Turnout and social fractionalisation
| Author | Ignacio Lago,Marc Guinjoan,Sandra Bermúdez,Pablo Simón |
| Date | 01 May 2018 |
| Published date | 01 May 2018 |
| DOI | 10.1177/0263395716686598 |
| Subject Matter | Articles |
research-article2016
Article
Politics
2018, Vol. 38(2) 113 –132
Turnout and social
© The Author(s) 2017
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
fractionalisation
https://doi.org/10.1177/0263395716686598
DOI: 10.1177/0263395716686598
journals.sagepub.com/home/pol
Ignacio Lago and Sandra Bermúdez
Universitat Pompeu Fabra, Spain
Marc Guinjoan
Universitat Autònoma de Barcelona, Spain
Pablo Simón
Universidad Carlos III de Madrid, Spain
Abstract
Social fractionalisation has been omitted in most influential cross-sectional studies on turnout,
and when it has been included, evidence is, at best, mixed. This article addresses this gap from
two perspectives. First, using aggregated data from 22 countries we show that turnout is inversely
related to ethnolinguistic fractionalisation, even after controlling for institutional, political
and socioeconomic determinants. Second, we rely on data from elections in two multilingual
territories, Catalonia and Quebec, to examine both the direct and indirect causal mechanisms for
which voting and the sense of duty of vote are affected by the individuals’ aversion to the opposite
ethnicity and the relative size of ethnicities. Analyses show that those relatively more averse to
mixing with others who are different to themselves have a lower propensity to vote and are less
likely to construe voting as a civic duty when they belong to the minority group.
Keywords
civic duty, heterogeneity, social capital, social fractionalisation, turnout
Received: 19th February 2016; Revised version received: 18th November 2016; Accepted: 24th November 2016
Introduction
One of the most well-established propositions in the ample literature on quality of govern-
ment is that ‘more homogeneous communities foster greater levels of social-capital produc-
tion’ (Costa and Kahn, 2003: 103) and consequently ‘ethnolinguistically homogeneous
countries have better government than heterogeneous ones’ (La Porta et al., 1999: 265). Both
Corresponding author:
Ignacio Lago, Department of Political and Social Sciences, Universitat Pompeu Fabra, Ramon Trias Fargas
25-27, Barcelona 08005, Spain.
Email: ignacio.lago@upf.edu
114
Politics 38(2)
the argument and the empirical evidence are pervasive. More homogeneous communities
have a higher level of social interaction, leading to more social capital, and social capital
influences economic outcomes and public policies (Alesina et al., 1999; Alesina and La
Ferrara, 2000; Alesina and Zhuravskaya, 2011; Easterly and Levine, 1997; La Porta et al.,
1999). Costa and Kahn (2003: 104) summarise the evidence in the following way: ‘over the
past five years, at least 15 different empirical economic papers have studied the consequences
of community heterogeneity, and all these studies have the same punch line: heterogeneity
reduces civic engagement. In more-diverse communities, people participate less’.
The literature on economics, civil war or public goods clearly discerns between ethnic,
linguistic and religious fractionalisation to study its impact on different outcomes (Alesina
et al., 2003; Alesina and La Ferrara, 2000; Fearon and Laitin, 2003). Thus, despite the
traditional way of conceiving heterogeneity among individuals as purely one of income
inequality, a growing literature is focusing on social fragmentation. In our article, in line
with previous research, social fragmentation is understood as the probability that two
randomly drawn individuals belong to two different groups in terms of ethnicity, lan-
guage or religion.1
Surprisingly, social fragmentation has been omitted in most influential cross-sectional
studies on turnout. For instance, in the books by Blais (2000) and Franklin (2004) on the
determinants of turnout, the level of fractionalisation of the country is absent from the
empirical models. Similarly, when Alesina et al. (2003) or Alesina and Zhuravskaya
(2011) study the impact of fractionalisation and segregation on more than 20 dependent
variables capturing economic success and the quality of democratic institutions, turnout
is not considered. The empirical results of the few studies that do include social fragmen-
tation are mixed: Geys (2006: 653), in a review article of aggregated-level research on
turnout levels, revealed that there was ‘no relation […] between turnout and population
concentration and homogeneity’; however, in a recent update of this review, Cancela and
Geys (2016: 267) have suggested that the evidence ‘appear to be moving towards the
conclusion that political fragmentation in general has little direct, independent relation to
voter turnout’. Finally, Martinez i Coma (2016) has also recently found a negative impact
of ethnic diversity on turnout in democratic and parliamentary elections. The longstand-
ing lack of consensus on the issue may be due to the fact that these studies have not con-
sidered the mechanisms operating behind this relationship. Thus, the purpose of this
article is not only to revisit the relationship between social fractionalisation and turnout
but also to delve into the mechanisms that may explain this link.
To do so, we use an innovative research design that combines aggregate and individ-
ual-level analysis. Our contribution is twofold. First, we begin by discussing why frac-
tionalisation might cause turnout to be lower. We test our arguments empirically by using
Franklin’s (2004) models, which account for the level of turnout in lower chamber elec-
tions in 22 countries from 1945 to 1999, but we add ethnic, linguistic and religious frag-
mentation to the specifications. The results indicate that turnout is negatively correlated
with ethnic and linguistic fragmentation, but not with religious diversity. Second, as this
aggregate association in not enough to pin down the origin of this finding, we then use
individual data from elections in Catalonia (Spain) and Quebec (Canada), two multilin-
gual territories, to examine to what extent social fractionalisation affects voting and the
sense of duty to vote. By changing the majority ethnic group across regional and national
elections, we show that those individuals who are relatively more averse to mixing with
others different to themselves have a lower propensity to vote and are less likely to con-
strue voting as a civic duty when they belong to the minority group.
Lago et al.
115
The article continues as follows. In the following section, we discuss the arguments for
which fractionalisation should cause turnout to be lower. The third section describes the
methods for the aggregated and individual-level data, as well as it presents the results of
the two empirical analyses. The fourth and last section presents our conclusions.
Arguments
The determinants of turnout have been extensively studied from different approaches
(Blais, 2000; Franklin, 2004; Herrera et al., 2015; Jackman, 1987; Powell, 1986). Scholars
have devoted their attention to understanding, among other aspects, the impact of differ-
ent socioeconomic, institutional and political factors (Blais and Dobrzynska, 1998); the
effect of rules such as the electoral system (Blais and Aarts, 2006; Blais and Carty, 1990;
Cox, 2015); or how findings may apply to non-Western democracies (Kostadinova,
2003). Among the studies centred on the social composition of societies, social fraction-
alisation – mostly ethnic and linguistic – has been considered as a potential determinant,
although the empirical evidence is far from conclusive (Geys, 2006: 645).
Several authors have argued that the relationship between turnout and social fraction-
alisation is positive. For instance, according to Anduiza (1999: Chapter 1), mobilisation
and then turnout increase with cleavage anchoring and cultural segmentation, given that
social heterogeneity fosters party–voter linkages. In a similar vein, Oliver (1999: 191)
claims that:
local politics should be more contentious (and participation highest) in economically diverse
cities as these places have more groups pursuing contradictory goals. Conversely, a greater
consensus over local policy in economically homogeneous cities should limit competition,
citizen interest, and participation. (Oliver, 1999: 191)
However, the empirical evidence is, at best, mixed (Anduiza, 1999: Chapter 4).
Other studies have stressed exactly the opposite effect, expecting social fractionalisation
to cause turnout to dip. As Cohen (1982: 259) summarises, ‘planners who are concerned
with building communities […] have accepted the fact that social homogeneity is a neces-
sary prerequisite of community cohesion’. As cohesion increases group solidarity, social
pressure and information, political participation in communities with a high degree of social
homogeneity should be higher than in areas where it is not the case (Geys, 2006: 644–645;
see also Fiorina, 1990; Gerber et al., 2008; Oliver, 2000; Yamamura, 2011). The negative
impact of heterogeneity on turnout could be exacerbated by political mobilisation. In areas
of racial diversity, for instance, Whites might increase the intentional efforts to demobilise
Blacks by adopting more restrictive procedures relating to voter turnout, such as increased
stringency of voter registration requirements (Hill and Leighley, 1999: 280).
Finally, for Campbell (2006), there is a U-shaped relationship between heterogeneity
and turnout. In his work, both in the communities characterised by...
To continue reading
Request your trial