Who Really Represents Me? The Case of Afro-Latinx Bureaucratic Representation in New York City Public Schools

As demographic groups ’ heterogeneity increases, questions emerge about how elected and unelected political representatives respond to such diversity. Representative bureaucracy scholarship suggests that representatives will rely on shared values and interests with clients of their demographic group to make decisions or implement policies that improve the group ’ s status. However, differences in immigration histories, demographic characteristics, language, and discrimination experiences within racial and ethnic groups are points of diversion that could affect representation. We explore the relationship between race and ethnicity to understand how within-group differences may disrupt the traditional assumptions of representation. Centering on the experiences of Afro-Latinx students, we ask, What effect do within-group differences have on bureaucrat-client representation? ” Afro-Latinx students share a racial identity with Black education bureaucrats and an ethnic identity with Latinx education bureaucrats but may also differ from both groups in their language acquisition, culture, norms, and interests. We find that Black representatives offer Afro-Latinx students substantive representation, while Latinx representatives do not when we consider their racial identity. The research holds implications for understanding the boundaries of representation and may offer insight into the importance of disaggregating groups in representation studies.


Introduction
The racial and ethnic makeup of the United States is ever-changing, as shifts in immigration patterns and legislation cause different minority populations to grow at different rates.Since 2000, the Latinx population in the United States has grown by over 70 percent and now makes up approximately 19 percent of the total U.S. population (Krogstad, Passel, and Noe-Bustamante 2022).About a third of the Latinx community is foreign-born, though immigrants are a shrinking portion of the Latinx population in the United States (Noe-Bustamante and Flores 2019; Noe-Bustamante 2019).The Black community in the United States is also growing and changing.Over the last several decades, immigration has accounted for an increasing share of this population growth, and we can expect this trend to continue through the next several decades.U.S. Census Bureau projections indicate that by 2060, the Black immigrant population will reach 9.5 million, or 16 percent of the total Black population (Tamir and Anderson 2020).Most Black immigrants originate from Africa or the Caribbean, with a smaller share from Latin American countries such as Guyana, Mexico, and Honduras (Tamir 2021).
Occurring alongside these population changes are changes in racial and ethnic self-identification.In the last 20 years, an increasing number of Black Americans have identified as both Black and Hispanic, or what the Census categorizes as "Black Hispanic" and what is known as "Afro-Latinx" in the race and ethnic politics literature (Tamir 2021).From 2000 to 2019, the number of people identifying as Black Hispanic (henceforth Afro-Latinx) more than doubled to reach 2.4 million.As simultaneously racially Black and ethnically Hispanic, these individuals sit at a complex intersection of identity which creates ambiguity around group membership and representation in electoral politics and other public spaces.Are Afro-Latinx individuals more likely to identify and find representation within the Black community, Latinx community, or neither?
Political science research suggests that descriptive representation is important for Black and Latinx people and can often translate to substantive representation in the legislative sphere.Black and Latinx constituents' perceived policy commonalities within their racial and ethnic groups drive their preference for legislative representation by members of these groups (Casellas, Gillion, and Wallace 2019).For Black and Latinx people, descriptive representation translates to substantive representation along various measures.Compared to white legislators, Black and Latinx legislators spend significantly more time advocating for policies favorable to Black and Latinx individuals (Minta 2009;Lowande, Ritchie, and Lauterbach 2019) and are also more likely to co-sponsor bills of high salience to these communities (Wallace 2014).Outside of the legislature, Black and Latinx bureaucrats also substantively represent Black and Latinx clients, respectively, in the context of social services such as family and child welfare agencies (Watkins-Hayes 2009), education (Nicholson-Crotty, Grissom, and Nicholson-Crotty 2011;Grissom, Kern, and Rodriguez 2015;Lindsay and Hart 2017), policing (Lasley et al. 2011;Headley and Wright 2020), and prisons (Wade-Olson 2019; Johnston and Holt 2021).It is clear that racial and ethnic identity matters for representation, but in the case of Afro-Latinx representation, does one identity take precedence over the other?Does shared ethnicity with other Latinxs translate into effective representation by co-ethnics?Does a shared Black racial identification mean that Black people better represent Afro-Latinxs?
We draw on the intersection of two main streams of literature to help answer these questions: representative bureaucracy and race and ethnic politics.Scholars of representative bureaucracy contend that representation among unelected bureaucrats is often as crucial for outcomes as representation among elected officials because elected institutions offer incomplete representation of minority groups (Long 1952;Van Riper 1958).Majority rule provisions complicate political representation for a diverse population with varied interests, so representation in bureaucracies provides minority groups an opportunity to have their interests, preferences, and demands reflected in policy implementation and outcomes (Meier and Capers 2012;Long 1952).As Afro-Latinx people constitute both a racial and ethnic minority group subject to lower access to political representation, we center our research in the bureaucracy, an institution in which they are more likely to experience greater levels of representation.However, research on representation, both bureaucratic and political, typically relies on a single shared identity (e.g., race or ethnicity or gender) between the representative and the represented.An intersectional ethnic-racial identity complicates the assumptions of representation research (but see Vinapol 2020; Fay et al. 2021;Capers and Smith 2021).
Because representative bureaucracy research has not yet fully addressed questions about ethnic-racial intersectional identities, we turn to the race and ethnic politics literature to shape our expectations for Afro-Latinx bureaucratic representation.Specifically, we look to extant work on the concepts of linked fate and racial group consciousness to help us understand how within-group differences might impact group membership and feelings of commonality based on race and ethnicity.Finally, research at the intersection of public administration and race and ethnic politics informs our understanding of the importance and challenges of disaggregating the Latinx community when assessing the impact of representation on policy outcomes.The group is comprised of a web of shared yet distinct languages, religions, racial identities, immigration experiences, and cultural practices that may also influence representation.
Using the case of teachers representing Afro-Latinx students in New York City schools as a measure of bureaucratic representation, our findings reveal that while ethnic representation is effective for the broader Latin American group, it is less effective for people who identify as Afro-Latinx, and instead, they may gain the most effective representation from Black representatives who share their racial identity.The results hold implications for understanding the salience of race within ethnic groups and strengthen the importance of disaggregating pan-ethnic groups when seeking to meet their unique interests.They also demonstrate how U.S. institutions like the bureaucracy respond to the increased diversity within demographic groups.

Representative Bureaucracy
Studies of representation in bureaucracies posit that when unelected officials hold a shared demographic background with the clients they serve, bureaucracies are more responsive to the public's broad interests and needs (Kingsley 1944).This premise relies on several assumptions: 1) people of the same demographic background share a common history, similar life experiences, values, and norms, and in turn, share political interests and expectations; 2) bureaucrats rely on their values, experiences, norms, and socialization to make decisions; and therefore, 3) bureaucrats engage in actions that best reflect those shared values and interests for clients that share their demographic background.Researchers term such actions active representation.Like substantive representation, active representation occurs when a bureaucrat uses his/ her position to "press for the interests and desires of those whom he/she is presumed to represent" (Mosher 1968, 11).This process is vital for marginalized and minoritized groups because it offers recourse in representation they may not find in elected bodies.Bureaucratic representation also directly affects the group's immediate access to goods and services, perhaps more so than elected representation.
Nevertheless, descriptive representation among bureaucrats that leads to active, substantive gains for minoritized groups is not a guaranteed process.Scholars point to several factors or conditions that strengthen the likelihood of descriptive representation yielding substantive benefits for clients that share the identity of a bureaucrat.In addition to Mosher's (1968) contention that bureaucrats must "press for the interests and desires of those whom he/she is presumed to represent," Meier (1993) and Thompson (1976) also contend that the bureaucrats must hold positions in which they can use their discretion, support, or even mobilization efforts to influence the outcomes of policy in a way that will reflect the interests and desires of those represented.When bureaucrats serve in institutions that can actually affect the day-to-day outcomes of the represented groups and that deal with issues salient to their represented group such as the redistribution of social provisions, their descriptive representation may also more readily lead to substantive benefits for clients that share their identity (Thompson 1976).Last, if bureaucrats view themselves as advocates of marginalized groups' rights or believe that representing one group will not isolate, restrict, or harm another group, that is, their actions do not hold zero-sum consequences, they are more likely to use their position to substantively represent clients of their demographic group (Lim 2006;Selden 1997).
Employees of government agencies first come to mind as arbiters of representation in bureaucracies, but education bureaucrats, or teachers, make up the largest body of bureaucrats nationwide (U.S. Bureau of Labor Statistics 2022) who we can reasonably expect to have great influence on the day-to-day and long-term outcomes of their clients, students (Lipsky 1980).The nature of teachers' role in K-12 education meets Meier (1993) and Thompson's (1976) preconditions for active representation in that it requires high-touch, prolonged contact with their students and involves relatively high levels of discretion and autonomy (Grissom et al. 2015).Students spend considerable time with their teachers, and school settings offer repeated educational and social interactions that influence outcomes.Additionally, teachers of color in particular often view themselves as "advocates of students of color" who understand the students' culture, experiences, and academic challenges and therefore have a heightened investment in their success (Griffin 2018).
When teachers share the racial or ethnic identity of their clients (i.e., students), they have more positive perceptions of the students' performance (Dee 2005), students have higher levels of academic achievement, (Gershenson et al. 2018;Capers and Smith 2021) and fewer absences from school (Holt and Gershenson 2019).Latinx students are more likely to be recommended for gifted education and less likely to experience harsh, punitive disciplinary sanctions when there is greater Latinx representation among school bureaucrats (Meier and Stewart 1991;Meier 1993).Additionally, descriptive representation among Latinx bureaucrats yields positive emotional (symbolic) gains for Latinx clients.Latinx students feel more connected to their school and hold higher educational expectations when there are more Latinx teachers in their school (Atkins, Fertig, and Wilkins 2014).Indeed, minority teachers can and do use their positions to actively serve as representatives for their students and increase their access to opportunities and resources (Atkins, Fertig, and Wilkins 2014;Nicholson-Crotty, Grissom, and Nicholson-Crotty 2011).
Scholars point to similar relationships between bureaucrats and clients outside of education as well.For example, increases in the number of federal bureaucrats of color are associated with more positive loan application decisions in the U.S. Farmers Home Administration (Selden 1997), more EPA enforcement actions in local communities with high levels of race-related social vulnerability and severe environmental inequities (Liang, Park and Zho 2020), more discrimination investigations and charges filed on behalf of Black and Latinx people through the Equal Employment Opportunity Commission (Hindera 1993), and increases in federal contracts to minority-owned firms (Brunjes and Kellough 2018).Law enforcement bureaucracies with more racial and ethnic descriptive representation tend to engage in less intensive immigration enforcement of Latinx immigrants (Chand 2020;Lewis et al. 2013), are viewed as more trustworthy, fairer, and more legitimate among communities of color (Riccucci, Van Ryzin and Jackson 2018), and have officers of color that are less likely to use severe force against Black people (Headley and Wright 2020).While bureaucratic agencies differ from one another in their specific forms and functions, these consistent findings suggest that insights gained in one type of bureaucracy may well be relevant for others.
In all, the representative bureaucracy literature shows that a shared background on a single demographic factor is sufficient for substantive bureaucratic representation.It suggests that a shared identity with one's representative should be beneficial for clients, even if clients vary on other demographic characteristics.For Latinx clients, having a shared pan-ethnicity, that is, a shared "social origin" of common language, culture, and heritage, is assumed to mean that representatives and clients have some common ground-shared experience, values, attitudes, and interests-which the representatives can use to better serve the clients than representatives who do not share one's ethnicity or "social origin" (Taylor et al. 2012;Lopez 2013;Jang et al. 2022).As scholars point to some commonalities among people with a Latinx pan-ethnicity such as the use of Spanish, familiarity with Spanish surnames, and a common religion of Catholicism, and a combined 78 percent of Latinxs agree that Latinxs from different countries have common values, it stands to reason that such similarities and common values may also influence bureaucrats' decision-making when serving clients in which they share a pan-ethnic Latinx identity (Lopez and Espiritu 1990;Jang et al. 2022;Lopez 2013).What's more, pan-ethnicity is often contextually activated, structurally determined, and externally defined through experiences with political and bureaucratic institutions, and Latinx education bureaucrats are not exempt from these influences (Jang et al. 2022).In fact, they may be more inclined to use their position to better serve or represent the interests of Latinx students given their professional understanding of pan-ethnic congruence and the policy or political implications of a shared identity.

Racial /Ethnic Group Identity
While the theory of representative bureaucracy suggests that representation will best occur in a space of homogeneity or shared identities, experiences, and/or values and norms and assumes that congruence on a single identity is sufficient for bureaucratic representation to occur, the theory flattens the salience of clients' multiple identities.It has yet to investigate how representation operates for clients such as Afro-Latinxs who have multiple intersecting identities that hold consequences for access to goods, services, and interactions with American institutions.The race and ethnic politics literature on the political implications of group identity helps us understand what to expect when a Latinx or a non-Latinx Black bureaucrat serves an Afro-Latinx client in which the conditions of representative bureaucracy are not neatly met, and bureaucrats must contend with serving clients of multiple salient identities.Here, group identity and concepts related to group identity such as racial group consciousness and linked fate may explain bureaucrats' decision calculus when representing Afro-Latinx clients.Rogers (2006) defines group identification as "a self-awareness of membership in a group and a psychological sense of attachment to the group" (p.176).Minority group identity then contends that commonalities among a "minority" group can result in shared group identification and consciousness among said minorities (Austin Middleton and Yon 2012).Two concepts derived from the minority group identity thesis, racial group consciousness and linked fate, are frequently used to understand political behavior among racial and ethnic groups and may also explain bureaucratic behavior.
Racial group consciousness is a politicized in-group identification rooted in beliefs about one's racial group used to explain political behavior among minoritized groups (Miller et al. 1981;Sanchez and Vargas 2016).It suggests that 1) when individuals identify as a member of a racial minority group, 2) prefer to interact with members of the said racial group over out-group members, 3) tend to view one's racial group status through a comparative lens of resources relative to other groups, and 4) views systemic discrimination as a key explanatory factor for the racial group's position, they demonstrate some level of racial group consciousness (Miller et al. 1981, 500;Austin, Middleton, and Yon 2012).Researchers also point to shared interests, political ideologies, leadership preferences, and a belief in the benefits of collective action as sources of solidarity among distinct racial minority groups (Masuoka 2006;Austin, Middleton, and Yon 2012).
Experiences of racism and discrimination in a racialized social system also offer a mechanism of shared identity (Austin, Middleton, and Yon 2012;Bonilla-Silva 1997).Discrimination or perceptions of it emerge as the foundation of group consciousness for Black and Latinx people (Sanchez and Vargas 2016); individuals who report experiences of racial discrimination tend to indicate a greater sense of racial group consciousness, which in turn leads them to support policies that affect the racial group as a whole (Capers and Smith 2016).As scholars have found racial group consciousness to be important for understanding group behavior, it is likely equally helpful in understanding bureaucratic behavior.
A second concept of group identity is linked fate.Originally derived to explain uniformity in Black political behavior, linked fate centers on how non-whites operate in concert to resist racism, pursue their political interests, and influence American politics (Dawson 1995, Rogers andKim 2021).It emphasizes the collective agency of nonwhites in a racialized political system that subordinates Black people to the lowest rungs of the American racial hierarchy.The micro-level portion of Dawson's theory contends that among individuals within a group that there is a determinative link between their own well-being and that of the group as a whole, and because of this link, individuals tend to consider the well-being of all members of their racial group when making political decisions, even if such decisions are less beneficial to the individual decision-maker (Allen, Dawson, and Brown 1989;Dawson 1995;Rogers 2006;Sanchez and Vargas 2016).Group-based interests guide political decision-making within demographic groups; that is, it serves as a proxy or information shortcut for simplifying political decisions and should efficiently help Black people determine their own interests as long as race continues to shape their life chances (Dawson 1995, 61).As bureaucrats hold membership in such groups, they may also rely on a sense of linked fate to make decisions in implementing public policies.
Both concepts have been applied to various demographic groups, but scholars reach mixed conclusions on their applicability.While research consistently finds both concepts useful in explaining Black political behavior (Miller et al. 1981;Dawson 1995;Shingles 1981;Uhlaner, Cain and Kiewiet 1989;Tate 2010;Smith 2013;Shaw, Foster, and Combs 2019), scholars are more cautious to conclude their presence among the pan-ethnic Latinx community as research findings are mixed, inconclusive, or vary based on Latinx subgroup (Stokes 2003;Sanchez 2006b;Sanchez and Masuoka 2010;Masuoka 2006).As Junn and Masuoka (2008) point out, predicting group-based political behavior based on a shared racial classification is not simple or obvious for groups with immigration-related population growth (729), and McClain and colleagues ( 2009) caution against applying theories and measures of Black group-based political behavior to other groups on similar grounds.Such differences between Black and Latinx group identity may also appear among Black and Latinx bureaucrats when representing clients of a shared demographic, complicating the representation each offers to Afro-Latinx people who share identities with both demographic groups but also differ in the extent to which they meet the conditions of representative bureaucracy.

Latinx Group Identity, Heterogeneity and Representation
Latinx is a heterogeneous pan-ethnic category whose members may or may not share the traits or experiences that are assumed to foster bureaucratic representation and are known to motivate linked fate and group consciousness.Subgroups of Latinxs often have diverging immigration histories, different demographic characteristics, language differences, and experiences with discrimination, which make them more heterogeneous than commonly assumed (Umaña- Taylor et al. 2002).Such heterogeneity complicates ideas of homogeneity for bureaucratic representation and ideas of racial or ethnic solidarity rooted in a sense of group consciousness or feelings of a linked fate that could also facilitate bureaucratic representation to the benefit Afro-Latinx people.Consequently, Latinx bureaucrats may not substantively represent Afro-Latinx clients.The research on Latinx group identity and behavior offers some insight on the barriers to substantive representation for Afro-Latinx clients.
As previously noted, researchers reach mixed conclusions on racial group consciousness and linked fate among Latinxs (Stokes 2003;Sanchez 2006b;Sanchez and Masuoka 2010;Masuoka 2006).Research that supports a level of ethnic (racial) group consciousness and feelings of linked fate among Latinx people suggests a panethnic identity is likely salient for Latinxs, and therefore, Latinx bureaucrats may also lean on this group ethnic identity to make decisions regarding representation and/or the distribution of goods and services.On the other hand, scholars point to a more complex group identity and fail to find evidence of coherent group consciousness or feelings of linked fate (Sanchez and Masuoka 2010;Shaw, Foster, and Combs 2019).Because experiences with discrimination, a common driver of linked fate among African Americans, vary among Latinx subgroups, it does not meaningfully contribute to their sense of linked fate and stands in the way of group solidarity (Sanchez and Masuoka 2010).Economic marginalization and proximity to immigrant experiences appear to influence ethnic group consciousness among Latinx people, but again, these experiences vary among the group and Afro-Latinx members tend to find themselves on the lowest end of the economic spectrum.
Beyond external experiences of discrimination, discrimination within the Latinx group also serves as a barrier to group solidarity and likely minimizes ethnic representation for Afro-Latinx clients (Monforti and Sanchez 2010;Carey et al. 2013;Mallet and Pinto-Coelho 2018).Internal discrimination occurs when members of a minority group experience discrimination from other members of that group, or "co-ethnics."Internal division and discrimination among Latinxs may occur for several reasons, including conflict based on nativity, language differences, or phenotype.Monforti and Sanchez (2010) found that perceived internal discrimination varied by national origin, with Dominicans being the most likely to see internal discrimination as a "big problem" and Puerto Ricans being the least likely.Latinx immigrants with limited English language proficiency also report higher levels of internal discrimination by third or fourth-generation immigrants who are English-dominant (Lavariega Monforti and Sanchez 2010).However, language-based discrimination also occurs in the opposite direction with Spanishdominant Latinxs negatively viewing Latinxs with limited Spanish language proficiency (Howard 2018).For example, Haywood (2017) finds that Spanishdominant peers "looked down upon" and made fun of English-dominant Afro-Latinxs (p.773).In sum, experiences with discrimination-both external and internal to the Latinx group-may also hamper bureaucratic representation for Afro-Latinxs in which Latinx bureaucrats may fail to substantively represent them.
Phenotypical and skin color differences also stand out as barriers to group identity, a weakened sense of racial group consciousness and linked fate, and likely influence how Latinx bureaucrats represent Afro-Latinx clients (Sanchez and Masuoka 2010).As a group, Latinxs are commonly referred to as "Brown," to homogenize the group and distinguish it from White and Black on the White-Black racial paradigm, but historical racial mixing has led Latinxs to hold a broad range of physical characteristics in skin color and phenotype (Telles 2018;López 2008).Often one's actual skin color holds implications for in-group acceptance, one's sense of linked fate, experiences, and social and economic outcomes and privileges (Clealand and Gutierrez 2022;Chavez-Dueñas, Adames and Organista 2014).Latinx people with darker skin experience higher levels of workplace discrimination, lower occupational prestige, lower incomes, less educational attainment, and poorer health outcomes (Espino and Franz 2002;Noe-Bustamante et al. 2021;Arce et al. 1987; Golash-Boza and Darity 2008; Aja 2016).They are more likely to be negatively stereotyped or exoticized in media depictions, and associated with "laziness, backwardness, lethargy, and neglect," (Soler Castillo & Pardo Abril, 2009, p. 135;Chavez-Dueñas, Adames and Organista 2014).Consequently, darker-skinned Latinos are more likely to report having a linked fate to Black people and feel less close to other Latinos relative to Latinos who identify as White (Clealand and Gutierrez 2022).
Latinxs' racial identity, often shaped by one's skin color, also serves as a barrier to a sense of a Latinx-linked fate, group solidarity, and opportunities for bureaucratic representation that can substantively benefit Afro-Latinx people.Race is such a focal point in the United States that it may "overwhelm and foreclose : : : other options for group identification" and serves as perhaps the most significant barrier to a shared group identity (Rogers 2006, 171).Latinx immigrants quickly recognize the role of race and anti-Black discrimination in America, leading some to intentionally distance themselves from the African American community and identity in an effort to shore up their status in their new homeland (Rogers 2006).When queried on their racial identity, Latinx people are least likely to self-identify as Black and over half identify as White (Noe-Bustamante et al. 2021.)It may come as no surprise that when asked to identify their racial identity using a color scale, only three percent of Latinx individuals used one of the four darkest skin colors to self-identify their racial identity, while 80 percent relied on the one of the four lightest shades to identify (Noe-Bustamante et al. 2021) 1 .Lighter-skinned Latinx people frequently identify more strongly with White people, Latinx people with medium or brown skin tones tend to identify with a mixed race category, and darker-skinned Latinxs feel more commonality with Black people and identify with a Black or mixed race category (Clealand and Gutierrez 2022;Noe-Bustamante et al. 2021;Howard 2018).Because racial group consciousness and linked fate rely on one to first identify as a member of a racial/ethnic minority group, such behaviors limit a sense of an ethnic group identity.They also contribute to internal discrimination and intra-group conflict as Latinxs who identify as Black are much more likely to report race as a leading cause of conflict among co-ethnics (Mallet and Pinto-Coelho 2018).The findings are consistent with previous research finding anti-Black prejudice among Latinx groups (McClain et al. 2006).
In total, the barriers to an ethnic group identity may restrict ethnic representation for Afro-Latinx people and limit the extent to which Latinx teachers represent Afro-Latinx students.Representative bureaucracy rests on a notion of pan-ethnicity that obscures the diverse experiences and values of Latinx people in the United States, but research on linked fate and racial group consciousness among panethnic Latinxs provides some clarity on the factors that impede cohesion among Latinx communities.Feelings of linked fate among Latinxs are closely tied to socioeconomic and immigration status and vary by policy area instead of racial solidarity as the concept original derives (Sanchez 2006a;Sanchez and Masuoka 2010;Sanchez and Madeiros 2016).Additionally, racial distinctions in skin color and identity weaken group identity and solidarity.Skin color and racial identity often place darker-skinned and Afro-Latinxs at a disadvantage to their fairer-skinned co-ethnic peers; darkerskinned Latinxs and Afro-Latinxs experience anti-Black discrimination within their ethnic group and therefore identify more closely with African Americans and other Black people in the United States.Such intra-group conflicts, internal discrimination, and exclusionary practices may diminish the degree to which one Latinx person represents another, even though they share an ethnic identity (Howard 2018;Haywood 2017;Lavariega Monforti and Sanchez 2010;Clealand and Gutierrez 2022).Internal discrimination increases feelings of intra-group competition and minimizes group closeness and feelings of linked fate, which may decrease Latinx support for pan-ethnic goals and break down social cohesion (Carey et al. 2013;Mallet and Pinto-Coelho 2018).
Additionally, Black Latinxs or those with darker skin's links to Black Americans can further alienate them from fair skinned, or White-identifying Latinx people, so while Afro-Latinx clients share an ethnic identity with White, "Brown," and nonracially identified Latinx bureaucrats and potentially nationality or language, they do not share a racial identity with them, they tend to have different life experiences and norms given their racial identity, and this may in fact lead other Latinx bureaucrats to also view Afro-Latinxs as non-members or exclude them from the public benefits of bureaucratic representation.In sum, varied experiences of discrimination, internal group conflict, anti-Blackness, and racial identity complicate the assumptions of representative bureaucracy that a shared common history, similar life experiences, values, and norms will translate into shared political interests, expectations and favorable bureaucratic actions.

Black Consciousness, Linked Fate, and Representation
Although racial group consciousness and linked fate are more nuanced and fleeting among Latinxs, they are well documented as powerful mechanisms of Black political behavior and may serve as mechanisms of bureaucratic representation as well.Because race so powerfully shapes African Americans' experience, their group identity goes beyond simple affinity to become an important political heuristic (Rogers 2006).Survey research shows African Americans tend to believe that "what happens to Black people in the United States will have something to do with what happens in your life," indicating a sense of linked fate to the racial group (Capers and Smith 2016).This sense of a linked fate is related to African Americans' consciousness of their group's position in society and perceptions of discrimination (Sanchez and Vargas 2016), and scholars link higher levels of racial group consciousness to greater political participation among Black people (Miller et al. 1981;Shingles 1981;Uhlaner, Cain, and Kiewiet 1989).Studies show group consciousness among African Americans leads them to support policies perceived to benefit the racial group or address racial inequality, to oppose policies that have been explicitly racialized, such as police brutality, and to rally around immigrant causes that affect the members of the racial group (Smith 2014, Kim 2000, Candelario 2007;Tate 2010;Lopez Bunyasi and Smith 2019).
Researchers also point to shared interests, political ideologies, leadership preferences, and a belief in the benefits of collective action as sources of solidarity among distinct racial minority groups (Masuoka 2006;Austin, Middleton, and Yon 2012).Experiences of racism and discrimination in a racialized social system also offer a mechanism of shared identity (Austin, Middleton, and Yon 2012;Bonilla-Silva 1997).Individuals who report experiences of racial discrimination tend to indicate a greater sense of racial group consciousness, which in turn leads them to support policies that affect the racial group as a whole (Capers and Smith 2016).
African Americans' group consciousness also leads them to express racial solidarity with Black people of differing nationalities and ethnicities.Austin et al.'s (2012) research shows that African Americans have a shared racial group consciousness with Black immigrants, including Afro-Cubans.Similarly, Smith (2013) finds that African Americans and Black immigrants have relatively similar levels of group consciousness and linked fate to other Blacks and such group consciousness leads them to support racialized policies such as reparations that can benefit the broader racial group.While Nunnally (2010) concludes that African Americans feel less linked to Caribbean and African people in the United States, she also finds that more frequent experiences of racial discrimination and living in the northern region of the United States where one likely has more cross-ethnic interactions enhances feelings of a linked fate to Black people of varying nationalities.These findings suggest that African Americans are generally more amenable to accepting ethnic differences when race remains a salient identity.
It stands to reason that this sense of group consciousness and the idea that their individual well-being is contingent on the well-being of other Black people may also appear in the actions of Black bureaucrats.As such, African American teachers' racial group consciousness, sense of a linked fate, and shared political ideology, interests, and racial experiences may lead them to overlook their ethnic difference from Afro-Latinx students and the unmet conditions of representative bureaucracy to represent Afro-Latinx students' interests.

Expectations
The broad literature on representative bureaucracy and racial/ethnic identity helps us to develop two expectations of from whom Afro-Latinx clients, or students, may experience bureaucratic representation.As a baseline, the representative bureaucracy literature leads us to contend that Latinx bureaucratic representatives, or teachers, will better represent Latinx clients (students) than non-Latinx representatives; having a shared pan-ethnic identity will yield positive policy outcomes because a shared background on a single demographic factor is sufficient for substantive bureaucratic representation (H1).Here we expect all Latinx students, including Afro-Latinx students, to experience bureaucratic representation through their Latinx teacher as one's pan-ethnic identity is assessed.
However, Latinx teachers are a part of a diverse, pan-ethnic group that varies in experiences, values, interests, and expectations, and therefore, representation based on pan-ethnicity may be more difficult than representation bureaucracy research suggests.Research on linked fate and racial group consciousness among pan-ethnic Latinxs acknowledges such complications and leads us to consider that Latinx bureaucratic representatives, or teachers, may not represent Afro-Latinx clients better than non-Latinx representatives; having a shared ethnic identity will yield null or negative policy outcomes for Afro-Latinx clients (H2).We expect Latinx teachers to have a negative or empirically insignificant relationship with Afro-Latinx students when their racial identity is assessed.
Nevertheless, Afro-Latinx clients are not without some recourse because linked fate and racial group consciousness research on African Americans in the United States suggests that African Americans' sense of group consciousness and notion that their well-being is linked to the broader racial group's well-being extends to ethnic minorities within the racial group.We predict that having a shared racial identity with African Americans will yield positive policy outcomes for Afro-Latinx; Black bureaucratic representatives, teachers, will represent Afro-Latinx clients (H3).Assessing race will benefit Afro-Latinx students when examining Black teachers' representation.

Data
We use the case of public education to assess our hypotheses.Public education is a large bureaucracy in which teachers are the street-level, unelected policy implementers and students are the clients (Pitts 2007).The public education system is one of the most common settings in which scholars test aspects of representative bureaucracy theory (Bishu and Kennedy 2020;Keiser et al. 2002;Meier & Stewart, 1991;Meier, Wrinkle, & Polinard, 1999;Nicholson-Crotty et al. 2011;Pitts 2007;Roch & Pitts 2012).It is an ideal setting for testing representative bureaucracy theory because of the direct and prolonged interactions between clients (students) and bureaucrats (teachers).It is also a venue in which the conditions of descriptive representation can lead to favorable outcomes for same-identity clients.Although state agencies and school districts govern basic curricula requirements, education bureaucrats have high levels of discretion over their task and service environment, and they hold influence over social and/or political issues salient to the group (e.g., classroom curriculum implementation and content, student achievement).They determine how they will implement policies, their teaching method, and classroom management strategies (Roch, Pitts, and Navarro 2010).Bureaucrats of color tend to serve broadly in schools districts, holding positions at every level of the institution, from teachers to superintendents, which enhances the opportunities of representation for marginalized groups.Because education bureaucrats tend to have lower levels of formal socialization within their institution or schools compared to other street-level bureaucrats such law enforcement officers, they are more likely to bring their values and sense of group identity into their work.These factors make it an appropriate venue to test our theoretical expectations as well.
We use restricted, individual-level data from New York City Public Schools (NYC) to assess our research questions.The NYC school system is the largest and one of the most racially and ethnically diverse school systems in the U.S. It educates nearly 200,000 immigrants annually (New York City Independent Budget Office 2016).Its database of student performance includes the statewide standardized testing results of third through eighth-grade students in English/Language Arts (ELA), math, and science that links to student demographic data.To assess our hypotheses, we combine student demographic, teacher demographic, and student performance data from 2006-2007 to the 2015-2016 school terms.We limit our analysis to the 3.6 million foreign-born and Puerto Rican students in the third through eighth grade who are linked to at least one teacher (n = 3,623,143). 2Fortyfive percent of foreign-born students identify a Latin American country as their place of birth, and among those students, twelve percent identify as Black.Table 1 includes descriptive statistics for all key variables.

Dependent Variable
We use foreign-born and Puerto Rican students' raw scores on the annual NY state assessment in ELA to assess our hypotheses (New York City Department of Education 2018). 3Only students who have been in the United States for over one year take the ELA test in their grade (New York City Department of Education 2018).While standardized assessments do not capture students' entire learning experience and much controversy surrounds their use, they remain a standard metric for assessing the extent to which students acquire basic academic skills.They continue to be used in teachers' performance appraisals and are important to various education stakeholders (McNeil 2000;Meier and O'Toole 2001).Consequently, teachers (and schools) have an extra incentive to perform well, and scholars continue to use it as a suitable outcome measure.
We use an indicator of a student's country of birth to designate a student as foreign-born and identify those born in Latin American 4 countries who identify as Black as Afro-Latinx (1= Afro-Latinx).We also include select Caribbean nations 5 as a part of Latin America based on location, demographics, language, and/or historical connection to chattel slavery.

Independent Variable
The main independent variable is the race of a student's English Language Arts teacher.In models assessing the effect of pan-ethnic representation and ethnic representation we code Latinx teachers as "1" and non-Latinx teachers as "0."We also rely on a binary variable to assess racial representation, 1= Black teacher.Approximately 18 percent of the teachers in the sample are Latinx; their race and place of birth are unknown.Black teachers make up about 17 percent of the sample.Non-Black and non-Latinx teachers make up roughly 80 percent of the sample.

Controls
Fully specified models include student, teacher, and school level variables related to student performance in the extant education research.We account for a student's socioeconomic status (1= under poverty level) as education scholars frequently find that students with more resources, greater access to educational materials, and opportunities outside of school tend to perform better (Jencks and Phillips 1998).Students who do not identify English as their native language or the language most often spoken at home also struggle to perform well on standardized assessments (Abedi et al. 2003), so we control for a student's English language status (ELL=1).We control for student gender (female=1) as well because scholars have found significant gender differences in reading achievement and writing ability-two skills frequently assessed in ELA assessments (Reilly, Neumann and Andrews 2019;Reynolds et al. 2015).
We account for teacher characteristics that may influence performance.Education literature suggests that teaching experience can improve students' performance, particularly African American students' performance (Kukla-Acevedo 2009; Rockoff 2004), so we account for teachers' years of experience at their current school as well as years of employment in NYC public schools.Teacher salary also serves as a measure of teacher experience, quality, and school resources (Hedges and Greenwald 1996;Evans, Murray, and Schwab 1997;Houtenville and Conway 2008).Finally, we control for school-level factors that may also affect performance, including school racial and SES demographics and achievement profile.We incorporate the proportion of Latinx and Black teachers in a school to account for the effect of representation at the organizational level, consistent with much of the representative bureaucracy research.We control for the proportion of Black, Latinx, and low-income students (Jencks and Phillips 1998).We standardize the average performance score on the ELA assessment to control for peer performance effects (Hanushek et al. 2003).Finally, we include a measure of student-teacher ratio as a proxy for school resources because schools with greater human and financial resources are associated with higher student performance (Evans, Murray, and Schwab 1997;Greenwald, Hedges, and Laine 1996;Hedges and Greenwald 1996;Houtenville and Conway 2008).

Methodology
Given the panel nature of the data, we estimate a series of fixed effects models that predict a student's performance based on the ethnic or racial congruence one has with the classroom teacher to better understand bureaucratic representation for Afro-Latinx clients.Fixed effect modeling allows us to account for the effect of timefixed, unobserved, measurable student factors that may be correlated with the independent variables or the outcome measure and bias our findings.This is particularly important for our study because assessment scores, our outcome measure, can be fraught with biases that may disrupt empirical findings.Standardized assessments tend to unfairly disadvantage students of color, lowincome students, rural students, and girls when assessment test items include content that fails to consider their experiences or access and in turn negatively affects their performance.Assessments that use language or viewpoints steeped in stereotypes or that include insensitive content can also influence student performance (Popham 2006).Similarly, teachers' backgrounds may also alter the empirical results.In general, teachers hold slight anti-Black implicit biases that also correlate with student performance (Jacoby-Senghor, Sinclair, and Shelton 2016; Chin et al. 2020).Finally, students may vary in their own unobserved inputs such as motivation or parental support that we are unable to measure with our data.Fixed effects modeling helps us move closer to addressing these potential sources of bias as well as minimizing the possibility of omitted variable bias. 6Fixed effects modeling also better serves our theoretical argument as we seek to isolate the effect of representation on Afro-Latinx clients; that is, we want to know the outcome of a student when the student experiences a change in teacher race.We conduct a Hausman test to determine if the unique errors and the regressors are correlated; however, our results lead us to reject this premise and support the appropriateness of fixed effect modeling for our tests.Fully specified models include year and student fixed effects.
We use robust standard errors to address issues of heteroskedasticity in which the error term variance is not constant, creating inefficient and underestimated coefficients.Robust standard errors allow for unequal variances as it does not assume constant variance in the error term.Instead, it uses the regression's residuals to estimate the variance and therefore offers more reliable coefficient estimates and confidence intervals.While robust standard errors are not the only method to address heteroskedasticity, it is a more flexible approach that does not require additional assumptions.Equation (1) shows the general form of the fixed effect models: where Y ijst is the assessment score of student i, assigned to teacher j, at school s, during year t.X jkst is a vector of teacher characteristics, including years of experience, TchrRace ijst , is an indicator for the race or ethnicity of the classroom teacher (Latinx or Black based on model), and Z st is a vector of school characteristics, including the racial demographics, average performance, and student-teacher ratio of the school.φ t is a school year fixed effect; γ i is a student fixed effect, and µ ijst is the error term.

Results
We seek to assess the effect of within-group differences on bureaucratic representation and to understand the role of such differences in service delivery for Afro-Latinx people.We test the assertions of representative bureaucracy, group heterogeneity, and racial group consciousness.Model 1 of Table 2 shows the effect of holding a shared pan-ethnic identity with a bureaucrat.Recall that the research on descriptive representation and representative bureaucracy suggests that having a shared pan-ethnic identity with a bureaucrat should bode well for Latin Americanborn clients, including those who may have different racial identities because bureaucrats are likely to rely on the shared identity and its correlates (i.e.shared values, experiences, interests) to make decisions and serve clients.Using the relationship between teachers (bureaucrats) and students (clients) as our case, we find that when the average Latin American-born student has a Latinx teacher, the student performs better on the ELA assessment.The Latin American-born student experiences a .61percent increase on the assessment than when the student does not have a Latinx teacher.The findings support our first hypothesis and align with the arguments of descriptive representation and representative bureaucracy.At a baseline, bureaucratic representation yields substantive benefits for clients, particularly when one's pan-ethnic identity is shared.
While our focus is not on student performance, it is noteworthy to acknowledge the other factors that predict students' assessment scores.Latin American-born students who attend schools with more experienced teachers and resources perform better on the assessment, though as teachers gain more experience in the New York City school district, they appear to be less effective in improving student achievement on the ELA assessment.Students who are English Language Learners, those who have performed poorly on previous assessments, and those identified as below the poverty level appear to perform lower on the ELA assessment.Students who attend schools with larger Latinx teacher populations, larger Black and Latinx student populations, lower average test scores, and higher student-teacher ratios Second, we test the group heterogeneity hypothesis that racial differences within the ethnic group may influence representation and limit its substantive benefits.Model 2 of Table 2 shows that Afro-Latinx students do not differ statistically in their performance on the ELA assessment under the tutelage of a Latinx teacher than when they do not have a Latinx teacher.Having ethnic representation does not appear to benefit Afro-Latinx clients substantively, supporting our second hypothesis.Racial differences appear to complicate bureaucratic representation, and in turn, clients who may not share the racial background of the bureaucrat do not experience the substantive benefits of a shared ethnic background.This appears to be especially true when a client's racial identity is emphasized.The control variable findings are consistent with those in Model 1, except for the proportion of Latinx teachers and a student's SES status.In Model 2, both variables are insignificant and do not appear to affect Afro-Latinx students' performance on the ELA assessment.
Finally, Model 3 of Table 2 offers a test of racial representation in which a shared racial identity will influence representation and yield substantive benefits for racial group members.Our model supports the assertion; Afro-Latinx students perform 0.79 percent higher on the ELA assessment under the instruction of a Black teacher than a non-Black teacher.Having racial representation appears to benefit Afro-Latinx clients; Black bureaucrats provide Afro-Latinx clients substantive representation.Control variables remain consistent with Models 1 and 2, but we also find unexpectedly that increases in the proportion of Black teachers decrease Afro-Latinx students' ELA assessment scores.

Discussion
As the demographics of America shift, the politics of America may shift as well.This study examines bureaucratic representation for Afro-Latinxs to understand the implications of diversity within racial and ethnic groups.As previously noted, Afro-Latinx people often identify racially with the Black community but ethnically with the Latinx community.Consequently, we ask, are Afro-Latinxs more likely to gain representation through Latinx representatives, Black representatives, both, or neither?The research literature on bureaucratic representation and racial and ethnic group identity offers varying lenses to understand Afro-Latinx representation and develop our expectations.We test the three perspectives in turn to examine the complexities of bureaucratic representation for Afro-Latinx clients.
Our findings reveal the answer to our question, "who represents Afro-Latinxs?"We find that Afro-Latinxs experience substantive representation with Black bureaucrats, whereas Latinx bureaucrats do not appear to provide substantive representation to Afro-Latinxs when their racial identity is highlighted.Latinx representation only appears significant when the model does not consider racial identity.The findings are consistent with Capers and Smith (2021): "similar experiences within a racialized social system serve to overcome differences in social origins," (p.717).Although Afro-Latinxs and Black representatives may differ in nationality, language, immigration status, culture, and ways of knowing, Black bureaucrats appear to overlook such differences to represent members of their racial group.
On the other hand, this racial difference appears to be a salient barrier to ethnic representation for Afro-Latinx people.Our finding that Latinx bureaucrats do not statistically differ from non-Latinx bureaucrats in their representation of Afro-Latinx clients does not align with most of the previous research on Latinx bureaucratic representation.This is not to suggest that Latinx bureaucrats will not or cannot represent Afro-Latinx clients; however, the findings suggest that racial identities complicate ethnic representation, and within-group racial versus ethnic differences may pose a problem for consensus building in pan-ethnic groups and service delivery.Race remains both a unifying and dividing factor in political representation, and our results indicate the substantive implications of race's effect in politics.For pan-ethnic representation among Latinx-identifying individuals to occur, the group must overcome racial differences and its antecedents.Being members of a shared pan-ethnic group is not enough to elicit substantive representation as representative bureaucracy research traditionally suggests.Our findings reveal that bureaucrats likely consider more than just a shared racial or ethnic identity in their decision-making, particularly when they hold identities that they deem equally or more salient than a shared ethnic identity.As Latino Politics scholars note, neither racial/ethnic identity nor experiences of discrimination are leading factors of group identity for Latinx people (Sanchez and Masuoka 2010;Sanchez and Madeiros 2016), and our findings suggest that they may not determine Latinx bureaucrats' behavior as well.Future research on Latinx bureaucratic representation should begin thinking beyond a pan-ethnic Latinx identity to challenge the assumptions of representative bureaucracy and more thoroughly assess the mechanisms that drive bureaucratic representation.Group heterogeneity will become increasingly important for scholars to consider in representation studies as all racial groups continue to diversify.
Our research brings Afro-Latinx representation to the forefront of representation scholarship, but it is not without limitations that scholars must consider as they develop future studies in this area.The race and ethnic politics literature serves as the basis of our empirical studies, but our data does not allow us to identify the racial background of Latinx bureaucrats to the extent that we are able to identify the racial and ethnic background of foreign-born Latinx clients.As such, we can only assess the race or ethnicity of teachers, and we cannot identify the likely "ideal" representative of Afro-Latinx students, an Afro-Latinx teacher.Nevertheless, our research allows us to highlight the importance of racial diversity within the Latinx community and the effect that race has on representation.We hope that future researchers (and data providers) seek ways to identify the diversity within racial and ethnic groups better.
Relatedly, our inability to discern more demographic information about the teachers (e.g., racial identity, nationality, language) prevents us from more appropriately disentangling the mechanisms of representation, group heterogeneity, and linked fate.We are unable to directly test the level of group consciousness or linked fate among Black bureaucrats, for example.We are also unable to query Latinx bureaucrats on the salience of their racial identity, ethnic identity, or immigration status in their decision making because we rely on restricted, administrative data.Instead, we are left to make assumptions about the potential relationships we observe.In addition to better data collection, future research may also consider shifting to qualitative studies of representation and group consciousness (see Althaus and O'Faircheallaigh 2022) or direct query or observation of representative behaviors and attitudes to discern the factors shaping our empirical findings better (see Kiang, Wilkinson, and Juang 2021).Examining more directly if Black bureaucrats do in fact overlook ethnic differences to represent the interests of members of their racial group is a promising direction for future research.Our research provides future research with a starting point to explore the complex puzzle of racial and ethnic identity in representation in bureaucratic agencies and a window of insight into the potential challenges of an increasingly racially diverse nation.
Supplementary material.To view supplementary material for this article, please visit https://doi.org/10.1017/rep.2024.6 Competing interests.The authors declare none.

Notes
1 Note that Census Bureau identifies only 1.9% of the Latino population as Black; Pew Research Center identifies 2% of Latinos as self-identifying as Black, but 24% consider themselves Afro-Latino (Lopez, Gonzalez-Barrera, and Arditi 2021).The disparities are likely due to respondents' interpretation of the categories; "Black or African American" is often interpreted as a selected category for Black people born in the U.S. or with ancestry in the U.S. instead of those with ancestry abroad or Latin American norms that discourage identifying with the Black category applied to U.S. surveys (see Cruz-Janzen 2007;Mitchell 2018). 2 The NYCDOE administrative data only allows students to identify under a single race category, so we are unable to account for students born in the U.S. who identify as Latinx and Black or Afro-Latinx.Another 27,614 students are excluded from the analysis due to missing data on relevant variables.3 Students with disabilities may take alternate versions of the assessments if it is written into the student's Individualized Education Plan (IEP) (NYC Department of Education 2018a; NYC Department of Education 2018b).4 Latin American countries include those in South America, Central America, and Mexico (UN-DESA 2021).5 Countries often identified as both Caribbean and Latin American include: Belize, Cuba, Dominican Republic, Guyana, Haiti, and Surinam.Respondents from these countries are also included in the study.6 Fixed effect modeling restricts the empirical analysis to only include students who experience a change in teacher representation; students who consistently have a same race teacher over time do not contribute to the estimation and are therefore excluded.

Table 2 .
The substantive effect of descriptive representation for Latin American-born students