: 12,920 BIC: 10,419.8 Black White Bl-Wh diff Bill topics Labor market  Employment 0.793 0.819 −0.026  Licenses (work related) 0.856 0.857 0.001 Civil rights  Police 0.707 0.810 −0.103***  Voting 0.489 0.640 −0.151***  Education 0.767 0.837 −0.070**  Omnibus 0.448 0.565 −0.118*** Other topics  Public benefits 0.620 0.629 −0.008  Budget 0.748 0.810 −0.062  Driver’s license 0.856 0.871 −0.015  Public safety 0.785 0.889 −0.104***  Misc 0.652 0.697 −0.045 N: 12,920 BIC: 10,419.8 Note: Probabilities estimated from two-level model with interactions of legislator race with bill topic, main effects for district characteristics and unified Republican government, party polarization and party cohesion, and state fixed effects and year fixed effects. (Coefficients reported in Supplemental Table iii.) See text for explanation of how marginal probabilities are calculated. ***p < 0.001 **p < 0.01 *p < 0.05 Table 4. Probability of Supporting Restrictive Immigration Legislation by Bill Topic and Legislator Race Black White Bl-Wh diff Bill topics Labor market  Employment 0.793 0.819 −0.026  Licenses (work related) 0.856 0.857 0.001 Civil rights  Police 0.707 0.810 −0.103***  Voting 0.489 0.640 −0.151***  Education 0.767 0.837 −0.070**  Omnibus 0.448 0.565 −0.118*** Other topics  Public benefits 0.620 0.629 −0.008  Budget 0.748 0.810 −0.062  Driver’s license 0.856 0.871 −0.015  Public safety 0.785 0.889 −0.104***  Misc 0.652 0.697 −0.045 N: 12,920 BIC: 10,419.8 Black White Bl-Wh diff Bill topics Labor market  Employment 0.793 0.819 −0.026  Licenses (work related) 0.856 0.857 0.001 Civil rights  Police 0.707 0.810 −0.103***  Voting 0.489 0.640 −0.151***  Education 0.767 0.837 −0.070**  Omnibus 0.448 0.565 −0.118*** Other topics  Public benefits 0.620 0.629 −0.008  Budget 0.748 0.810 −0.062  Driver’s license 0.856 0.871 −0.015  Public safety 0.785 0.889 −0.104***  Misc 0.652 0.697 −0.045 N: 12,920 BIC: 10,419.8 Note: Probabilities estimated from two-level model with interactions of legislator race with bill topic, main effects for district characteristics and unified Republican government, party polarization and party cohesion, and state fixed effects and year fixed effects. (Coefficients reported in Supplemental Table iii.) See text for explanation of how marginal probabilities are calculated. ***p < 0.001 **p < 0.01 *p < 0.05 To test hypotheses 3 and 4 that district characteristics further condition the relationship between bill topic and legislator race previously examined in table 4, we estimate separate models for the labor market bills and civil rights bills and include an interaction between legislator race and district characteristics for each model. We report the results of the tests for Black-White differences in probabilities in table 5 and figures 1–6. (See Supplemental Table iii for coefficients used to estimate probabilities.) In the graphs, the effects of a selected independent variable differ for Black and White legislators when the confidence intervals (represented by shaded bands) for each group do not overlap. Table 5. Probability of Supporting Restrictive Immigration Legislation, by Demographic Characteristics and Legislator Race: Labor Market Topics and Civil Rights Topics Labor market topics Civil rights topics Black White Bl-Wh diff Black White Bl-Wh diff District Characteristics  % Change Latino pop   0% 0.836 0.866 −0.030 0.698 0.737 −0.039   50% 0.867 0.845 −0.022 0.696 0.753 −0.058**   100% 0.854 0.869 −0.015 0.693 0.770 −0.076***   150% 0.863 0.870 −0.007 0.691 0.785 −0.095***  % Latino   0% 0.861 0.871 −0.010 0.692 0.770 −0.078***   5% 0.856 0.865 −0.009 0.693 0.767 −0.073***   10% 0.851 0.859 −0.008 0.694 0.763 −0.069**  % Black   0% 0.924 0.886 0.048 0.704 0.807 −0.103*   30% 0.880 0.874 0.011 0.696 0.776 −0.080***   75% 0.781 0.854 −0.043* 0.687 0.742 −0.055*  % Immigrant occ   0% 0.733 0.817 −0.086 0.599 0.620 −0.006   20% 0.852 0.860 −0.013 0.689 0.762 −0.070***   40% 0.909 0.928 0.031 0.769 0.869 −0.103**  % Agriculture industry   0% 0.858 0.866 −0.009 0.685 0.754 −0.067**   10% 0.809 0.851 −0.042 0.747 0.837 −0.090   20% 0.751 0.835 −0.083 0.801 0.899 −0.097  % Unemployed   0% 0.807 0.854 −0.047 0.784 0.771 0.136   10% 0.845 0.867 −0.022 0.702 0.766 −0.064**   25% 0.893 0.884 0.009 0.559 0.759 −0.200*** N (votes) 4,618 5,187 N (legislators) 882 888 BIC 3989.2 4458.6 Labor market topics Civil rights topics Black White Bl-Wh diff Black White Bl-Wh diff District Characteristics  % Change Latino pop   0% 0.836 0.866 −0.030 0.698 0.737 −0.039   50% 0.867 0.845 −0.022 0.696 0.753 −0.058**   100% 0.854 0.869 −0.015 0.693 0.770 −0.076***   150% 0.863 0.870 −0.007 0.691 0.785 −0.095***  % Latino   0% 0.861 0.871 −0.010 0.692 0.770 −0.078***   5% 0.856 0.865 −0.009 0.693 0.767 −0.073***   10% 0.851 0.859 −0.008 0.694 0.763 −0.069**  % Black   0% 0.924 0.886 0.048 0.704 0.807 −0.103*   30% 0.880 0.874 0.011 0.696 0.776 −0.080***   75% 0.781 0.854 −0.043* 0.687 0.742 −0.055*  % Immigrant occ   0% 0.733 0.817 −0.086 0.599 0.620 −0.006   20% 0.852 0.860 −0.013 0.689 0.762 −0.070***   40% 0.909 0.928 0.031 0.769 0.869 −0.103**  % Agriculture industry   0% 0.858 0.866 −0.009 0.685 0.754 −0.067**   10% 0.809 0.851 −0.042 0.747 0.837 −0.090   20% 0.751 0.835 −0.083 0.801 0.899 −0.097  % Unemployed   0% 0.807 0.854 −0.047 0.784 0.771 0.136   10% 0.845 0.867 −0.022 0.702 0.766 −0.064**   25% 0.893 0.884 0.009 0.559 0.759 −0.200*** N (votes) 4,618 5,187 N (legislators) 882 888 BIC 3989.2 4458.6 Note: Probabilities estimated from two-level logit models run separately for labor market and civil rights topics, and controlling for unified Republican state govt. and party polarization and party cohesion within the legislator’s chamber, with state and year fixed effects. (Coefficients reported in Supplementary Table iv.) See text for explanation of how marginal probabilities are calculated. ***p < 0.001 **p < 0.01 *p < 0.05. Table 5. Probability of Supporting Restrictive Immigration Legislation, by Demographic Characteristics and Legislator Race: Labor Market Topics and Civil Rights Topics Labor market topics Civil rights topics Black White Bl-Wh diff Black White Bl-Wh diff District Characteristics  % Change Latino pop   0% 0.836 0.866 −0.030 0.698 0.737 −0.039   50% 0.867 0.845 −0.022 0.696 0.753 −0.058**   100% 0.854 0.869 −0.015 0.693 0.770 −0.076***   150% 0.863 0.870 −0.007 0.691 0.785 −0.095***  % Latino   0% 0.861 0.871 −0.010 0.692 0.770 −0.078***   5% 0.856 0.865 −0.009 0.693 0.767 −0.073***   10% 0.851 0.859 −0.008 0.694 0.763 −0.069**  % Black   0% 0.924 0.886 0.048 0.704 0.807 −0.103*   30% 0.880 0.874 0.011 0.696 0.776 −0.080***   75% 0.781 0.854 −0.043* 0.687 0.742 −0.055*  % Immigrant occ   0% 0.733 0.817 −0.086 0.599 0.620 −0.006   20% 0.852 0.860 −0.013 0.689 0.762 −0.070***   40% 0.909 0.928 0.031 0.769 0.869 −0.103**  % Agriculture industry   0% 0.858 0.866 −0.009 0.685 0.754 −0.067**   10% 0.809 0.851 −0.042 0.747 0.837 −0.090   20% 0.751 0.835 −0.083 0.801 0.899 −0.097  % Unemployed   0% 0.807 0.854 −0.047 0.784 0.771 0.136   10% 0.845 0.867 −0.022 0.702 0.766 −0.064**   25% 0.893 0.884 0.009 0.559 0.759 −0.200*** N (votes) 4,618 5,187 N (legislators) 882 888 BIC 3989.2 4458.6 Labor market topics Civil rights topics Black White Bl-Wh diff Black White Bl-Wh diff District Characteristics  % Change Latino pop   0% 0.836 0.866 −0.030 0.698 0.737 −0.039   50% 0.867 0.845 −0.022 0.696 0.753 −0.058**   100% 0.854 0.869 −0.015 0.693 0.770 −0.076***   150% 0.863 0.870 −0.007 0.691 0.785 −0.095***  % Latino   0% 0.861 0.871 −0.010 0.692 0.770 −0.078***   5% 0.856 0.865 −0.009 0.693 0.767 −0.073***   10% 0.851 0.859 −0.008 0.694 0.763 −0.069**  % Black   0% 0.924 0.886 0.048 0.704 0.807 −0.103*   30% 0.880 0.874 0.011 0.696 0.776 −0.080***   75% 0.781 0.854 −0.043* 0.687 0.742 −0.055*  % Immigrant occ   0% 0.733 0.817 −0.086 0.599 0.620 −0.006   20% 0.852 0.860 −0.013 0.689 0.762 −0.070***   40% 0.909 0.928 0.031 0.769 0.869 −0.103**  % Agriculture industry   0% 0.858 0.866 −0.009 0.685 0.754 −0.067**   10% 0.809 0.851 −0.042 0.747 0.837 −0.090   20% 0.751 0.835 −0.083 0.801 0.899 −0.097  % Unemployed   0% 0.807 0.854 −0.047 0.784 0.771 0.136   10% 0.845 0.867 −0.022 0.702 0.766 −0.064**   25% 0.893 0.884 0.009 0.559 0.759 −0.200*** N (votes) 4,618 5,187 N (legislators) 882 888 BIC 3989.2 4458.6 Note: Probabilities estimated from two-level logit models run separately for labor market and civil rights topics, and controlling for unified Republican state govt. and party polarization and party cohesion within the legislator’s chamber, with state and year fixed effects. (Coefficients reported in Supplementary Table iv.) See text for explanation of how marginal probabilities are calculated. ***p < 0.001 **p < 0.01 *p < 0.05. Figure 1. View largeDownload slide Probability of supporting bills with a labor market topic by legislator race and % change in the Latino population Figure 1. View largeDownload slide Probability of supporting bills with a labor market topic by legislator race and % change in the Latino population Figure 2. View largeDownload slide Probability of supporting bills with a labor market topic by legislator race and % district employed in immigrant occupations Figure 2. View largeDownload slide Probability of supporting bills with a labor market topic by legislator race and % district employed in immigrant occupations Figure 3. View largeDownload slide Probability of supporting bills with a labor market topic by legislator race and % unemployed in district Figure 3. View largeDownload slide Probability of supporting bills with a labor market topic by legislator race and % unemployed in district Figure 4. View largeDownload slide Probability of supporting bills with a civil rights topic by legislator race and % change in the Latino population Figure 4. View largeDownload slide Probability of supporting bills with a civil rights topic by legislator race and % change in the Latino population Figure 5. View largeDownload slide Probability of supporting bills with a civil rights topic by legislator race and % district employed in immigrant occupations Figure 5. View largeDownload slide Probability of supporting bills with a civil rights topic by legislator race and % district employed in immigrant occupations Figure 6. View largeDownload slide Probability of supporting bills with a civil rights topic by legislator race and % unemployed in district Figure 6. View largeDownload slide Probability of supporting bills with a civil rights topic by legislator race and % unemployed in district When we consider the labor market legislation, none of the district threat characteristics significantly affect the votes of either Black or White Democratic legislators (figures 1–3 and table 5). These results support hypothesis 3 that there will be no racial differences in how threat conditions affect legislators’ support for labor market bills. In contrast, when civil rights–related issues are on the floor, district threat characteristics evoke significantly stronger support among White Democratic legislators compared to their African American colleagues. Figure 4 shows that our key demographic indicator of “threat”—the % change in the district Latino population—increases the likelihood of supporting restrictive civil rights bills only among White legislators. A rapid rise in % Latino slightly decreases support for civil rights legislation among African American Democratic legislators. These Black-White differences in the effect of changes in the district Latino population are statistically significant, supporting hypothesis 4 (table 5). Consistent with group threat theories, larger percentages of constituents employed in immigrant occupations are associated with increased support for restrictive civil rights bills among both White and African American Democrats. However, this positive effect is significantly stronger among White Democratic legislators compared to their Black colleagues (table 5, figure 5). This too supports hypothesis 4. The effects of district unemployment also differ according to legislator race, but not in ways entirely anticipated by our hypotheses. Increases in district unemployment have no significant effect on the likelihood that White Democratic legislators will support restrictive civil rights bills. In contrast, higher rates of constituent unemployment decrease support for such civil rights measures among African American Democratic legislators (table 5, figure 6). This negative effect among Black Democratic legislators is the opposite of what we would expect in anticipation of a “threat” response. Given the possibility that districts with relatively high unemployment rates include even greater numbers of Black and/or poor residents vulnerable to voter ID laws and enhanced law enforcement, our results suggest that, in this context of proposed restrictions on civil rights, what is often considered an indicator of threat could actually be a source of commonality among many “Nuevo South” African American Democratic legislators. Conclusion Overall, the results offer considerable support for our hypotheses and their underlying logic. When the bill topic restricts immigrant employment, Black and White Democratic legislators maintain strong, equal support for these labor market “threat” measures (or “initiatives”). But when the bill topic signals that civil rights may be at stake, Black Democratic legislators are less likely to support the measure than White Democrats. Further, district-level “threat” indicators appear to affect African American and White legislators differently. More often than not, White Democratic legislators react strongly, as if threatened, by increasing their support for such restrictions. Black legislative responses are relatively weak and inconsistent, suggesting some uncertainty or ambivalence about what constitutes a threat or a commonality. Nevertheless, Black Democratic legislators consistently support restrictive immigration bills that touch on civil rights issues at lower rates than their White counterparts. Although the majority of African American Democratic legislators support restrictive immigration legislation on all but two topics (omnibus and voting), they do so at lower rates than White Democratic representatives. Indeed, if White Democrats voted like Black Democrats, the restrictive immigration legislation would still pass, given that Republicans were typically the majority party from 2005 to 2012 in every state but Arkansas and Louisiana. Despite the influence (or lack thereof) of Democratic votes on immigration bills, our study significantly advances our understanding of how the US racial hierarchy influences state policymaking. Our findings support theories of group threat and symbolic politics, suggesting that Black Democratic legislators’ perceptions of their group position in relation to Whites and Latinos, respectively, can engender legislative responses in line with group competition or with common goals grounded in the continued struggle for racial justice. We contend that the sociological significance of our study to issues of immigration, race, and politics extends beyond theories of competition and commonalities, however. Omi and Winant’s (1994) theory of racial formation suggests that immigration policy contributes to the construction of the US racial hierarchy (Ngai 2004). By legally defining who “belongs” and who does not, immigration policy creates legal and social boundaries between racialized categories and produces meanings behind those categories (Ngai 2004). For instance, the Trump administration’s proposal to build a wall at the US-Mexico border symbolically and legally designates the racialized group of “Latinos” as “not belonging” (Brown 2012; Masuoka and Junn 2013). To date, scholarship on immigration policy and race focuses on how White elites use immigration policy to maintain their dominant social position in the US racial hierarchy (Ngai 2004). Our analyses take this literature in a new direction, showing how African American Democratic legislators, representing a subordinate racial group, participate in this process of racialization through state policymaking. We show how race and symbolic politics are connected in shaping elite behavior toward immigration policy, and we identify an empirical tool (analyses by bill topic) to test hypotheses about these connections. Our focus on roll call votes to assess the role of legislator race in immigration policymaking does carry certain limitations. Only legislation that can garner sufficient support to pass is brought to the floor for a vote. Thus, the set of bills in our study is highly selective and constrained, representing just a small part of the legislative process. However, there are also advantages to analyzing roll call votes. In the public record, roll call votes are the most visible aspect of legislative behavior. Indeed, public interest groups maintain scorecards on legislator voting records on immigration legislation. Also, the outcome of roll call votes is arguably the most consequential. Enacted immigration laws institutionalize and codify symbolic boundaries, directly affecting the lives of immigrants (Ngai 2004). To the extent that these laws and their enforcement operate through targeting Latinos, restrictive immigration policies contribute to the positioning of Latinos in the US racial hierarchy (Masuoka and Junn 2013; Ngai 2004). Our analyses of roll call votes provide important direction for future research outside the “Nuevo South.” In particular, future research should address the questions of what circumstances are likely to push Black legislators toward a commonalities response for economic bills, and what circumstances are likely to push Black legislators toward a competition response to bills covering civil rights issues. One of the most important political conditions to test in future work involve the presence and power of Latino elected officials and the communities they represent. Notes 1 Given that almost half (44 percent) of all Black Democratic state legislators in the United States are located in our nine states, these states are of substantive importance as well, representing an epicenter of Black state immigration politics. 2 The two most politically impactful ways that White Democrats differ from White Republicans are the differences in political ideology and the much higher presence of Black constituents in the districts represented by White Democrats compared to White Republicans. 3 “Independence” occurs when Black Democratic legislators do not give Congressional testimony on issues promoted by the Latino Caucus. Our roll call data do not allow us to determine an “independence” effect. 4 A full 86 percent of Black representatives in our sample represented majority-Black districts; 11 percent of White representatives in our sample also represented majority-Black districts. 5 We also ran analyses including all driver’s license topics as “competition” measures. This changed the coefficients for licenses, but did not alter the main findings regarding the competition bills that we report in the paper. 6 Because the geographic boundaries for the state legislative districts and the census blocks changed between 2000 and 2010 in the census, we constructed our measure of % change in the Latino and total population by creating equivalent district boundaries. 7 The percent foreign born and percent Hispanic measures are correlated at 0.87 in our dataset. 8 The 2008 data most closely match the composition of the state assembly in these years for these states. 9 We thank an anonymous reviewer for pointing this out to us. 10 We report results from additional tests to check for the robustness of our findings in the Supplemental Methods Appendix. Supplementary Material Supplementary material is available at Social Forces online. About the Authors Irene Browne is Associate Professor of Sociology at Emory University. Her research areas include intersectionality (race, class, gender); immigration; and labor market inequality. She is editor of the book Latinas and African-American Women at Work: Race, Gender and Economic Inequality. She has also published articles in the American Sociological Review, Annual Review of Sociology, Social Forces, Sociological Quarterly, American Behavioral Scientist, and other journals. Beth Reingold is Associate Professor of Political Science and Women’s, Gender, and Sexuality Studies at Emory University. Her research on gender, race, and the politics of representation and identity in the United States has appeared most recently in the American Journal of Political Science, Political Research Quarterly, Politics & Gender, and Representation: The Case of Women (edited by Maria Escobar-Lemmon and Michelle Taylor-Robinson, 2014). Anne-Kathrin Kronberg is an Assistant Professor in the Sociology Department at the University of North Carolina–Charlotte. Her research focuses on organizations, work, and social inequality. Her work has appeared in Social Forces, Work and Occupations, and Mobilization and has been supported by the German Research Foundation. References Abascal , Maria . 2015 . “ Us and Them: Black-White Relations in the Wake of Hispanic Population Growth .” American Sociological Review 80 ( 4 ): 789 – 813 . Google Scholar CrossRef Search ADS Black , Earl , and Merle Black . 1989 . Politics and Society in the South . Cambridge, MA : Harvard University Press . Blalock , Hubert . 1967 . Toward a Theory of Minority-Group Relations . New York : Wiley . Blumer , Herbert . 1958 . “ Race Prejudice as a Sense of Group Position .” Pacific Sociological Review 1 : 3 – 7 . Google Scholar CrossRef Search ADS Bobo , Lawrence , and Vincent Hutchings . 1996 . “ Perceptions of Racial Group Competition: Extending Blumer’s Theory of Group Competition in a Multiracial Social Context .” American Sociological Review 61 ( 6 ): 951 – 72 . Google Scholar CrossRef Search ADS Bositis , David . 2011 . “Resegregation in Southern Politics?” Research Report, November. Washington, DC: Joint Center for Political and Economic Studies. Brown , Hana . 2013 . “ Race, Legality and the Social Policy Consequences of Anti-Immigration Mobilization .” American Sociological Review 78 ( 2 ): 290 – 314 . Google Scholar CrossRef Search ADS Canon , David . 1999 . Race, Redistricting, and Representation: The Unintended Consequences of Black Majority Districts . Chicago : University of Chicago Press . Ceobanu , Alin , and Xavier Escandell . 2010 . “ Comparative Analyses of Public Attitudes Toward Immigrants and Immigration Using Multinational Survey Data: A Review of Theories and Research .” Annual Review of Sociology 36 : 309 – 28 . Google Scholar CrossRef Search ADS Chavez , Jorge , and Doris Provine . 2009 . “ Race and the Response of State Legislatures to Unauthorized Immigrants .” Annals of the American Academy of Political and Social Science 623 ( 1 ): 78 – 92 . Google Scholar CrossRef Search ADS Chavez , Leo . 2013 . The Latino Threat Narrative , 2nd ed . Palo Alto, CA : Stanford University Press . Citrin , Jack , Donald Green , Christopher Muste , and Cara Wong . 1997 . “ Public Opinion toward Immigration Reform: The Role of Economic Motivations .” Journal of Politics 59 : 858 – 81 . Google Scholar CrossRef Search ADS Citrin , Jack , Beth Reingold , and Donald Green . 1990 . “ American Identity and the Politics of Ethnic Change .” The Journal of Politics 52 : 1124 – 1154 . Google Scholar CrossRef Search ADS Clawson , Daniel . 1989 . “ Interlocks, PACS, and Corporate Conservatism .” American Journal of Sociology 94 ( 4 ): 749 – 73 . Google Scholar CrossRef Search ADS Diamond , Jeff . 1998 . “ African-American Attitudes Towards United States Immigration Policy .” International Migration Review 32 : 51 – 70 . Google Scholar CrossRef Search ADS Facchini , Giovanni , and Max Friedrich Steinhardt . 2011 . “ What Drives U.S. Immigration Policy? Evidence from Congressional Roll Call Votes .” Journal of Public Economics 95 ( 7 ): 734 – 43 . Google Scholar CrossRef Search ADS Gaillard , Frye . 2004 . Cradle of Freedom: Alabama and the Movement That Changed America . Tuscaloosa : University of Alabama Press . Gay , Claudine . 2006 . “ Seeing Difference: The Effect of Economic Disparity on Black Attitudes towards Latinos .” American Sociological Review 50 ( 4 ): 982 – 97 . Gonzalez , Jorge , and Nipoli Kamdar . 2000 . “ Do Not Give Me Your Tired, Your Poor! Determinants of Legislator Voting on Immigration Issues .” Eastern Economic Journal 26 ( 2 ): 127 – 43 . Griffin , John , and Brian Newman . 2008 . Minority Report: Evaluating Political Equality in America . Chicago : University of Chicago Press . Google Scholar CrossRef Search ADS Grose , Christian . 2011 . Congress in Black and White . New York : Cambridge University Press . Google Scholar CrossRef Search ADS Guo , Guang , and H. X. Zhao . 2000 . “ Multilevel Modeling for Binary Data .” Annual Review of Sociology 26 : 441 – 62 . Google Scholar CrossRef Search ADS Hainmueller , Jens , and Daniel Hopkins . 2014 . “ Public Attitudes towards Immigration .” Annual Review of Political Science 17 : 225 – 49 . Google Scholar CrossRef Search ADS Hero , Rodney , and Robert Preuhs . 2013 . Black-Latino Relations in US National Politics: Beyond Conflict or Cooperation . New York : Cambridge University Press . Hopkins , Daniel . 2010 . “ Politicized Places: Explaining Where and When Immigrants Provoke Local Opposition .” American Political Science Review 104 ( 1 ): 40 – 60 . Google Scholar CrossRef Search ADS Hutchings , Vincent , and Cara Wong . 2014 . “ Racism, Group Position, and Attitudes about Immigration among Blacks and Whites .” Du Bois Review 11 ( 2 ): 419 – 42 . Google Scholar CrossRef Search ADS Jones-Correa , Michael. 2011 . “Commonalities, Competition, and Linked Fate.” In Just Neighbors? Research on African-American and Latino Relations in the United States , edited by Edward Telles , Mark Q. Sawyer , and Gaspar Rivera-Salgado , 63 – 95 . New York : Russell Sage Foundation . Juenke , Eric Gonzalez , and Robert Preuhs . 2012 . “ Irreplaceable Legislators? Rethinking Minority Representatives in the New Century .” American Journal of Political Science 56 ( 3 ): 705 – 15 . Google Scholar CrossRef Search ADS Kaufmann , Karen . 2003 . “ Cracks in the Rainbow: Group Commonality as a Basis for Latino and African-American Political Coalitions .” Political Research Quarterly 56 ( 2 ): 107 – 26 . Google Scholar CrossRef Search ADS King-Meadows , Tyson , and Thomas Schaller . 2007 . Devolution and Black State Legislators: Challenges and Choices in the Twenty-First Century . New York : SUNY Press . Long , J. Scott , and Jeremy Freese . 2014 . Regression Models for Categorical Dependent Variables Using Stata , 3rd ed . College Station, TX : Stata Press . Lovato , Roberto . 2008 . “Juan Crow in Georgia.” The Nation, May 26. http://www.thenation.com/article/juan-crow-georgia. Marrow , Helen . 2011 . “Intergroup Relations: Reconceptualizing Discrimination and Hierarchy.” In Being Brown in Dixie: Race, Ethnicity, and Latino Immigration in the New South , edited by Cameron Lippard and Charles Gallagher , 53 – 76 . Boulder : First Forum Press . Masuoka , Natalie , and Jane Junn . 2013 . The Politics of Belonging: Race, Public Opinion and Immigration . Chicago : University of Chicago Press . Google Scholar CrossRef Search ADS McAdam , Douglas . 1990 . Freedom Summer . New York : Oxford University Press . McCarty , Nolan , Keith Poole , and Howard Rosenthal . 2016 . Polarized America: The Dance of Ideology and Unequal Riches . Cambridge, MA : MIT Press . McClain , Paula , Victoria DeFrancesco Soto , Monique Lyle , Gerald Lackey , Jeffrey Grynaviski , Kendra Cotton , Shayla Nunnally , Thomas Scotto , and Allan Kendrick . 2008 . “Black Elites and Latino Immigrant Relations in a Southern City: Do Black Elites and the Black Masses Agree?” In New Race Politics in America , edited by Jane Junn and Kerry Haynie , 145 – 65 . Cambridge : Cambridge University Press . Google Scholar CrossRef Search ADS McClain , Paula , Monique Lyle , Efren Perez , Jessica Johnson Carew , Eugene Walton , Candis Watts , Gerald Lackey , Danielle Clealand , and Shayla Nunnally . 2009 . “Black and White Americans and Latino Immigrants: A Preliminary Look at Attitudes in Three Southern Cities.” Presented at the Annual Meeting of the American Political Science Association, Toronto, Canada. McDermott , Monica . 2011 . “Black Attitudes and Hispanic Immigrants in South Carolina.” In Just Neighbors? Research on African-American and Latino Relations in the United States , edited by Edward Telles , Mark Sawyer , and Gaspar Rivera-Salgado , 242 – 66 . New York : Russell Sage Foundation . McKanders , Karla . 2010 . “ Sustaining Tiered Personhood: Jim Crow and Anti-Immigrant Laws .” Harvard Journal of Racial and Ethnic Justice 26 : 163 . Monogan , James . 2013 . “ The Politics of Immigrant Policy in the 50 US States, 2005–2011 .” Journal of Public Policy 33 ( 1 ): 35 – 64 . Google Scholar CrossRef Search ADS Motel , Seth , and Eileen Patten . 2013 . Statistical Portrait of the Foreign-Born Population in the United States, 2011. PEW Hispanic Center. http://www.pewhispanic.org/files/2013/01/PHC-2011-FB-Stat-Profiles.pdf. NAACP . 2014 . “Born Suspect: Stop-and-Frisk Abuses and the Continued Struggle to End Racial Profiling in America.” Report. https://action.naacp.org/page/-/Criminal%20Justice/Born_Suspect_Report_final_web.pdf.. NAACP . 2011 . “NAACP Immigration Fact Sheet and Talking Points.” http://www.naacp.org/wp-content/uploads/2016/04/Immig%20Factsheet%20Tlkng%20Pts%20Final%20July%202011.pdf. Ngai , Mae . 2004 . Impossible Subjects: Illegal Aliens and the Making of Modern America . Princeton, NJ : Princeton University Press . NHCSL . 2010 . “In Response to Draconian Immigration Law, Black and Hispanic State Legislators Pull Joint Conference from Arizona Venue.” Press release. http://www.altoarizona.com/documents/Black_Hispanic_State_Legislators_Pull_Joint_Conference_From_Arizona_Venue.pdf. Nicholson-Crotty , Jill , and Sean Nicholson-Crotty . 2011 . “ Industry Strength and Immigrant Policy in the American States .” Political Research Quarterly 64 ( 3 ): 612 – 24 . Google Scholar CrossRef Search ADS Nteta , Tatishe . 2013 . “ United We Stand? African-Americans, Self-Interest and Immigration Reform .” American Politics Research 41 ( 1 ): 147 – 72 . Google Scholar CrossRef Search ADS Odem , Mary , and Elaine Lacy . 2009 . Latino Immigrants and the Transformation of the US South . Athens : University of Georgia Press . Omi , Michael , and Howard Winant . 1994 . Racial Formation in the United States: From the 1960s to the 1980s , 2nd ed . New York : Routledge and Kegan-Paul . Pager , Devah , and Lincoln Quillian . 2005 . “ Walking the Talk? What Employers Say Versus What They Do .” American Sociological Review 70 : 355 – 80 . Google Scholar CrossRef Search ADS Passel , Jeffrey , and D’Vera Cohn . 2011 . Unauthorized Immigrant Population: National and State Trends, 2010 . Washington, DC : Pew Hispanic Center . Pew Research Center . 2013 . “Mapping the Latino Population, by State, County and City.” http://www.pewhispanic.org/files/2013/08/latino_populations_in_the_states_counties_and_cities_FINAL.pdf. Pew Research Center . 2015 . “U.S. Foreign Born Population Trends.” http://www.pewhispanic.org/2015/09/28/chapter-5-u-s-foreign-born-population-trends/. Pruehs , Robert . 2006 . “ The Conditional Effects of Minority Descriptive Representation: Black Democratic Legislators and Policy Influence in the American States .” Journal of Politics 68 ( 3 ): 585 – 99 . Google Scholar CrossRef Search ADS Ramakrishnan , S. Karthick , and Tom Wong . 2010 . “Partisanship, Not Spanish: Explaining Municipal Ordinances Affecting Undocumented Immigrants. In Taking Local Control: Immigration Policy Activism in US Cities and States , edited by Monica W. Varsanyi , 73 – 93 . Palo Alto, CA : Stanford University Press . Raudenbush , Stephen , and Anthony Bryk . 2002 . Hierarchical Linear Models: Applications and Data Analysis. Thousand Oaks, CA : Sage Publications . Raudenbush , Stephen , Randall Fotiu , and Yuk Fai Cheong . 1999 . “ Synthesizing Results from the Trial State Assessment .” Journal of Educational and Behavioral Statistics . 24 ( 4 ): 413 – 38 . Google Scholar CrossRef Search ADS Shafer , Byron , and Richard Johnston . 2009 . The End of Southern Exceptionalism: Class, Race, and Partisan Change in the Postwar South . Cambridge, MA : Harvard University Press . Shor , Boris , and Nolan McCarty . 2011 . “ The Ideological Mapping of American Legislatures .” American Political Science Review 105 ( 3 ): 530 – 55 . Google Scholar CrossRef Search ADS Tate , Katherine . 2003 . Black Faces in the Mirror: African-Americans and their Representatives in Congress . Princeton, NJ : Princeton University Press . Thurber , James , and Antoine Yoshinaka , eds. 2015 . American Gridlock: The Sources, Character, and Impact of Political Polarization . New York : Cambridge University Press . Google Scholar CrossRef Search ADS Wilkinson , Betina Curaia . 2015 . Partners or Rivals? Charlottesville : University of Virginia Press . Williams , Kim , and Lonnie Hannon III . 2016 . “ Immigrant Rights in a Deep South City: The Effects of Anti-Immigrant Legislation on Black Elite Opinion in Birmingham, AL .” Du Bois Review 13 ( 1 ): 139 – 57 . Google Scholar CrossRef Search ADS Author notes We thank Tim Dowd, Helen Marrow, Joya Misra, Alex Hicks, and Kim Williams for their comments on the paper. We are grateful to Dr. Rob O’Reilly and Shannon McClintock for their expert statistics and data advice. We were fortunate to have an exceptionally skilled group of research assistants who assisted in the data collection and coding: Shilpi Agrawala, Alex Reibman, Joanna Chang, Tiffany Chen, In Young Park, Yordanos Agajyelleh, and Jeffeline Ermilus. We thank the Emory SIRE and RISE Undergraduate Research Programs for supporting these students. © The Author(s) 2018. Published by Oxford University Press on behalf of the University of North Carolina at Chapel Hill. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Social ForcesOxford University Press

Published: Feb 16, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off