Race, representation, and the case of Asian Americans

Research in political science has focused extensively on the descriptive representation of marginalized groups, considering why and how it matters. In defining work on the topic, Pitkin (Reference Pitkin1972) characterized descriptive representatives as political officeholders who “look like their constituents.” Since Pitkin, scholars have highlighted the importance of descriptive representation for a variety of outcomes such as overcoming group marginalization in the policy-making process (Mansbridge Reference Mansbridge1999; Phillips Reference Phillips1998; Sapiro Reference Sapiro1981). While descriptive representation does not always lead to influence in politics and policy, there is a consensus that, at a minimum, the election of political officials from marginalized groups is necessary for the full consideration of group perspectives (Dovi Reference Dovi2002; Mansbridge Reference Mansbridge1999; Williams˙Reference Williams2000). In practice, descriptive representation often leads to substantive representation of group interests (Cameron, Epstein, and O’Halloran Reference Cameron, Epstein and O’Halloran1996; Grose Reference Grose2011; Hero and Tolbert Reference Hero and Tolbert1995; Lowande, Ritchie, and Lauterbach Reference Lowande, Ritchie and Lauterbach2019).

How do minority politicians get elected? The most prominent explanation argues that voters with shared racial or ethnic identity with a candidate are more likely to support the candidate. By this logic, representation of minority groups depends on their share of the population and whether they form a critical mass of voters in particular constituencies, if not the polity at large. Consistent with this expectation, studies of descriptive representation find that constituency racial demographics are a strong predictor of minority representation. For example, the likelihood of electing an African American or Latino legislator increases with the share of an electoral district’s population from that minority group (Canon Reference Canon1999; Casellas Reference Casellas2010; Grofman and Handley Reference Grofman and Handley1989; Hardy-Fanta et al. Reference Hardy-Fanta, Lien, Pinderhughes and Sierra2016; Lublin Reference Lublin1997; Lublin et al. Reference Lublin, Brunell, Grofman and Handley2009; Parker Reference Parker1990; Preuhs and Juenke Reference Preuhs and Juenke2011). In fact, minority legislators are particularly likely to be elected in districts where most residents (or at least a plurality) are from the same racial group as the candidate (Cameron, Epstein, and O’Halloran Reference Cameron, Epstein and O’Halloran1996; Hicks et al. Reference Hicks, Klarner, McKee and Smith2018; Lublin et al. Reference Lublin, Handley, Brunell and Grofman2020; Preuhs and Juenke Reference Preuhs and Juenke2011). A growing literature also considers “supply-side explanations” for the emergence of women and minority candidates, finding that alongside institutional factors, district demographics are strong predictors of whether non-white candidates appear on the ballot (Branton Reference Branton2009; Canon Reference Canon1999; Juenke Reference Juenke2014; Shah Reference Shah2014; Shah, Scott, and Juenke Reference Shah, Scott and Juenke2019).Footnote ³

Scholarship focusing specifically on the case of Asian American officeholders has produced mixed findings, suggesting that theories of minority population size may not fully explain their representation in political office. In contrast to work studying African American and Latino representation, Lai et al. (Reference Lai, Cho, Kim and Takeda2001) find that congressional districts with the largest Asian American populations in the late 1990s and early 2000s were not represented by Asian Americans. In turn, they argue that Asian American candidates are most successful when they have “cross-over appeal” to members of other racial groups. Emergent research (Lublin and Wright Reference Lublin and Wright2023) has revisited this question in a more recent time period (2011–2020), using data on Asian American representation in state legislatures as well as Congress. The authors find that Asian American candidates have cross-over appeal, drawing more support from non-Asian racial minorities than white candidates do, but uncover evidence of co-ethnic voting by Asians as well. In a similar vein, Lien and Filler (Reference Lien and Filler2022) find that Asian American representation on California city councils grows with their share of the population, albeit neither consistently nor uniformly, using data from several points in time between 1980 and 2020 (p. 99).Footnote ⁴

Other studies suggest that the link between Asian population and representation might be a conditional one. For example, in an analysis of the factors that shape state legislative representation for Asian Americans and Latinos, Dyogi Phillips (Reference Phillips2021) argues that a complex interaction of race and gender “simultaneously constrains and facilitates access to electoral opportunity for distinct groups of men and women” (p. 9). She finds that majority–minority districts benefit Asian American and Latino male candidates, but not their female counterparts. Lien and Filler (Reference Lien and Filler2022) also find gendered differences in Asian American representation across levels of political office. In another study focused on the emergence of minority candidates for state legislative positions, Fraga, Juenke, and Shah (Reference Fraga, Juenke and Shah2020) find that Asian American candidates are more likely to run as the percentage of Asian Americans in the district increases, but this association weakens in districts represented by an Asian American incumbent in Congress.Footnote ⁵ Others find similar effects at the local level in Orange County, California, particularly when Asian American candidates employ co-ethnic cues (Uhlaner and Le Reference Uhlaner and Le2017).

Related research has studied Asian American co-ethnic representation from the perspective of voters as opposed to office-holders or candidates. For example, studies of California voters during the late 2010s (Sadhwani Reference Sadhwani2022a, Reference Sadhwani2022b) find that Asian Americans are more likely to vote for Asian American than non-Asian candidates for Congress and the state assembly, but show that shared national origin is a strong predictor of Asian Americans’ vote choices. These studies support theories of co-ethnic voting as well as the idea that pan-ethnicity is not the primary axis of group identity for most Asian Americans (see also Leung Reference Leung2022).

Immigrant assimilation and political incorporation

While much of the literature on minority representation focuses on co-ethnicity, a separate body of work (much of it by sociologists) considers what factors shape immigrants’ social and political experiences in the US with implications for these groups’ representation in political office.Footnote ⁶ One factor that has received more attention from scholars of immigrant socialization in sociology than political science is immigrant assimilation and political incorporation. We outline the key theoretical arguments and findings from this research and then consider how these factors may shape Asian American political representation.

Classic sociological research on immigrant incorporation argues that immigrant groups adapt to American life by adopting the dominant social and cultural practices in their host society (Alba and Nee Reference Alba and Nee2003; Gordon Reference Gordon1964; Warner and Srole Reference Warner and Srole1945). These accounts center the process of “straight-line assimilation,” whereby immigrants and their children adopt American social, economic, and cultural practices – and lose those from their country of origin – over time and generations in a linear fashion (Warner and Srole Reference Warner and Srole1945). The degree to which immigrant groups must reject their own social practices differs across theories, ranging from “Anglo-conformity, which demanded the complete renunciation of the immigrant’s ancestral culture,” to “cultural pluralism,” in which some aspects of immigrant cultures become incorporated into mainstream American culture (Gordon Reference Gordon1964, p. 8). These theories generally conceive of assimilation as a process of economic and social mobility from one generation to the next (Ramakrishnan and Espenshade Reference Ramakrishnan and Espenshade2001, p. 876). As such, common measures include, “socioeconomic status, spatial concentration, language assimilation, and intermarriage” (Waters and Jiménez Reference Waters and Jiménez2005, p. 105). Although these theories were developed to explain the incorporation of European immigrant groups before 1965 legal reforms altered the racial composition of immigrants entering the US, some scholars argue that they also apply to racially diverse groups of new immigrants, including Asian Americans (Alba and Nee Reference Alba and Nee2003; Waters and Jiménez Reference Waters and Jiménez2005).

Our account aligns with other work which updates these theories by acknowledging that straight-line assimilation is not a foregone conclusion, especially for immigrants from racial minority groups. For example, research on “segmented assimilation” explores the possibility that the second-generation children of Latino and Black immigrants may not uniformly experience upward socioeconomic mobility (Gans Reference Gans1992; Portes and Rumbaut Reference Portes and Rumbaut2001; Portes and Zhou Reference Portes and Zhou1993). Instead, these authors argue there is variation in second-generation assimilation outcomes along lines of race, class, and residential settlement. While many of these accounts focus on Latino and Black children of immigrants, segmented assimilation is also a possibility for the children of Asian American immigrants who grow up in ethnic enclaves or low-income households (Fu and Hatfield Reference Fu and Hatfield2008; Hiller and Chow Reference Hiller, Chow and Bass2005; Zhou and Xiong Reference Zhou and Xiong2005).

It is important to note that theories of segmented assimilation argue that there are a range of assimilation outcomes for the children of immigrants, not that downward mobility is the only possible outcome (Zhou and Xiong Reference Zhou and Xiong2005). In this vein, Alba and Nee (Reference Alba and Nee2003) acknowledge that segmented assimilation may occur, arguing that it does not “rule out the possibility that assimilation in the form of entry into the mainstream has a major role to play in the future” (p. 9). In fact, they take positive socioeconomic outcomes and high rates of intermarriage among Asian Americans as evidence that theories of assimilation apply in a modern context. However, these are not the only measures of social and political assimilation. Others argue that Asian American assimilation is an inherently racialized process, as racial identity continues to play an important role in their lives, especially as it relates to politics, “regardless of their socioeconomic status and levels of acculturation” (Lee and Kye Reference Lee and Kye2016, p. 253).

Research in political science considers the implications of these theories for the civic incorporation of immigrant groups including Asian Americans. For example, several indicators of immigrant assimilation, such as generational status, years spent in the US, naturalization, and English language ability increase the likelihood of Asian American political participation and partisan acquisition (Chan, Nguy, and Masuoka Reference Chan, Nguy and Masuoka2024; Cho Reference Cho1999; Phan and Garcia Reference Phan and Garcia2009; Ramakrishnan and Espenshade Reference Ramakrishnan and Espenshade2001; Wong Reference Wong2000; Wong et al. Reference Wong, Ramakrishnan, Lee and Junn2011). However, this work also raises questions about assumptions of linear assimilation or the accumulation of political resources over time. For example, Asian American and Latino political engagement may not consistently increase across years and generations spent in the US (Carlos Reference Carlos2018; Jones-Correa Reference Jones-Correa1998; Lee et al. Reference Lee, Ramakrishnan and Ramírez2006; Lien Reference Lien2004; Ramakrishnan and Espenshade Reference Ramakrishnan and Espenshade2001). Moreover, socioeconomic status, another key measure of assimilation, does not consistently predict higher rates of civic engagement among Asian Americans (Cho Reference Cho1999; Junn Reference Junn1999; Lien et al. Reference Lien, Collet, Wong and Ramakrishnan2001). Finally, factors distinctive to the immigrant experience, such as the political regime type in countries of origin, political engagement before migration, and transnational political ties shape Asian American civic engagement (Collet and Lien Reference Collet and Lien2009; Lee et al. Reference Lee, Ramakrishnan and Ramírez2006; Lien Reference Lien2010; Ramakrishnan Reference Ramakrishnan2005; Wong et al. Reference Wong, Ramakrishnan, Lee and Junn2011).

Building on this literature, we argue that immigrant assimilation (i.e., the integration of an ethnic group into a broader society) should lead to greater representation in political office, although this outcome is unlikely to be uniform or immediate. While immigrant groups such as Asian Americans may in general become more incorporated in the political realm over time and generations, this need not always be the case (Fu and Hatfield Reference Fu and Hatfield2008; Zhou and Xiong Reference Zhou and Xiong2005). Therefore, we argue that Asian American population size does not immediately lead to increased political representation. Rather, we build on theories of segmented assimilation and argue that the gradual acquisition of civic resources among members of the Asian American community produces descriptive representation over a period of time.

According to such theories of immigrant assimilation and political incorporation, as members of immigrant populations become more incorporated within a society, they should have greater opportunity to engage in politics for reasons of legal status (e.g., enfranchisement due to citizenship), acculturation (e.g., growing recognition of salient political cleavages, increased knowledge of the political system, exposure to norms of political participation), and resource acquisition (e.g., acquiring levels of education and wealth that permit greater participation in politics). As immigrant groups become incorporated in these ways, they will be more likely to produce candidates for public office as well as be able to support these candidates via voting, organizing, donating, and other forms of civic participation. For these reasons, the more incorporated an immigrant community is in a constituency, the more likely they should be represented in office.

Drawing on these arguments, we develop several predictions about the relationship between immigrant incorporation and representation in state legislatures. The first is that Asian American population size alone does not predict legislative representation, as theories of shared racial or ethnic identity and representation predict. Instead, the historic population hypothesis predicts that the likelihood of Asian American representation in state legislatures increases as the local Asian American community becomes more established, which we measure in terms of population size in prior decades (while holding constant contemporaneous population size).

However, since theories of segmented assimilation argue that time alone is not sufficient for political incorporation, we also consider whether the acquisition of civic resources increases the likelihood of representation. The civic resource acquisition hypothesis predicts that the likelihood of Asian American state legislative representation increases when a higher percent of the Asian American population possesses key resources. In our analyses, we consider the proportion of the Asian American community who are citizens, fluent in English, multiracial, college educated, or high income.

In testing whether these theories explain Asian American representation in state legislatures, we draw on Lee and Kye’s (Reference Lee and Kye2016) account of “racialized assimilation,” which centers the continuing significance of race for Asian Americans despite relative socioeconomic and social assimilation. Contrary to straight-line theories of assimilation, we do not expect that integration leads to full social incorporation into “mainstream [white] society” such that the interests of Asian Americans are no different from those of whites. Rather, even assimilated Asian American immigrants who have access to civic resources maintain a distinct Asian American racial identity that shapes their life experience in the US (Kim Reference Kim1999; Lee and Kye Reference Lee and Kye2016).Footnote ⁷ In turn, regardless of their level of assimilation, Asian Americans may benefit from descriptive representation and vote for co-ethnic representatives. By this account, assimilation and the acquisition of civic resources leads to political incorporation, providing Asian Americans with the tools for civic engagement.

While European immigrants largely followed a linear assimilation trend (Gordon Reference Gordon1964), this racialized model of assimilation and political incorporation may also apply to diverse immigrants from ethno-racial minority groups beyond the Asian American community. Some examples include Black and Latino immigrants from Africa, the Caribbean, Central America, and South America. However, it is important to note that patterns of immigrant assimilation and political representation in other immigrant communities may differ from those for Asian Americans because these other groups are often highly concentrated into districts where many voters are from the same racial group and have different social experiences in the US. Furthermore, as Black immigrants assimilate, they may become incorporated into a politically active non-immigrant African American community (Rogers Reference Rogers2006). Future work may explore the extent to which these arguments apply to diverse non-white immigrant communities.

Data on Asian American representation in the states

Does social integration lead to a greater Asian American presence in state legislatures? How? Answering these questions requires data on Asian American representation. Below, we describe a new dataset of Asian American political representation in state legislatures across time, and detail two complementary approaches we adopt to measuring assimilation and social integration.

The primary dataset we use, originally collected by Klarner (Reference Klarner2022) and shared with the authors, contains the racial and ethnic identities of all American state legislators elected to office between 1970 and 2020. The dataset was constructed by a team of researchers using rosters maintained by interest groups, archival biographical records published by state legislatures, and (in recent years) state legislative and candidate websites. More details on the procedures for identifying legislator race and ethnicity as well as validation of the codings can be found in Section S1 of the Supplementary Material.

During this time period, there are 476 unique state legislators with an Asian or Pacific Islander background.Footnote ⁸ In Table S2 in the Supplementary Material, we provide additional data on the nationality and ethnicity of these legislators, the states they hold office in, and their political parties. Of particular note from the table is the partisan split: Democratic Asian American state legislators outnumber their Republican counterparts by approximately a 4 to 1 margin.Footnote ⁹

When and where does Asian American representation in the states occur? In Figure 1, we display two plots showing the growth of Asian American representation over time as well as the location of Asian American state representatives in recent years. The top panel compares Asian population growth in the United States between 1970 and 2020 as a percent of total US population, alongside the percent of all state legislators with an Asian American background. The timing of population growth versus representation is consistent with an assimilation-based story of political incorporation; from 1970 onwards the Asian share of the population grows steadily. However, Asian American representation does not begin to increase meaningfully until approximately 2000, at which point it too starts growing. This suggests that the political incorporation of Asian Americans does not follow theories of minority population size, as representation does not increase immediately nor linearly with population size.

Figure 1. Temporal and geographic dynamics of Asian American representation in the states.

Note: The top panel displays the growth of Asian American population in the United States (dashed line; data from decennial US Censuses) and Asian American representation in state legislatures (solid line) between 1970 and 2020, each expressed as the percent of total (population/legislators). The bottom panel shows lower chamber state legislative districts in the contiguous US for the 2012–2021 redistricting cycle (data on state legislative district boundaries come from the US Census Bureau). Solid black dots indicate state legislative districts with at least one Asian American representative during this time period.

The bottom panel shows the location of Asian American state representatives in the most recent full redistricting cycle (2012 to 2021). As the map reveals, there are areas where Asian American representation is particularly concentrated, such as in the Northeast, California, and large urban areas in the Midwest.

Our main analyses take advantage of variation in Asian American representation across both time and districts to test our hypotheses about immigrant political incorporation and representation in political office. The first analysis examines how Asian American political representation is associated with Asian population growth, both contemporaneous and in the near past, using data on Asian American population size by state dating back to the 1960s. We collect decennial US census data on the Asian population in each state by decade, as well as auxiliary variables likely associated with Asian American representation. The hypothesis we test (the historic population hypothesis) is whether Asian American population growth in earlier decades is associated with contemporaneous Asian American representation after conditioning on contemporaneous Asian American population size. In other words, when comparing two states with a similar presence of Asian Americans in the electorate, do we observe greater Asian American representation in the state where Asian Americans arrived earlier? This state-level analysis offers an initial test of theories of immigrant assimilation and political incorporation alongside theories of co-racial representation.

The second analysis we conduct uses finer-grained geographic data (at the expense of examining a shorter time period) to test expectations about access to civic resources in immigrant communities and representation in political office. For this analysis, we evaluate the determinants of Asian American representation at the state legislative district level. To do so, we collect data on state legislative districts from the American Community Survey (ACS). From 2006 onwards, the ACS surveys contain information not just about the size of the Asian population in each state legislative district, but also the citizenship, language, education, and income of the district at large as well as the Asian community. We use these variables, which permit a deeper examination of how assimilation along several dimensions (citizenship, language use, education, etc.), to test the civic resource acquisition hypothesis. We conclude with a brief comparison of co-ethnic versus pan-ethnic representation at the state legislative district level.

Population growth and Asian American representation

We begin our empirical analyses by looking at how the timing of Asian population growth in individual US states corresponds to representation in state legislatures. Our strategy is to look at states with varying Asian population levels over time, and determine if population growth is associated with immediate and/or future representation of Asian Americans in the state legislature. Explanations of minority representation that highlight co-ethnic voting or critical minority population mass predict we should find the former, while theories of immigrant assimilation and incorporation suggest we should observe the latter.

Because historical data on the Asian population in each state comes from decennial censuses, the unit of observation in these regressions is a state by decade combination.Footnote ¹⁰ In the simplest model, the specification can be given as

$$ {\mathrm{Asian}\ \mathrm{R}\mathrm{e}\mathrm{p}\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}_{s,d}={\beta}_1{\mathrm{Asian}\ \mathrm{P}\mathrm{o}\mathrm{p}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}_{s,d}+{\mathbf{X}}_{s,d}\gamma +{\alpha}_{\mathrm{s}}+{\delta}_{\mathrm{d}}+{\epsilon}_{s,d}, $$

where Asian Representation _s,d refers to the percent of state legislative seats in state s and decade d held by Asian Americans, and Asian Population _s,d the Asian percent of the population in the state at the beginning of the decade. Put differently, for each state, we regress the percent of Asian American legislators in the 1980s on the Asian population in the year 1980, the percent of Asian American legislators in the 1990s on the population in the year 1990, and so on.

In the equation above, α_s and δ_d represent state and decade fixed effects, respectively. The inclusion of state fixed effects ensures the model uses solely variation in population size within states (not across states) to estimate the relationship to representation. The decade fixed effects control for the secular growth in Asian American state legislators over time, accounting for any nation-wide changes in Asian American representation. In addition to these fixed effects, we control for factors that vary across time within a state (X _s,d above). The control variables used are the state median income, the percent college educated in the state, the percent of other racial minorities in the state, and the Democratic vote in the most recent presidential election.

While the simplest model uses only the Asian population in the state at the beginning of the decade, additional models evaluate whether co-ethnic population in earlier decades predicts Asian American representation in later decades, independent of the contemporaneous population. To accomplish this, lagged versions of the Asian Population variable are included along with the contemporaneous one. For instance, with one lag included the equation becomes

$$ {\displaystyle \begin{array}{c}{\mathrm{Asian}\ \mathrm{R}\mathrm{e}\mathrm{p}\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}_{s,d}={\beta}_1{\mathrm{Asian}\ \mathrm{P}\mathrm{o}\mathrm{p}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}_{s,d}+{\beta}_2{\mathrm{Asian}\ \mathrm{P}\mathrm{o}\mathrm{p}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}_{s,d{\textstyle \hbox{-}}1}\\ {}+{\mathbf{X}}_{\mathrm{s},\mathrm{d}}\gamma +{\alpha}_{\mathrm{s}}+{\delta}_{\mathrm{d}}+{\epsilon}_{s,d}.\end{array}} $$

In this model, β ₂ can be interpreted as the predicted difference in Asian American representation associated with the state’s Asian population size at the beginning of the previous decade, holding constant the population at the beginning of the current decade. Intuitively, the model allows us to compare two states, each with an equal Asian population size at the beginning of the decade but one with a larger Asian population the decade prior. If the latter type of state has more Asian American legislators in office, this suggests that it is not just the size of the Asian population that matters but also how long they have been members of the community.

Table 1 displays the results for four such models, beginning with no lagged population variables and then gradually adding one-, two-, and three-decade lagged population variables.Footnote ¹¹ The first model, in Column 1, confirms that larger Asian American populations are indeed associated with higher levels of Asian American representation, as theories of co-racial representation predict.

Table 1. Asian American population growth and representation in state legislatures

Note. Standard errors clustered by state shown in parentheses. p-values from two-tailed tests: *p < 0.05; **p < 0.01

Columns 2–4 of Table 1 qualify this finding considerably, however. These columns reveal that representation depends on how established local Asian American populations are, as the historic population hypothesis predicts. The models in these columns, which include one or more lagged population variables, do a much better job of predicting patterns in Asian American representation, as evidenced by the increasing adjusted R ² value across columns. The coefficient estimates reveal that it is not just the size of the contemporaneous Asian population that matters, but when that population growth occurred, in line with theories of immigrant assimilation and incorporation. In fact, when the lagged population variables are included, the estimated effect of contemporaneous population is negative. This suggests that after taking into account the presence of Asian population in prior decades, recent Asian population growth – due to immigration or other reasons – may actually have a negative impact on Asian American representation. Such a pattern is consistent with multiple explanations rooted in existing literature, such as a backlash effect against immigrants among non-immigrant community members (Abrajano and Hajnal Reference Abrajano and Hajnal2017), low rates of political engagement among Asian American newcomers (Lien et al. Reference Lien, Collet, Wong and Ramakrishnan2001), or weak pan-ethnic racial identity among less-incorporated minority populations (Lien, Conway, and Wong Reference Lien, Conway and Wong2004).

In contrast, the relationship between earlier Asian population levels and contemporaneous representation is consistent and positive. In Columns 2–4 the coefficients from all of the lagged population variables are all positive, and most are significant. In Column 4, where three lagged variables are simultaneously included, the coefficient sizes imply that Asian population levels two decades earlier are the strongest predictors of how many Asian American state legislators currently hold office.

To illustrate these relationships more clearly, Figure 2 compares the contemporaneous and lagged Asian American population to Asian American state legislative representation in the four states with the largest Asian population share in recent years. As can be seen in the figure, in all four states there is a sizeable representation gap between the contemporaneous Asian population and the share of state legislative seats held by Asian Americans. However, there is a much closer relationship between the share of current state legislative seats and the Asian American population from two decades prior. In fact, in two states (California and Washington) Asian Americans hold a slightly greater percent of seats in the state legislature by 2010 than their population share in 1990.

Figure 2. Population and representation dynamics in four example states.

Note: The figure displays Asian American state legislative representation as well as contemporaneous and lagged Asian American population in the four states – California, New Jersey, New York, and Washington – with the largest Asian share of the population between 1980 and 2010.

To summarize, an analysis of Asian American representation and population dynamics shows population growth does lead to political incorporation, but that this is a process takes place over multiple decades.

District demographics and Asian American representation

Our second analysis moves from representation in the aggregate to dyadic representation, evaluating which characteristics of state legislative districts – in particular, civic resources possessed by the Asian population in a district – are associated with having an Asian American state representative or senator. As mentioned above, we collect data on state legislative districts from the American Community Survey beginning in 2006, the first year ACS data are available. In addition to looking at the size of the Asian population in each state, we consider how indicators of immigrant assimilation and social integration along multiple dimensions (e.g., legal, social, and socioeconomic) – citizenship, language, multiracial status, education, and income – are associated with co-racial representation of the Asian population.

We begin by regressing a binary indicator for electing an Asian American representative in a given district in a given election cycle on the Asian percentage of the population in the district, to establish a baseline for the relationship between Asian population and co-racial political representation.

Second, we estimate the same regression but separate the Asian population by citizenship status, including as independent variables the population percentage of US-born Asian citizens, the population percentage of foreign-born naturalized Asian citizens, and the population percentage of foreign-born Asian non-citizens. Our expectation is that the percentage of native-born Asian citizens will be most strongly associated with pan-ethnic representation and the percentage of Asian non-citizens the least strongly associated (Cho Reference Cho1999; Junn Reference Junn1999; Wong et al. Reference Wong, Ramakrishnan, Lee and Junn2011). In the latter case, given that citizenship is a legal requirement for voting, a greater number of Asian non-citizens (holding constant the number of Asians who are US citizens) should not be associated with pan-ethnic voting or candidacies. We expect US-born Asian citizens are more likely to be represented by Asian Americans than naturalized Asian citizens as the former have spent a greater portion of their lives in the United States and are more likely to be more incorporated in the political culture (Wong et al. Reference Wong, Ramakrishnan, Lee and Junn2011).

Third, we separate the Asian population in the district into the percent that report speaking English as their primary language at home versus those that report speaking another language. English proficiency is an indicator of how socially integrated Asian Americans are, with greater proficiency revealing greater assimilation (Cho Reference Cho1999; Lee and Kye Reference Lee and Kye2016; Waters and Jiménez Reference Waters and Jiménez2005). Furthermore, a lack of English proficiency should make it more difficult for individuals to absorb information about politics (advertisements, debates, signs, etc.) and may be associated with less engagement in a political system that does not prioritize language access (Nguyen Reference Nguyen2022a).

Fourth, we separate the Asian population into those identifying solely as Asian versus those who identify as another race or ethnicity (e.g., white, Latino, Black).Footnote ¹² Those who identify as Asian American alongside other racial groups are more likely to be incorporated in American society and make an intentional decisions to identify both with the Asian American community as well as one or more non-Asian ethnoracial groups (Davenport Reference Davenport2018; Masuoka Reference Masuoka2017).

Fifth, we separate the Asian population into those who report having a four-year college degree or more education versus those with less than a four-year degree. Higher education is an indicator of social incorporation, and those with a college degree should be better prepared to understand and participate in American politics. Although research on Asian American civic engagement does not find that income and education consistently predict turnout (Junn Reference Junn1999; Lien et al. Reference Lien, Collet, Wong and Ramakrishnan2001), in studies of African American and women’s representation, education (as well as income) has been found to be an important predictor of representation (Karnig Reference Karnig1979; Marschall and Ruhil Reference Marschall and Anirudh2006; Robinson and Dye Reference Robinson and Dye1978; Smith, Reingold, and Owens Reference Smith, Reingold and Owens2012; Thomsen and Swers Reference Thomsen and Swers2017; Wright and Zhu Reference Wright and Zhu2021).

Accordingly, we also separate the Asian population into low income households ($50,000 of household income per year or less), middle income households (between $50,000 and $100,000 of annual household income), and high income households (anything above $100,000 of annual household income). The logic here is straightforward; as individuals have greater economic resources, following and participating in politics becomes easier. Thus we expect a stronger association between the percent Asian population in a district and pan-ethnic representation among middle and high income households.

We evaluate these expectations, which stem from the civic resource acquisition hypothesis, using our data on the race of state legislators and the socio-demographic characteristics of the districts they represent, estimating a series of logistic regression models decomposing the Asian population in each district into different categories based on the hypotheses above.Footnote ¹³ In these models, the unit of observation is a state legislative district in a specific election year, and the dependent variable is whether the elected representative is Asian American or not.

In addition to the above-discussed characteristics of the Asian population, all models control for variables that might also influence the likelihood a district is represented by an Asian American. These include the percent of a district’s entire population (i.e., non-Asians and Asians collectively) that is college-educated, the median income of the district, and the percent of a district’s residents are Black, Hispanic or Latino, or from another non-white category, relative to the omitted baseline category (the percent of a districts’ residents who are white).Footnote ¹⁴ We also include the measure of district conservatism developed by Tausanovitch and Warshaw (Reference Tausanovitch and Warshaw2013) to control for differing propensities to elect an Asian American across party and ideology. Finally, and importantly, we include state-election year fixed effects to control for any differences across years and states in the probability of Asian Americans getting elected to state office. Consequently, these regressions compare demographic differences across districts within a state within a particular election year.

The results of these models are displayed graphically in Figure 3. To establish a baseline, the first model simply includes the Asian share of the population without any further decomposition. Coefficient estimates above this baseline indicate that Asian individuals in that group are more likely to be represented by an Asian American legislator than Asian residents writ large, while a smaller coefficient implies the inverse.

Figure 3. Predictors of district representation by an Asian American state legislator.

Note: Shapes indicate coefficient estimates from logistic regression models with control variables (general district ideology, racial composition, income, education) and state-year fixed effects included. Standard errors clustered at the district level. Solid lines display 95% confidence intervals. Data span 2006 to 2020. The full set of estimates, including coefficients for control variables, are provided in Table S6 in the Supplementary Material.

The subsequent results largely align with our expectations and support the civic resource acquisition hypothesis. Beginning with the citizenship variables, we find that the percent of Asian citizens in a district is associated with representation by an Asian American, while the percent of Asian non-citizens has effectively no relationship with the likelihood of co-ethnic representation. There does not appear to be much difference in pan-ethnic representation for US-born versus naturalized citizens, however. One explanation for this is that naturalized citizens have made a conscious and oftentimes costly decision to acquire citizenship (Jones-Correa Reference Jones-Correa1998). Given naturalized citizens often have made the difficult decision to renounce allegiance to their country of origin, it is unsurprising that they might also make the conscious decision to participate in American politics.

Turning next to the language variables, we find some evidence that English-speaking Asians are more likely to be represented by an Asian American than non-English speaking Asians. The coefficient estimate for the English-speaking Asian percent of a district’s population is almost twice as large as the coefficient for the non-English speaking Asian population. The relationship between language use and representation, while not precisely estimated (the p-value from a two-tailed F-test for difference of coefficients is 0.064) is in the direction predicted by our theory.

Moving next to the racial identification variables, we find that Asians who identify with another race are more likely to be represented by an Asian legislator in office than those who identify as Asian alone (p = 0.097). This might be surprising from one perspective, as we might expect those who identify with other racial and ethnic groups in addition to their Asian heritage to support those other racial and ethnic groups politically, perhaps to the detriment of Asian candidates. However, taken as an indicator of Asian social assimilation, the relationship between multiracial identification and Asian representation accords with our theoretical expectations (Masuoka Reference Masuoka2017).

The results for the education variables are also in line with expectations. The coefficient for pan-ethnic representation of college educated Asians is almost twice as large as the baseline coefficient. In contrast, there is essentially no relationship between the percent of non-college educated Asians in a district and the likelihood of representation by an Asian American.

For income, we find evidence of a non-linear relationship. While the percent of middle-income Asian Americans in a district is more predictive of co-ethnic representation than the number of low-income Asian Americans (p = 0.088), the relationship between the number of high-income Asian Americans in the district and Asian American representation is negative. This latter finding, while unexpected, may be connected to the districts where high-income Asian Americans live, if they are quite different than districts where low- and middle-income Asian Americans live. This also aligns with findings that increasing income does not consistently predict higher turnout rates among Asian American voters (Junn Reference Junn1999; Lien et al. Reference Lien, Collet, Wong and Ramakrishnan2001).

To summarize, we find that the districts represented by Asian Americans are, as theories of shared racial identity and representation suggest, are those with a high concentration of Asian residents. However, the social incorporation of that Asian population also matters. Places with large numbers of middle-income, English-speaking, multi-cultural, Asian American citizens are most likely to elect an Asian American to represent them in their state legislature.

Co-ethnic versus pan-ethnic representation

Up until this point, we have examined Asian legislators and residents as a single aggregate group. However, Asian as a pan-ethnic racial group covers a wide variety of different nationalities and ethnicities. Previous research on Asian American political behavior argues and finds that co-ethnic preferences are particularly strong, although there is evidence a pan-ethnic Asian American identity also exists (Cho Reference Cho2001; Le Espiritu Reference Le Espiritu1992; Junn and Masuoka Reference Junn and Masuoka2008; Masuoka Reference Masuoka2006).

To what extent are the above findings driven by co-ethnic versus pan-ethnic representation? We evaluate this question by re-estimating the regression models of the previous section, but including separate population and representation variables for the largest ethnic subgroups in our data: Chinese, Indian, Korean, Filipino and Japanese. The results are shown in Section S7 of the Supplementary Material. As Supplementary Figure S7 shows, while the representational link is generally strongest for legislators and residents of the same Asian ethnic subgroup, there is evidence of cross-ethnic representation for Asian American state legislators as well. In four out of the five ethnic subgroups, there is a strong, positive, and significant relationship between ethnic subgroup population size and the likelihood of being represented by a member of that ethnic subgroup. In all five cases, after controlling for ethnic subgroup population the relationship between the broader Asian population in the district and pan-ethnic representation is weaker than the relationship with co-ethnic representation.Footnote ¹⁵ This corroborates theories of co-ethnic voting and highlights the strength of bonds between Asian American voters and representatives from the same ethnic subgroup (Leung Reference Leung2022; Sadhwani Reference Sadhwani2022a). However, the results also support the idea that the formation of a pan-ethnic Asian political identity and coalitional group may be a consequence of social incorporation (Le Espiritu Reference Le Espiritu1992).