Representation Despite Discrimination

• DOI: 10.1177/1065912913511863
• Date: 01 June 2014
• Author: Alex Street
• Published date: 01 June 2014
• Subject Matter: Articles

511863PRQXXX10.1177/1065912913511863StreetStreet
research-article2013
Article
Political Research Quarterly
2014, Vol. 67(2) 374 –385
Representation Despite Discrimination:
© 2013 University of Utah
Reprints and permissions:
sagepub.com/journalsPermissions.nav
Minority Candidates in Germany
DOI: 10.1177/1065912913511863
prq.sagepub.com
Alex Street1
Abstract
Immigrant-origin minorities are underrepresented in many democratic legislatures. This paper evaluates the direct
effects of voter discrimination on the electoral performance of minority political candidates in Germany. Using evidence
from both a survey experiment and actual election data, the paper tests two mechanisms of discrimination—negative
attitudes toward minority groups and assumptions about candidate ideology—and shows that neither results in a
substantial penalty for the small numbers of minority candidates who actually compete for office. Minority candidates
in Germany typically run for political parties that discriminating voters would not have supported in any case.
Keywords
minority representation, discrimination, elections
There is now extensive evidence on the breadth and depth
elections in 2005 and 2009 provides little evidence that
of negative attitudes toward immigrant-origin minorities
immigrant-origin candidates incur a penalty at the polls.
in Western Europe (e.g., Adida, Laitin, and Valfort 2010;
Bringing together the two kinds of data calls attention to
Hainmueller and Hangartner 2013; Sides and Citrin
the ways in which the effects of candidate identity may be
2007). Scholars have debated the sources of these atti-
mediated by voter characteristics and the context of party
tudes, in particular the importance of perceived economic
competition. Substantively, the findings help to explain
or cultural threats (e.g., Dancygier and Donnelly 2013;
why some members of stigmatized minority groups win
Sniderman, Hagendoorn, and Prior 2004). Much less is
representation, even in contexts with high levels of voter
known, however, about the ways in which these attitudes
prejudice. The results suggest a more nuanced view of the
influence political outcomes such as elections and public
political effects of discrimination than might be inferred
policy.
from the alarming evidence of public hostility toward
This paper examines how voter discrimination affects
immigrant-origin minorities in Western Europe.
the electoral performance of immigrant-origin political
Discrimination serves as a constraint but does not entirely
candidates. Evidence from a survey experiment shows
determine the electoral prospects of political candidates
that German voters, particularly those who feel threatened
from minority groups.
by immigrants and Muslims, are less supportive of identi-
The next section of the paper reviews theories and evi-
cal candidates with Turkish rather than German names.
dence on candidate identity and voter discrimination. I
However, discrimination is observed mostly among sup-
then describe the research site, data, and methods. Findings
porters of parties of the Right. In Germany, the small num-
from the survey experiment and from the analysis of elec-
bers of minority group members who stand for office tend
tion data are presented in separate sections. I discuss the
to run on the Left. The fact that discriminatory voters are
implications of the findings for future research, drawing
found mainly on the Right, and minority candidates on the
on secondary sources to consider how voter discrimina-
Left, limits the direct effects of discrimination.
tion may indirectly constrain the representation of minor-
Survey experiments have clear advantages for identi-
ity groups. The final section concludes.
fying causal effects, but doubts are often raised over the
external validity of the findings (Gaines, Kuklinski, and
1Max Planck Institute for the Study of Religious and Ethnic Diversity,
Quirk 2007). This paper provides an example of how evi-
Göttingen, Germany
dence from a survey experiment can be combined with
Corresponding Author:
observational data—an approach that is remarkably rare
Alex Street, Max Planck Institute for the Study of Religious and Ethnic
in the existing literature (but see Philpot and Walton
Diversity, Hermann-Föge-Weg 11, Göttingen, 37073, Germany.
2007). A parallel analysis of vote shares from German
Email: street.alex@gmail.com

---

Street
375
Candidate Identity and Voter
ideology are not inherently negative, they can still lead to
Discrimination
discrimination. Stereotypes about ideology may be inac-
curate in the aggregate and should not be assumed to
Members of ethnic, racial, and religious minorities, as
apply in any given case (Dolan 2010).
well as women, are underrepresented in many democratic
Although conceptually distinct, in fact these two
legislatures, compared to the size of these groups in the
mechanisms may interact. For example, Berinsky and
relevant populations (Bird, Saalfeld, and Wüst 2011;
Mendelberg (2005) find that a socially acceptable stereo-
Bloemraad and Schönwälder 2013; Krook and O’Brien
type (“Jews are liberal”) is linked in voters’ minds to an
2010). A number of explanations have been proposed to
unacceptable stereotype (“Jews are shady”). References
account for this fact, including voter discrimination, bar-
to the unacceptable stereotype—even if voters reject it—
riers within political parties, lack of resources for poten-
can affect voter intentions by activating the assumption
tial candidates, and the incentives established by voting
that Jewish candidates are liberal.
rules (Lawless and Fox 2010; Norris 2004; Sigelman
Much of the research on the psychological mecha-
et al. 1995). In this paper, I focus on the impact of voter
nisms of voter discrimination relies on experiments in
discrimination, defined as differential treatment of other-
which subjects are presented with information on hypo-
wise similar individuals depending on the social catego-
thetical candidates. This approach allows the manipula-
ries with which they are associated.
tion of a small number of candidate attributes—such as
Existing research has revealed two main psychologi-
ethnicity or gender—in a controlled environment. But the
cal mechanisms of discrimination against women and
artificial context also raises concerns over external valid-
minority candidates. The first of the proposed mecha-
ity. We might expect different responses to similar treat-
nisms is that voter attitudes toward particular candidates
ments in the real world, where more is at stake, where
are influenced by broader negative views of certain social
exposure to the treatment of interest may be repeated, and
groups. For example, Terkildsen (1993) finds that white
where many sources of information compete for voters’
voters with negative views of African Americans evalu-
limited attention (Gaines, Kuklinski, and Quirk 2007). In
ate African American candidates less favorably than sim-
addition, there is a particular reason to suspect that the
ilar white candidates. Greenwald et al. (2009) report that
two mechanisms described above will have limited direct
implicit measures of preference for white over black
effects on election outcomes. The key outcome is the
Americans predicted candidate choice in the 2008 elec-
behavior of potential supporters of minority candidates.
tion. Benson, Merolla, and Geer (2011) find evidence of
But candidate identity is not the only factor that deter-
bias against political candidates belonging to religious
mines the potential for support. Attitudes toward minority
minorities in the United States, though not against female
groups and ideological preferences vary significantly
or African American candidates. In all these studies, oth-
across the supporters of different political parties.
erwise identical candidates are less likely to receive voter
In Germany, as in many European countries, negative
support when negative group attitudes are activated.
views of immigrant-origin minority groups are more
The other commonly proposed psychological explana-
common among those on the Right of the political spec-
tion for discrimination against women and minority can-
trum (Alba, Schmidt, and Wasmer 2003; Hainmueller
didates is that voters use identity cues to infer the
and Hiscox 2009). Most prior research on minority candi-
ideological position of the candidate. Voters may penalize
dates has been conducted in the USA, where two parties
candidates who are assumed to be further from their own
compete for the support of the median voter, but in other
ideological positions. McDermott (1997) argues that,
electoral systems it may be easier for minority candidates
even accounting for partisanship, women are expected to
to avoid relying on potentially hostile voters. Under pro-
be more liberal than men and finds that female Democratic
portional representation, small parties can appeal to nar-
candidates fare better among liberal voters and worse
row segments of the electorate while still hoping to win
among conservatives than their male peers (see also
seats and even a place in a coalition government. If nega-
Huddy and Terkildsen 1993). Washington (2006, 975)
tive attitudes are more common on the Right but minority
suggests the belief that black Democratic candidates are
candidates run mainly on the Left, the set of potential
“far more liberal than their...