Can information campaigns impact preferences toward vote selling? Theory and evidence from Kenya
| Date | 01 June 2020 |
| DOI | 10.1177/0192512119836283 |
| Author | Aaron Erlich |
| Published date | 01 June 2020 |
| Subject Matter | Articles |
https://doi.org/10.1177/0192512119836283
International Political Science Review
2020, Vol. 41(3) 419 –435
© The Author(s) 2019
Article reuse guidelines:
sagepub.com/journals-permissions
DOI: 10.1177/0192512119836283
journals.sagepub.com/home/ips
Can information campaigns
impact preferences toward
vote selling? Theory and
evidence from Kenya
Aaron Erlich
McGill University, Canada
Abstract
What factors shape citizens’ willingness to engage in vote selling? This paper argues that providing voters with
information about the detrimental effect of vote selling (public service predation) or telling them that their
community members will look down on them if they engage in the practice (social sanctioning) can shape
vote-selling attitudes in emerging democracies. Using a nationwide randomized survey experiment carried out
between May and June of 2012 in Kenya, this study primes voters with theory-based informational messages
for voters to test whether such messages can potentially curtail vote-selling attitudes. The paper finds that
both public service predation and social sanctioning messages can reduce stated vote-selling preferences
as much as legal campaigns that have been tested previously. The study has important implications for
researchers and policy-makers because it suggests alternative methods to change vote-selling attitudes and
even behavior in the short- to medium-term.
Keywords
Kenya, vote selling, survey experiment, political messaging, information treatment, democratization
Introduction
Beyond the structural determinants of vote selling, it is theoretically important to examine citizens’
attitudes towards vote selling and to test whether they can be changed through targeted messaging.
Changing attitudes is important because these attitudes are likely linked to vote-selling behavior
and shape vote-trading practices. From a policy perspective, testing the impacts of messaging on
attitudinal change towards vote selling is critical because other solutions to curbing vote selling,
such as promoting economic development at the individual and aggregate levels and strengthening
the rule of law, are difficult, long-term endeavors.
Corresponding author:
Aaron Erlich, Department of Political Science, McGill University, 855 rue Sherbrooke Ouest, Montréal, Québec H3A
2T7, Canada.
Email: aaron.erlich@mcgill.ca
836283IPS0010.1177/0192512119836283International Political Science ReviewErlich
research-article2019
Article
420 International Political Science Review 41(3)
Vote selling—the voter side of vote trading—propagates a vicious cycle of bad governance: It
limits public goods provision (Vicente, 2014); reinforces ethnic-based voting (Wantchekon, 2003);
and decreases democratic legitimacy (Ziblatt, 2009). While structural factors—poverty (Mares,
2015), social networks (Stokes et al., 2013), and institutional design (Kam, 2017)—all influence
vote trading, recent research casts doubt on the assumption that these are the only drivers (e.g.
Corstange, 2012). Cross-national evidence demonstrates that vote trading varies significantly
across countries at the same level of economic development and similar formal institutions (e.g.
Jensen and Justesen, 2014).
Another source of variation in vote-selling behavior is attitudinal. If vote-selling attitudes are
linked to vote-selling behaviors, changing these attitudes can improve governance; to date,
research examining short-term changes in vote-selling behavior, beliefs, and attitudes has
focused on reminding citizens that vote trading is illegal (Vicente, 2014). However, I argue that
legal information campaigns might not be the most efficacious type of campaign in resource
constrained environments because voters in politically fragile environments know that there is
no enforcement mechanism.
I hypothesize that, in emerging democracies, individuals’ attitudes towards vote selling are
shaped by access to information about its potential consequences and by local norms, including the
social acceptability of opportunistic behavior. Access to information and social norms can be lever-
aged to change voter attitudes, this study hypothesizes, through informational interventions tai-
lored to the problem. This study focuses attention on the sociological aspects of electoral attitudes
that the literature on clientelism has hitherto neglected, and argues that providing citizens with
information on the negative consequences they will suffer in their day-to-day lives and about social
perceptions of vote selling will change their attitudes towards vote selling.
To test the argument, I develop and deploy three types of information campaigns with distinct
mechanisms. Campaign 1 tells respondents specifically how vote selling is detrimental to public
service provision because it siphons money from public coffers (Public service predation treat-
ment (PSPT)); Campaign 2 reinforces the idea that vote selling is looked down upon by respond-
ents’ communities (Social sanctioning treatment (SST)); and Campaign 3 stresses that vote selling
is illegal (Legal treatment (LT)).
These three messages are embedded within a nationally representative survey of Kenya. Kenya
provides a good place to test such messaging because there are high levels of vote selling, and the
process is well understood. Moreover, genuine competition between candidates exists in Kenya,
and ballots are generally secret. Within the survey in Kenya, respondents received one of the three
randomized messages. The outcome variable—respondents’ attitude toward vote selling—was
then measured using an index variable that aggregates three post-treatment questions.
To preview the findings, the analysis reveals that treatment messages that inform voters about
how vote selling is detrimental to public services or reminds citizens that their neighbors may look
down on their vote selling appear to be as effective as, if not more than, an LT. The PSPT appears
to work particularly well with urban and educated voters. While the LT decreases vote-selling
preferences, the findings are less robust than for the other two types of message.
The implications for policy suggested by these findings diverge from current practice among
both researchers and development practitioners. This paper suggests that finding methods to per-
suade voters that it is not in their interest to engage in vote selling or creating an environment where
they believe they will be socially sanctioned for this activity, may change attitudes and effectively
attenuate vote selling in the short term. This paper’s findings imply that there is room for optimism
about the efficacy of applying interventions at the micro level to mitigate citizens’ lack of considera-
tion of the effects of clientelism and to leverage social sanctioning instead of relying on long run
processes at the macro level to gradually reduce vote trading in developing democracies.
To continue reading
Request your trial