Across Western democracies, studies evaluating political discrimination have documented racial bias not only in politicians' responsiveness to citizens (Alizade and Ellger Reference Alizade and Ellger2022; Costa Reference Costa2017; Mendez and Grose Reference Mendez and Grose2018; White, Nathan, and Faller Reference White, Nathan and Faller2015) but also in the actual selection of new candidates (Brouard and Tiberj Reference Brouard and Tiberj2010; Dancygier et al. Reference Dancygier2021; Eriksson and Vernby Reference Eriksson and Vernby2021; Kalla, Rosenbluth, and Teele Reference Kalla, Rosenbluth and Teele2018; Soininen Reference Soininen, Bird, Saalfeld and Wüst2011; Tolley Reference Tolley2019). Together, this body of research points to a significant problem: the political system signals to minorities that it does not include their voices and that it does not welcome diversity, thus posing challenges to democratic legitimacy.
A way to offset this would be greater descriptive representation of marginalized groups among the political classes (Geese Reference Geese2022; Hayes and Hibbing Reference Hayes and Hibbing2017; Stout, Tate, and Wilson Reference Stout, Tate and Wilson2021). Nonetheless, research consistently finds that minorities appear to be underrepresented because they are not running for elected office in the first place (Gonzalez Juenke and Shah Reference Gonzalez Juenke and Shah2015; Scott Reference Scott2018; Shah Reference Shah2014; Shah, Juenke, and Fraga Reference Shah, Juenke and Fraga2022; Shah, Scott, and Juenke Reference Shah, Scott and Juenke2019). It is important to note that this is not necessarily due to a lack of interest in pursuing office among these groups (Dancygier et al. Reference Dancygier2021; Shah Reference Shah2014; Reference Shah2015).Footnote 1
When racial and ethnic minorities do hold elected office, tangible outcomes that improve the day-to-day lives of these groups are generated. For example, relationships between law enforcement and communities of colour ease (Aneja and Ritadhi Reference Aneja and Ritadhi2022; Christiani et al. Reference Christiani2022), political knowledge and participation increase among minority constituents (Fisher et al. Reference Fisher2015; Griffin and Keane Reference Griffin and Keane2006; Wolak and Juenke Reference Wolak and Juenke2021), and voters evaluate governmental responsiveness more positively (Arnesen and Peters Reference Arnesen and Peters2018; Banducci, Donovan, and Karp Reference Banducci, Donovan and Karp2004). Indeed, several studies have shown that experiences of discrimination and the desire to improve public policy can motivate members of underrepresented groups to take political action (Besco et al. Reference Besco2022; Oskooii Reference Oskooii2020; Pantoja, Ramirez, and Segura Reference Pantoja, Ramirez and Segura2001).
Before entering politics, a potential candidate's primary consideration is the non-monetary ‘cost’ of entry. Racial and ethnic minorities, who are often aware of the backlash they face from majority populations (see, for example, André and Dronkers Reference André and Dronkers2017; Lu Reference Lu2020), weigh the extent to which they would face discrimination if they were to enter politics (Brown and Lemi Reference Brown and Lemi2021; Phillips Reference Phillips2021). Alexandria Ocasio-Cortez's experience is an illustrative example of the costs minorities incur when entering politics. As a young Latina Member of Congress who assumed office without any prior political experience, she reflected in a GQ interview, ‘Others may see a person who is admired, but my everyday lived experience here is as a person who is despised … Imagine working a job and your bosses don't like you and folks on your team are suspicious of you. And then the competing company is trying to kill you.’Footnote 2 Ocasio-Cortez's comments demonstrate how unwelcome she feels in politics, not just by political opponents but also by senior members of her own party. While Ocasio-Cortez provides a vivid example of the non-monetary ‘costs’ of entry, underrepresented groups may also be the target of subtler behaviours that result in feeling unwelcome. Experimental research in the United States found that chairs from both major parties perceived Latin and black candidates as less viable and substantially less likely to win and, therefore, strategically chose to recruit and support other candidates (Doherty, Dowling, and Miller Reference Doherty, Dowling and Miller2019). In the Swedish system, where parties control nominations, an interview study confirmed that immigrant politicians have already experienced discrimination at the nomination stage (Blomqvist Reference Blomqvist2005, 90). More extensive studies using administrative data have also confirmed that immigrants are underrepresented in Swedish politics, in large part due to decisions made by party elites at the nomination stage (Dancygier et al. Reference Dancygier2015), which are key to candidate recruitment into politics (Soininen Reference Soininen2010). In Soininen and Qvists's (Reference Soininen and Qvist2021, 568, 570) interviews with members of nomination committees, respondents opine that immigrants ‘stand out’ when they ‘don't speak exactly like we do’ and ‘must be able to create contacts with other parts of a relatively white, established […] political culture’ to succeed.
In this article, we descriptively assess whether members of minoritized groups expect to face discrimination if they were to enter politics. We introduce a new measure of expected discrimination that captures how welcome citizens would expect to feel among other politicians if they were elected to office.Footnote 3 The basic logic underlying our measure is the assumption, drawn from the theory of ethnic homophily, that members of an ethnic group tend to prefer interacting with members of their own group. From this, it follows that if members of the ethnic majority are overrepresented in politics, we might expect non-majority groups to feel less welcome. This, in turn, could exacerbate adverse consequences on the ‘supply side’ of minority representation.
We draw on three large-scale public opinion surveys fielded in the United States and Sweden, where respondents answered similar questions about whether they or members of their group would anticipate feeling welcome among other politicians if they were to enter politics. While the political systems of the United States and Sweden differ in many respects, they are both cases where questions of descriptive representation are normatively pressing, given the large shares of minorities living in each country and the rapidly changing population demographics.Footnote 4 In Sweden, 26.3 per cent of the population is either foreign-born or has two parents who are foreign-born as of 2021, a rise from 19 per cent in 2010 (Statistics Sweden 2022), while the US non-Hispanic white population fell from 72.4 per cent in 2010 to 61.6 per cent in 2020 (United States Census 2021).
What is more, racial/ethnic minorities are underrepresented in politics across both country contexts. The literature on the underrepresentation of immigrant-origin people in Sweden finds that exclusion is not due to a dearth of resources or political interest but is driven by party gatekeepers who exclude immigrants from higher office (see, for example, Adman and Strömblad Reference Adman and Strömblad2015; Dahlstedt and Hertzberg Reference Dahlstedt and Hertzberg2007; Dancygier et al. Reference Dancygier2015; Dancygier et al. Reference Dancygier2021; Lindgren, Nicholson, and Oskarsson Reference Lindgren, Nicholson and Oskarsson2021; Lindgren and Österman Reference Lindgren and Österman2022). Similarly, the US race and ethnic politics literature finds that minority candidates emerge at lower rates than whites (Fraga, Juenke, and Shah Reference Fraga, Juenke and Shah2020) and that party elites disproportionately assist whites over minority candidates (Fraga and Hassell Reference Fraga and Hassell2021).
Across both country contexts, racial/ethnic minorities are less likely than their majority citizen counterparts to indicate that they would feel welcome among other politicians. These results persist for each minority group examined. We also extend our analysis in a number of ways. First, we replicate the results by adding controls for socioeconomic status, demographics, and political interest to examine if minority status per se makes minorities feel less welcome or if some potential correlate of minority status, such as education, is driving the observed effect. After adding controls, the Swedish results hardly change, and the US results remain statistically significant, though somewhat weakened. Second, we use a unique feature of the Swedish survey from 2021 to cross-validate the perceptions of minority and majority groups. We find that members of the majority group agree that minorities are less likely to feel welcome in politics. We then explore the connection between expectations of feeling welcome and interest in running for office. In two out of three surveys, there is a strong relationship between the two, attesting to the relevance of expected discrimination for political candidacy.Footnote 5 Using the US survey, we can show that, when controlling for measures that tap into past experiences of discrimination that are positively associated with interest in running for office, the impact of our new measure of expected discrimination remains virtually unchanged. Finally, we replicate our main analysis by restricting the sample to only include those with high political engagement, measured by their high level of political interest and involvement in civil society organizations. The results are substantively similar to the main results, suggesting that divergent perceptions of the inclusiveness of politics are also present for individuals who are a realistic part of the candidate pool.
Cases, Data and Methods
Our analysis draws on three surveys: one fielded in the United States in 2021 and two conducted in Sweden in 2017 and 2021.Footnote 6 Comparative politics scholars studying minority politics have long recognized that conceptualizations of minority status should be attentive to context and have struggled with valid cross-national measures to distinguish minorities from non-minorities (Bloemraad Reference Bloemraad2013).
We construct our ‘minority’ variables differently for Sweden and the United States, taking into account the unique contexts of each of our cases. Therefore, how a ‘minority’ is classified differs between our two country cases. While no established definition exists, American political scholars have, for historical (and practical) reasons, focused on ‘visible’ minorities, whereas European scholars have, for similar reasons, focused on the background of migration. We follow these traditions depending on the country context being examined. Our analyses focus on migration background in the Swedish context and racial background in the US context. Sweden and the United States are ‘most different systems’ in terms of their electoral systems and immigration histories, and if minorities in both countries expect to face discrimination if they were to enter politics, this suggests our results may generalize to other cases as well.
The US data comes from the 2020 Comparative Multiracial Post-Election Survey (CMPS). The CMPS was fielded between April and August 2021 and intentionally oversampled racial and ethnic minorities. Our analyses include 3,749 whites, 3,121 Asians, 4,363 blacks, and 3,071 Latinos. CMPS respondents were asked, ‘Imagine you are a new politician in the area where you live, would you feel welcome at meetings with other politicians?’ Response options ranged from ‘Yes, absolutely’, ‘Maybe’, ‘No, probably not’, to ‘No, absolutely not.’Footnote 7 Given the ordinal nature of the data, in our main analysis, we will focus on the share answering ‘Yes, absolutely’ in response to the above statement.Footnote 8
The first Swedish survey was conducted between May and September 2017. Similar surveys were sent to both politicians and eligible voters. Our analysis only relies on the latter sample since we are interested in potential (rather than actual) candidates. This sample includes 1,948 individuals who grew up in Sweden and 646 who grew up outside Sweden. Respondents were asked ‘Imagine you are a new politician in the municipal council, do you think you would feel welcome?’ The response options were ‘Yes, absolutely’, ‘Yes, maybe’, ‘No, probably not’, and ‘No, absolutely not.’
The second Swedish survey was conducted between September and December 2021. It included the question: ‘Thinking about the municipality where you live, do you think that a newly elected politician would feel welcome in meetings with other politicians if the newly elected politician …’ followed by the prompts, ‘Mainly grew up in Sweden’, ‘Mainly grew up in Europe’, and ‘Mainly grew up outside Europe.’ Here, the response options were ‘Yes, absolutely’, ‘Yes, maybe’, ‘No, probably not’, and ‘No, absolutely not.’ We then match this to information about where the respondents mainly grew up. This allows us to create a measure that reflects the extent to which respondents believe members of their own group would feel welcome among other politicians. The survey data analyzed comprised 1,292 individuals who mainly grew up in Sweden and 102 individuals who grew up outside of Sweden.
As is clear, all three surveys posed questions that measured respondents' expectations about who is welcome in politics. While the surveys used slightly different question wordings and cannot be directly compared, they enable us to assess whether a general pattern exists in how welcome minorities feel in politics across cultural and political contexts.Footnote 9 Appendix A.1 provides more details about each survey.
Main Results
Figure 1 displays the proportion of respondents who expect that they, or members of their ethnic group, would feel welcome among other politicians if elected to office. In the United States, around 25 per cent of white survey respondents indicated that they would feel welcome, whereas the corresponding figures for Latinos, Blacks and Asians were 21 per cent, 21 per cent, and 20 per cent, respectively. Turning to the 2017 Swedish survey, 25 per cent of respondents who grew up in Sweden reported that they would feel welcome, whereas the corresponding figures for respondents who grew up in Europe and outside Europe are 18 per cent and 14 per cent, respectively. In the 2021 Swedish survey, we asked respondents whether members of their own group (rather than themselves) would feel welcome among other politicians. With this alternative question wording, the ethnic gap in expected discrimination is larger.Footnote 10 The share who grew up in Sweden who think that members of their own group would feel welcome is over 60 per cent, whereas the corresponding figures for the respondents who grew up in Europe and outside Europe are 27 per cent and 28 per cent, respectively.
Figure 1. The share who indicate that they (U.S. Survey 1 and Swedish Survey 1) or members of their ethnic group (Swedish Survey 2) would feel welcome among other politicians, by minority status.
Note: The number of observations is N = 14,395 (US Survey), N = 2,594 (Swedish Survey 2017) and N = 1,394 (Swedish Survey 2021).
It is clear from these results that, in both the American and Swedish contexts, minorities expect more discrimination from fellow politicians than the majority group if they were to enter politics. Figure 2, in which we regressed our measures of feeling welcome on our indicators of minority status, further reinforces this finding. All estimates are negative and statistically significant. Moreover, many group differences are quite substantial. In the United States, Blacks are 16 per cent less likely to say they expect to feel welcome than Whites, and the corresponding figures for Latinos and Asians are similar.Footnote 11 In the 2017 Swedish survey, respondents who grew up in Europe and outside Europe were 25 per cent and 44 per cent less likely to feel welcome than the majority group. Finally, the likelihood of feeling welcome is more than halved when comparing respondents who mainly grew up outside Sweden to the majority group in the 2021 Swedish survey.
Figure 2. Group differences in the share who say they (US Survey 1 and Swedish Survey 1) or members of their ethnic group (Swedish Survey 2) would feel welcome among other politicians.
Note: Plots display coefficient estimates with 95 per cent confidence intervals from models that regress expectations of feeling welcome on race/ethnicity in the three surveys. The reference category in the upper left plot is ‘White.’ The reference category in the upper right and lower plot is “Grew up in Sweden.’ The number of observations is N = 14,395 (U.S. Survey), N = N = 2,594 (Swedish Survey 2017) and N = 1,394 (Swedish Survey 2021). Confidence intervals are based on robust standard errors.
Extensions: Potential Roots and Effects of Expected Discrimination
Next, we examine the potential roots and effects of expected discrimination in politics. We first explore the possibility that the relationship between minority status and feeling welcome, observed in Figure 2, is mediated by socioeconomic status, demographic attributes, or political interest. We control for these additional covariates because – when citizens form expectations about whether or not they would feel welcome among other politicians – these additional factors, which tend to correlate with minority status, may also play a part. For example, it could be that it is not minority status per se but low socioeconomic status that makes minorities feel less welcome. Appendix Figure A.2 displays these results. Overall, they mirror those in Figure 2 in so far as all coefficient estimates are negative and statistically significant. For the Swedish case, the estimates in Figure A.2 are nearly identical to the ones in Figure 2, though these estimates are relatively smaller for the US survey. This is suggestive evidence that minority status exerts an independent effect in shaping expectations of discrimination.
Second, the 2021 Swedish survey asked whether respondents expected members of their own group to feel welcome in politics and whether members of other groups could be expected to feel welcome. This feature allows us to infer whether it is only minorities who think they are less welcome in politics or if there is broader agreement that norms of exclusion exist in politics. Appendix Table A.4 shows that individuals who grew up in Sweden agree with those who grew up in the rest of Europe and outside Europe about how to welcome new politicians from different groups are expected to feel (p > 0.10 for all group-wise comparisons). This is further suggestive evidence that the results in Figure 2 reflect minorities picking up on signals that the political system does not always welcome diversity, signals which are also evident to non-minorities.
Third, we analyze how expected discrimination shapes office-seeking ambitions. So far, our results have shown that minority citizens across these two country contexts expect discrimination if they were to enter politics. Do these perceptions affect office-seeking ambitions? Scholarship has shown that the decision to enter politics is likely to be shaped by voter discrimination (Shah Reference Shah2014).Footnote 12 We argue that expectations of discrimination may play an important and overlooked role in shaping minority descriptive representation.Footnote 13 Our data largely confirms this. Appendix Table A.3 reveals a strong negative relationship between expected discrimination and interest in running for office both in the United States and in the 2017 and 2021 Swedish surveys. In the US case, the probability of expressing an interest in running for office increases by 20 percentage points when a respondent expects to feel welcome among other politicians. The corresponding figures in the 2017 and 2021 Swedish surveys are 18 percentage points and 9 percentage points, respectively.Footnote 14
In addition, previous literature has shown that past experiences of discrimination may increase political participation (Besco et al. Reference Besco2022; Oskooii Reference Oskooii2020; Pantoja, Ramirez, and Segura Reference Pantoja, Ramirez and Segura2001). At first glance, our result that expectations of discrimination dampen interest in running for office may appear to run counter to previous findings. This contradiction is only apparent, however. Conceptually, it is likely that, in many cases, experiences of discrimination have the potential to mobilize those affected by it, but this depends on how welcoming they perceive the political system to be. Most importantly, we can show empirically that when we control for experiences of discrimination (Oskooii Reference Oskooii2020), which has a positive impact on interest in running for office, the positive relationship between feeling welcome and interest in running for office hardly changes (see Table A.6), suggesting the importance of ‘feeling welcome’ as an independent and important variable that, on its own, contributes to shaping minority representation.Footnote 15
Finally, we re-run our main analysis, taking into account the fact that only the most politically interested and engaged are likely to overcome the hurdles associated with running for – and winning – political office. Thus, most people never become a realistic part of the actual candidate pool. Therefore, Appendix Tables A.1 and A.2 interact race/ethnicity with political engagement, as measured by high levels of political interest and involvement in civil society. Overall, these analyses do not alter our conclusions and show that both among the more and the less realistic pools of candidates, minorities feel less welcome than members of the majority population.
Conclusion
Our study turns to the underrepresentation of minorities in politics. It examines a key consideration that potential candidates weigh when deciding whether to run for political office – expected discrimination. Our findings suggest that the underrepresentation of minority candidates might stem from a reluctance to contend in the first place. We illuminate that the perceived unwelcomeness in politics, grounded in anticipated discrimination, is a potent deterrent for potential minority candidates and shapes the ‘supply side’ of minority candidacies. The results demonstrate that minorities who comprise large and ever-growing segments of the population expect to face discrimination if they were ever to enter politics. We find that minorities in the United States and Sweden anticipate feeling less welcome in politics compared to majority citizens. These expectations persist even when controlling for potential correlates of minority status, such as education. We also find that expected discrimination has political repercussions: those who do not anticipate feeling welcome among other politicians are less interested in running for office, all else being equal. Our findings underscore these perceptions’ consequential role in discouraging minorities from pursuing political office, thereby challenging conventional wisdom that heightened awareness or experience of discrimination that invariably galvanizes political participation. The fact that we uncover similar dynamics across two very different country contexts speaks to the potential generalizability of our findings.
These findings challenge the legitimacy of the democratic process. They undermine the fundamental notion that every citizen seeking redress for societal injustices should have equal access to the political arena. Moreover, our results suggest a vicious cycle where minority underrepresentation signals that the political system does not welcome diversity, which dissuades some potential candidates from running for political office. Therefore, parties interested in attracting a diverse roster of candidates should highlight how they welcome minorities into politics.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0007123424000073.
Data availability statement
Replication materials for this article can be found in Harvard Dataverse at: https://doi.org/10.7910/DVN/WF51RQ.
Acknowledgements
We thank Matt Barreto and Patrik Öhberg for their valuable comments on the survey questions used in th|e US 2021 and Sweden 2021 surveys. We also thank Karl-Oskar Lindgren, Pär Nyman and Rafaela Dancygier for generously sharing the Sweden 2017 data they collected jointly with Vernby. Finally, the comments from the editor and three anonymous reviewers significantly improved this article.
Financial support
This work was supported by the Swedish Research Council (grant number 2018-01887).
Competing interests
None.
Ethical standards
This research includes the use of third-party survey data. The US 2021 results draw on the Collaborative Multiracial Post-election Survey (CMPS). The principal CMPS investigators acquired ethical approval before fielding the survey. The Swedish 2021 results are drawn from a survey conducted by the SOM Institute. The principal investigators of the SOM survey acquired ethical approval prior to fielding the survey. The Swedish 2017 data was collected for the research project ‘The Underrepresentation of Immigrants in Politics.’ This survey's principal investigators acquired ethical approval before fielding the survey.
Open access
