Emerging Trends in The Social and Behavioral Sciences

Corruption and Electoral Accountability: Avenues for Future Research

Item

Title

Corruption and Electoral Accountability: Avenues for Future Research

Author

De Vries, Catherine E.

Solaz, Hector

Research Area

Social Institutions

Topic

Government Systems

Abstract

The vast majority of people across the globe lives in countries characterized by high levels of corruption, commonly defined as the public misuse of private gains. Although the exact costs of corruption are difficult to estimate, research suggests that corruption is bad for economic and social development. Not only has corruption been shown to have a detrimental effect on a country's economy, the ability to generate tax revenue and social equality, the political effects of corruption are also considerable. Owing to its association with weak state capacity, corruption may damage the ability of governments to craft and implement policies in areas in which continued intervention and investment is needed and their capacity to respond quickly and effectively to sudden shocks. Owing to these undesirable outcomes, elections are supposed to curb corruption because voters will throw the rascals out. Recent research, however, suggests that more often than not corruption goes unpunished at the ballot box. This essay sets out possible research avenues to find out why this is the case.

Related Essays

Economic Models of Voting (Political Science), Ian G. Anson and Timothy Hellwig

Electoral Authoritarianism (Political Science), Andreas Schedler

Presidential Power (Political Science), William G. Howell

Political Inequality (Sociology), Jeff Manza

Money in Politics (Political Science), Jeffrey Milyo

Information Politics in Dictatorships (Political Science), Jeremy L. Wallace

Lawmaking (Political Science), Jamie L. Carson and Mark E. Owens

News Framing Effects and Emotions (Political Science), Andreas R. T. Schuck and Alina Feinholdt

Political Advertising (Political Science), Erika Franklin Fowler

The Politics of Disaster Relief (Political Science), Alexander J. Oliver and Andrew Reeves

Racial Disenfranchisement (Political Science), Vincent L. Hutchings and Davin L. Phoenix

Perceptions of the Legitimacy of Collective Decisions (Political Science), John W. Patty

Partisan News: A Perspective from Economics (Economics), Daniel F. Stone

Identifier

etrds0444

extracted text

k
PSYHB etrds0444.tex

V2 - 10/22/2018

6:14 P.M.

Page 1

Corruption and Electoral
Accountability: Avenues for Future
Research
CATHERINE E. DE VRIES and HECTOR SOLAZ

Abstract

k

The vast majority of people across the globe lives in countries characterized by
high levels of corruption, commonly defined as the public misuse of private gains.
Although the exact costs of corruption are difficult to estimate, research suggests
that corruption is bad for economic and social development. Not only has corruption
been shown to have a detrimental effect on a country’s economy, the ability to
generate tax revenue and social equality, the political effects of corruption are also
considerable. Owing to its association with weak state capacity, corruption may
damage the ability of governments to craft and implement policies in areas in which
continued intervention and investment is needed and their capacity to respond
quickly and effectively to sudden shocks. Owing to these undesirable outcomes,
elections are supposed to curb corruption because voters will throw the rascals
out. Recent research, however, suggests that more often than not corruption goes
unpunished at the ballot box. This essay sets out possible research avenues to find
out why this is the case.

k

INTRODUCTION
Many academics, policy makers, and pundits assume that fair and periodic
elections will curb political malfeasance because of voters’ disdain of
corruption. Much of the recent empirical literature, however, contradicts this
view. On average, the effective electoral sanctioning of corruption is rare (De
Vries & Solaz, 2017; Fisman & Golden, 2017). Even if there is a clear incentive
for a society at large to root out corrupt practices, as this would increase
societal welfare, political elites, and voters, especially in high-corruption
contexts, face enormous coordination challenges that favor the status quo
(Klašnja, Little, & Tucker, 2016). An important question then arises: why
voters reelect politicians who steal from them?
Emerging Trends in the Social and Behavioral Sciences.
Robert A. Scott and Marlis Buchmann (General Editors) with Stephen Kosslyn (Consulting Editor).
© 2018 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.

1

k

k
PSYHB etrds0444.tex

2

k

V2 - 10/22/2018

6:14 P.M.

Page 2

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

In previous work, we have highlighted three stages that are important
in voters’ decision-making about how to deal with candidate corruption:
an information acquisition, causal attribution, and behavioral response stage
(De Vries & Solaz, 2017). Information acquisition refers to the ways in
which voters learn about corruption. Causal attribution is the process by
which voters assign blame to a politician (or bureaucrat) for corruption.
Finally, a behavioral response denotes that voters choose to act on their
corruption evaluations and blame attribution by throwing the rascals out.
A breakdown of any one or more of these stages can result in corruption
not being punished at the ballot box. Finding out how each of these stages
matter is important for the design of behavioral or institutional solutions to
increase accountability for corruption and by doing so increasing societal
welfare.
Much of the existing literature deals with the first and last stage. When
it comes to information acquisition, empirical work demonstrates that a
lack of punishment of corruption is the result of the quantity and quality of
information that voters receive. Information dissemination could be imperfect due to a lack of media reporting, political control of the media, high
levels of illiteracy, or the lack of credible sources (Botero, Cornejo, Gamboa,
Pavão, & Nickerson, 2015; Ferraz & Finan, 2008; Winters & Weitz-Shapiro,
2013). When it comes to the behavioral response stage, a wealth of evidence
suggests that voters trade-off corruption information against other benefits
when they decide how to cast their ballot. These benefits could be based on
side-payments (Fernández-Vázquez et al., 2016), performance in other areas,
or policy pledges (Klašnja & Tucker, 2013; Rundquist, Strom, & Peters, 1977).
Hence, voters may tolerate corrupt practices, or even approve of them,
because they value other things more.
The way in which causal attribution matters has received far less scholarly attention. Could it be the case that voters simply might not update their
beliefs about candidate quality even when presented with concrete evidence
of malfeasance? Some recent field and laboratory experimental evidence suggests that this is indeed the case. So far, this is based on only a handful of
studies and we do not know if these findings are generalizable. Moreover, to
date, we do not yet have a comprehensive understanding of the factors that
explain why voters fail to update their beliefs about candidate quality even
if they are confronted with unambiguous corruption information. Against
this backdrop, we suggest that the main challenge for future research is to
develop a coherent understanding of how causal attribution matters for the
electoral punishment of corruption.
In the ensuing sections, we first elaborate the three stages that are important in voters’ decision-making about candidate corruption and discuss existing work in terms of the stages they examine. Second, we highlight several

k

k

k
PSYHB etrds0444.tex

V2 - 10/22/2018

Corruption and Electoral Accountability: Avenues for Future Research

6:14 P.M.

Page 3

3

important avenues for future research regarding the role of causal attribution.
Finally, we conclude by highlighting why these avenues of future research
are so important for our understanding of corruption more broadly.
SHORT OVERVIEW OF EXISTING RESEARCH: THE IMPORTANCE
OF INFORMATION AND OTHER BENEFITS

k

The idea that voters hold politicians to account for corruption fits classical
models of retrospective voting. These models maintain that voters use elections to sanction low-performing politicians by selecting high-performing
ones (Besley, 2006). In order to win elections, candidates need to convince
voters that they are competent and thus able to promote and secure their
well-being. Voters form beliefs about candidate competence by paying attention to quality signals, such as developments in their living standards, the
current state of the economy or corruption scandals among other things (Fiorina, 1981). When voters indeed cast their ballots based on these beliefs about
candidate quality, we can expect that periodic and competitive elections help
to ensure that high-quality candidates get elected and are being reelected.
Information about candidate corruption is important, because it acts as a signal for voters that candidates are of low quality. As a result, we would expect
corrupt politicians to be punished at the ballot box, and the extent to which
this happens to vary by the quality of the signal. However persuasive this theoretical account may be, the real world turns out to be more complex. While
some studies find that corrupt activities indeed take a considerable electoral
toll on incumbents, especially when the information signal is clear or people’s
ability to distill it is high (Ferraz & Finan, 2008; Klašnja, 2016; Weitz-Shapiro &
Winters, 2017), other work suggests that voters might reelect politicians they
know to be corrupt even when confronted with strong evidence of malfeasance, because they prefer other benefits or due to group loyalty (Anduiza,
Gallego, & Munoz, 2013; Rundquist et al., 1977; Solaz, De Vries, & De Geus,
2018).
These conflicting findings might not be entirely surprising when we
consider that several conditions should be met in order for retrospective
voting to exist (Healy & Malhotra, 2013). In our previous work, we outlined three stages that are important for holding candidates accountable
for political malfeasance, an information acquisition, causal attribution, and
behavioral response stage. Information acquisition relates to voters needing
to be aware of corruption. In the first stage of information acquisition, it
follows that in order for voters to be able to punish corruption, they need
to know about it. Becoming aware of corruption is not always easy, because
elected public officials engaged in corrupt activities have an incentive to
hide information, and may try to influence the media or the judiciary not to

k

k

k
PSYHB etrds0444.tex

4

k

V2 - 10/22/2018

6:14 P.M.

Page 4

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

report on or prosecute their malfeasance. The literature suggests that people
learn about corruption either through personal experience or through
indirect perception by ways of media reporting for example (Klašnja,
Tucker, & Deegan-Krause, 2016). The second stage of causal attribution
refers to the fact that when voters are informed about candidate corruption,
they need to attribute responsibility for it and adjust their evaluations of
the candidate quality accordingly. Research has demonstrated that blame
attribution is far from straightforward, and often plagued by group-serving
biases. Whereas voters tend to attribute positive events or outcomes to their
preferred in-group, they absolve the in-group of blame for negative events
or outcomes (Anduiza et al., 2013; Solaz et al., 2018). The last stage refers
to behavioral responses. These blame attributions should also lead voters
to adjust their voting decisions by either voting for the opposition or by
refraining to cast a ballot in the first place. Voters have to consider a large
array of factors when deciding for whom to cast their ballot, and corruption
may simply not be a salient issue in a specific election campaign (Chang,
Golden, & Hill, 2010; Klašnja, Tucker, et al., 2016). Moreover, even if voters
find corruption important, they might be willing to trade it off against some
other benefit they care about more (Rundquist et al., 1977).
The majority of work to date has focused on the information acquisition
and behavioral response stage. When it comes to the former, studies suggest that the lack of electoral punishment of corruption is largely a function of the quantity of information available to voters (Chang et al., 2010;
Ferraz & Finan, 2008), and the quality of that information (Botero et al., 2015;
Winters & Weitz-Shapiro, 2013). A seminal study on the role of information conducted by Ferraz and Finan (2008) studies the effects of publicly
released audit reports about corruption practices in Brazilian municipalities.
The authors show that information dissemination increases the electoral punishment of corruption, and that these effects were most pronounced in those
municipalities where locally based radio stations disclosed this information.
Recent work in a similar vein suggests that not only the quantity of information matters but also the perceived quality of the source of information
(Botero et al., 2015; Winters & Weitz-Shapiro, 2013). The source deemed most
credible by voters may differ greatly from context to context. For instance,
a study from Brazil shows that respondents were more likely to punish a
candidate for corruption when the information stems from a federal audit
rather than from the opposition party (Winters & Weitz-Shapiro, 2013), while
a study conducted in Columbia found that corruption allegations made by
the leading national newspaper rather than a nongovernmental organization
or the judiciary engenders the strongest effects on punishment (Botero et al.,
2015).

k

k

k
PSYHB etrds0444.tex

V2 - 10/22/2018

Corruption and Electoral Accountability: Avenues for Future Research

6:14 P.M.

Page 5

5

Another set of studies suggests that corruption information might not lead
voters to punish malfeasance, because they care about other benefits more.
One of the first empirical papers on the topic by Rundquist and colleagues
featuring a set of survey experiments suggested that voters often engage in
a trade. Although they may generally disapprove of corrupt conduct, they
need to weigh their corruption evaluations against policy benefits, and this
may result in an overall positive balance in favor of reelection. As a result,
they might prefer a corrupt candidate who represents their policy positions to
a clean candidate who does not. Voters may trade off corruption against some
material gains (Fernández-Vázquez, Barberá, & Rivero, 2016), or turn a blind
eye in favor of other benefits such as ideological proximity (Rundquist et al.,
1977) or strong economic performance (Klašnja & Tucker, 2013). Overall, this
evidence suggests that because voters care about a multitude of things when
deciding to cast their ballot, oftentimes corruption becomes a secondary concern compared to other things like the state of the economy or a candidate’s
policy positions. If voters fail to punish corruption at the ballot box due to
instrumental calculations, such behavior might not be irrational, but still normatively undesirable, as it does not incentivize politicians to be clean.

k

AVENUES FOR FUTURE WORK: TOWARD A BETTER
UNDERSTANDING OF THE ROLE OF CAUSAL ATTRIBUTION
AND PRIOR BELIEFS
Notwithstanding the compelling evidence outlining the important role that
both information acquisition and voters’ decisions to trade off corruption
against other benefits play, to date we know much less about the intermediate stage of causal attribution. This is unfortunate, as we know that people
do not process information objectively. Social psychologists, for example,
have shown that people tend to make internal attributions (to their own
in-group) for positive events or outcomes and make external attributions
(to an out-group) for negative events or outcomes, something coined the
“ultimate attribution error” (Pettrigrew, 1979). While studies on retrospective
voting in political science have over the years paid considerable attention
to causal attribution (for an overview see Healy & Malhotra, 2013), to our
knowledge no study to date has explicitly examined causal attributions for
corruption as a dependent variable. In the following sections, we suggest
that understanding (i) how people attribute blame when exposed to information about corruption, and (ii) which factors hamper blame attribution
are two important areas that deserve more scholarly attention.
As highlighted in the previous section, existing work centers on the study
of how voters get informed about candidate corruption, and how they trade
this information off against other benefits they may care about. Yet, could

k

k

k
PSYHB etrds0444.tex

6

k

V2 - 10/22/2018

6:14 P.M.

Page 6

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

it be the case that voters simply do not update their beliefs about candidate
quality even when presented with unequivocal evidence of malfeasance?
When group loyalties, based on partisanship, religion, ethnicity, and so on,
are strong, corruption information of in-group members might not have
any effect on voter’s decision-making. Owing to the fact that corruption
allegations may pose a serious threat to in-group status and the in-group’s
hold on power, people might absolve in-group members of blame. The
most straightforward channel through which this would work is through
people’s in-group serving bias (Solaz et al., 2018). While voters tend to
reward the in-group for positive outcomes, they absolve it of blame for
negative outcomes. This would constitute a form of motivated reasoning
in which people, because of their automatic affective ties to their in-group,
selectively process information. Another channel through which this may
work is through in-group status. Evidence from behavioral economics shows
that people may gain some utility from the utility of others, in particular
from members of their own in-group (Charness & Rabin, 2002). In his work
on redistribution, for example, Shayo (2009) shows that individuals derive
utility from having high-status members in their group. Because of this,
poorer groups might not support redistribution from the rich to the poor
even though it would benefit them personally. If we extend this insight to
corruption-generated wealth, voters could gain some utility from the fact
that a candidate from their in-group stands to gain from corruption even if
they would personally lose out. In this case, voters might actually reward
candidates from their own group for being corrupt.
In a recent study, we started to examine these possible channels by exploiting a real-world corruption scandal in Spain that fell in the middle of a survey
fieldwork period and complementing it with laboratory experiments from
the United Kingdom (Solaz et al., 2018). By comparing Spanish respondents
before and after a corruption scandal was uncovered in a national newspaper,
we show that in-group loyalty based on partisanship decreased the likelihood of withdrawing support for the ruling party due to corruption, while
out-group identification increased it. In a next step, we replicated our results
in a laboratory setting that allowed us to isolate in-group effects more carefully by controlling the precise setting and information voters they received
as well as randomizing the group identity of the candidates that voters faced.
The laboratory results support our findings from the Spanish case: voters fail
to punish in-group candidates they know to be corrupt but punish out-group
candidates for the same indiscretions. These findings are important and lead
to a host of new questions. Which factors could condition the magnitude
of the effects of in-group serving biases? And what kind of group identities
might trigger these kinds of in-group effects, and which ones do not? These
are crucially important questions for future work to address.

k

k

k
PSYHB etrds0444.tex

V2 - 10/22/2018

Corruption and Electoral Accountability: Avenues for Future Research

k

Page 7

7

Another important avenue of future research is the importance of nature of
prior beliefs for understanding the degree to which voters, in fact, update
their evaluations of candidate quality in light of malfeasance revelations.
A recent study by Arias, Balan, Larreguy, Marshall, and Querubin (2017)
suggests that electoral punishment of corruption is rare when voters already
believe that candidates are of low quality. Owing to their negative priors, a
signal that the incumbent is corrupt will not be very informative. In fact, their
evidence stemming from a field experiment conducted in Mexico, where voters were informed about malfeasant municipal spending or not, suggests that
when priors were already negative, support for the incumbent party actually
increased after corruption was revealed because voters positively updated
their priors from information that was better than expected. These findings
suggest that remedying information asymmetries may not always lead to the
results that scholars or practitioners hope for. The evidence also leads to a
host of new questions, for example, which prior beliefs matter, those relating
to the incumbent as such or those relating to the prevalence of corruption
in society at large? And under which conditions do these different types of
priors matter most?
Concerning the latter, a study by Klašnja and Tucker (2013), comparing survey experimental evidence from Sweden and Moldova, for example, shows
that voters in Sweden, a low-corruption country, responded more to information about corruption than voters in Moldova, a high-corruption country,
do. What is more, voters in Moldova only responded to corruption information when economic conditions were poor. This perhaps suggests that while
on average due to their priors about the prevalence of corruption, voters in
high-corruption countries may not respond to one additional case of corruption, some heterogeneity in reactions most likely exists. This heterogeneity
could be the result of the overall state of the economy as the study by Klašnja
and Tucker suggests or individuals’ prior beliefs about candidate quality as
highlighted by Arias et al. (2017), but it could also be due to other prior beliefs
people might hold.
We think that people’s social priors might be especially important. Social
priors refer to people’s beliefs about society at large. The role of three sets
of social priors seem especially important in our view, people’s beliefs about
(i) the prevalence of corruption within society, (ii) the way other voters faced with
corruption allegations will act and (iii) the quality of the pool of future political
entrants. These social priors are likely going to condition the extent to which
voters will actually update their beliefs about candidate quality and blame
the incumbent for corrupt activities. The pervasiveness of these prior beliefs,
especially in high-corruption countries, may make it less likely that candidate
corruption will be punished at the ballot box. Let us address each of them
in turn.

k

6:14 P.M.

k

k
PSYHB etrds0444.tex

8

k

V2 - 10/22/2018

6:14 P.M.

Page 8

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

When people already think that corruption is widespread, they do not
punish corruption at the ballot box. This could be due to different reasons.
It might be due to the fact that when corruption is believed to be prevalent,
corruption allegations add little new information to voters’ decision-making
and therefore might carry very little weight in voters’ minds. This type
of reasoning is consistent with arguments that citizens residing in countries with higher quality of governance might display greater sensitivity to
accountability failures (Ashworth, Bueno de Mesquita, & Friedenberg, 2017).
It could also be the case that when corruption is more widespread, engaging
in corrupt practices might be deemed more acceptable by voters (Fisman
& Golden, 2017). Yet, it could also be that if voters think that corruption is
widespread, they also believe that it takes a corrupt politician to navigate the
system and to get things done (Bauhr & Charron, 2018). A recent study based
on conjoint experiments that was embedded in surveys in Argentina, Chile,
and Uruguay shows no effect of reminding people about the prevalence
of corruption in their country (Klašnja, Lupu, & Tucker, 2018). Yet, as the
authors themselves admit, empirically establishing the effect of people’s
perceptions about the pervasiveness of corruption is difficult, in part because
people’s understandings of what corruption might actually entail are likely
going to differ substantially between contexts in which corruption is more
or less widespread, and among individuals. Studying how people’s beliefs
about the prevalence of corruption matter is an important area for future
research.
Second, not only people’s perceptions about the extent of corruption in
society will likely affect their willingness to hold candidates to account
for corruption, but also their expectations about the behavior of others.
People’s prior beliefs about the behavior of other voters relate to the
importance of voter coordination in the punishment of corruption. When
people think that others are unlikely to punish existing politicians for
corruption, they are unlikely to do so as well (especially when there is
any form of side-benefit that they could lose by withdrawing support
from a corrupt incumbent). In order for voter coordination to occur, two
conditions need to be met: first, voters have to agree that corruption is
important for their vote, and second, they need to be able to agree to vote
for another clean candidate. A study by Chang et al. (2010) highlights the
importance of voter coordination. It is based on legislators in Italy’s lower
house between 1948 and 1994 and candidates of the two largest parties
(Christian Democracy and the Italian Communist Party). The authors show
that only in the 1992–1994 legislature did corrupt legislators face a serious
electoral penalty. They suggest that this structural break in the response to
corruption by Italian voters was due to the media coverage that grew out of
the Clean Hands operation in the early 1990s. The authors argue that a wider

k

k

k
PSYHB etrds0444.tex

V2 - 10/22/2018

Corruption and Electoral Accountability: Avenues for Future Research

k

Page 9

9

dissemination of information about corruption in the media not only makes
corruption more salient in voters’ minds but also provides a coordination
signal to voters, who realize that others will most likely punish corruption
as well.
Several recent studies examined the role of coordination signals through
different types of experimental treatments and found mixed results (Adida,
Gottlieb, Kramon, & McClendon, 2017; Arias, Larreguy, Marshall, &
Querubin, 2018; Yap, 2017). Citing evidence from a field experiment conducted in the context of legislative elections in Benin, Adida and colleagues
show that voters only punished bad performance when information was
accompanied with a message highlighting the importance of legislative performance for voter welfare and the fact that other villages in the constituency
received the same information. Laboratory evidence from Australia, Singapore, and the United States also highlights the importance of positive
coordination signals (Yap, 2017). Yet, evidence from a field experiment
revealing mayoral malfeasance before the 2015 Mexican municipal elections
suggests that accompanying leaflets detailing results of audit reports by
loudspeakers in order to signal to voters that others received the malfeasance
information as well did not significantly alter the effects of information on
incumbent vote shares (Arias et al., 2018). All in all, these results highlight
that understanding (i) which coordination signals matter and (ii) which
conditions may hamper their effects are two important areas for future
research.
Finally, voters’ beliefs about the quality of the pool of future political
entrants is likely important for their willingness to hold candidates to
account for corruption. When people think that the pool of possible political
entrants will also be corrupt, there is little reason to punish an existing
politician for corruption. In their theoretical work on corruption traps,
Klašnja, Little, et al. (2016) highlighted that overcoming corruption involves
coordination not only among voters and existing candidates but also among
possible entrants. Voters’ perceptions about the quality of the pool of
possible entrants will inform their behavior. When voters’ expect an alternative pool of candidates for office to be likely as corrupt as current ones,
why would they cast a ballot based on corruption? The pool of potential
candidates has been found to be important in other aspects of political
behavior as well, such as protest behavior (Meirowitz and Tucker, 2013).
When voters become informed about widespread corruption of incumbents,
this might influence their beliefs about the overall average quality of the
pool of potential governments, especially if they have little experience with
elections to begin with. While in established democracies, learning about
corruption may lead people to think that they should root out this one bad
apple, in newly established democracies this may not spark off pessimism

k

6:14 P.M.

k

k
PSYHB etrds0444.tex

10

V2 - 10/22/2018

6:14 P.M. Page 10

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

about all possible political candidates. Empirical examinations regarding
(i) the way in which people’s expectations about the quality of the pool of
possible alternatives affect the likelihood of punishing corruption, and (ii)
how these relationships differ based on people’s experience with democracy
constitute important avenues for future research.
CONCLUSION

k

In recent years, we have experienced somewhat of a boom in examining the
extent to which people hold politicians to account for corruption, not least
due to the widespread use of field, laboratory, and survey experiments. These
experimental tools help researchers to overcome some of the core problems
associated with studying corruption, such as the fact that politicians try to
hide malfeasance and people may not want to admit that they support corrupt politicians. Much of this recent work has centered on the importance of
corruption information and the willingness of voters, once they are informed
about corruption, to trade it off against other benefits. In this essay, we have
highlighted that the role of causal attribution has thus far received far less
scholarly attention. Based on a handful of recent studies, we suggested that
corruption might go unpunished at the ballot box simply because voters do
not update their beliefs about candidate quality in light of unambiguous
information about malfeasance. Recent work suggests that this is largely the
result of the weight voters attach to in-group status and their prior beliefs
about candidate quality. Yet, we do not know under which conditions these
factors are important, and if other beliefs matter as well. Moreover, most of
this early experimental work stems from one country context at one point in
time, so we do not know if we can generalize existing findings.
Overall, we have highlighted two key avenues for future research. First,
what is the role of role of attribution errors? To what extent do people fail to
blame their in-group candidate for corruption, and punish out-group candidates for the same indiscretion? Which factors may condition the magnitude
of these in-group effects? What is more, which kinds of group identities may
trigger these effects, and which ones do not? These are crucially important
questions for future work to address. Second, what is the role of people’s
beliefs about society at large? To what extent does the electoral punishment
of corruption hinge on people’s priors about the prevalence of corruption
within society, the way they expect other voters to act, and the quality of the
pool of future politicians? We believe that these social priors are crucial for
understanding the conditions under which corruption gets punished at the
ballot box. The extent to which they actually do, is an important topic for
future research.

k

k

k
PSYHB etrds0444.tex

V2 - 10/22/2018

Corruption and Electoral Accountability: Avenues for Future Research

6:14 P.M. Page 11

11

REFERENCES

k

Adida, C., Gottlieb, J., Kramon, E., & McClendon, G. (2017, November). Breaking the
clientelistic voting equilibrium: The joint importance of salience and coordination (Working Paper 48). Washington, DC: AIDDATA.
Anduiza, E., Gallego, A., & Munoz, J. (2013). Turning a blind eye: Experimental evidence of partisan bias in attitudes toward corruption. Comparative Political Studies,
46(12), 1664–1692.
Arias, E., Balan, P., Larreguy, H., Marshall, J., & Querubin, P. (2017). Priors rule: When
malfeasance revelations help and hurt incumbent parties? (Working Paper). Cambridge,
MA: Harvard University.
Arias, E., Larreguy, H., Marshall, J., & Querubin, P. (2018). Does the content and mode
of delivery of information matter for electoral accountability? Evidence from a field experiment in Mexico (Working Paper). Cambridge, MA: Harvard University.
Ashworth, S., Bueno de Mesquita, E., & Friedenberg, A. (2017). Accountability and
information in elections. American Economic Journal: Microeconomics, 9(2), 95–138.
Bauhr, M., & Charron, N. (2018). Insider or outsider? Grand corruption and electoral
accountability. Comparative Political Studies, 51(4), 415–446.
Besley, T. (2006). Principled agents? The political economy of good government. Oxford,
UK: Oxford University Press.
Botero, S., Cornejo, R. C., Gamboa, L., Pavão, N., & Nickerson, D. W. (2015). Says
who? An experiment on allegations of corruption and credibility of sources. Political Research Quarterly, 68, 493–504.
Chang, E. C., Golden, M. A., & Hill, S. J. (2010). Legislative malfeasance and political
accountability. World Politics, 62(2), 177–220.
Charness, G., & Rabin, M. (2002). Understanding social preferences with simple tests.
Quarterly Journal of Economics, 117(3), 817–869.
De Vries, C. E., & Solaz, H. (2017). The electoral consequences of corruption. Annual
Review of Political Science, 20, 391–408.
Fernández-Vázquez, P., Barberá, P., & Rivero, G. (2016). Rooting out corruption or
rooting for corruption? The heterogeneous electoral consequences of scandals.
Political Science Research and Methods, 4(2), 379–397.
Ferraz, C., & Finan, F. (2008). Exposing corrupt politicians: The effects of Brazil’s publicly released audits on electoral outcomes. Quarterly Journal of Economics, 123(2),
703–745.
Fiorina, M. P. (1981). Retrospective voting in American national elections. New Haven,
CT: Yale University Press.
Fisman, R., & Golden, M. (2017). Corruption: What everyone needs to know. Oxford:
Oxford University Press.
Healy, A., & Malhotra, N. (2013). Retrospective voting reconsidered. Annual Review
of Political Science, 16, 285–306.
Klašnja, M. (2016). Increasing rents and incumbency disadvantage. Journal of Theoretical Politics, 28(2), 225–265.
Klašnja, M., Little, A., & Tucker, J. A. (2016). Political corruption traps. Political Science
Research and Methods, 1–16.

k

k

k
PSYHB etrds0444.tex

12

k

V2 - 10/22/2018

6:14 P.M. Page 12

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Klašnja, M., Lupu, N., & Tucker, J. A. (2018). When do voters sanction corrupt politicians?
(Working Paper). Washington, D.C: University of Georgetown.
Klašnja, M., & Tucker, J. A. (2013). The economy, corruption, and the vote: Evidence
from experiments in Sweden and Moldova. Electoral Studies, 32(3), 536–543.
Klašnja, M., Tucker, J. A., & Deegan-Krause, K. (2016). Pocketbook vs. sociotropic
corruption voting. British Journal of Political Science, 46(1), 67–94.
Meirowitz, A., & Tucker, J. A. (2013). People power or a one-shot deal? A dynamic
model of protest. American Journal of Political Science, 57(2), 478–490.
Pettrigrew, T. F. (1979). The ultimate attribution error: Extending allport’s cognitive
analysis of prejudice. Personality and Social Psychology Bulletin, 5(4), 461–476.
Rundquist, B. S., Strom, G. S., & Peters, J. G. (1977). Corrupt politicians and their electoral support: Some experimental observations. American Political Science Review,
71(3), 954–963.
Shayo, M. (2009). A model of social identity with an application to political economy:
Nation, class, and redistribution. American Political Science Review, 103(2), 147–174.
Solaz, H., De Vries, C. E., & De Geus, R. (2018). In-group loyalty and the punishment
of corruption (Forthcoming). Comparative Political Studies.
Weitz-Shapiro, R., & Winters, M. S. (2017). Can citizens discern? Information credibility, political sophistication, and the punishment of corruption in Brazil. The Journal
of Politics, 79(1), 60–74.
Winters, M. S., & Weitz-Shapiro, R. (2013). Lacking information or condoning corruption: When do voters support corrupt politicians? Comparative Politics, 45(4),
418–436.
Yap, O. F. (2017). When do citizens take costly action against government corruption?
Evidence from experiments in Australia, Singapore, and the United States Journal
of East Asian Studies, 17(1), 119–136.

Catherine E. De Vries is a Professor of Political Behavior in Europe in the
Department of Political Science and Public Administration at the Vrije Universiteit Amsterdam. Her research spans a range of topics concerning public
opinion, political behavior and political corruption in the European context.
Hector Solaz is a lecturer at the John Stuart Mill College of the Vrije Universiteit Amsterdam. His research interests lie within behavioral science and
political economy. He uses lab experimental methods to study the foundations and implications of political and economic decision-making.
RELATED ESSAYS
Economic Models of Voting (Political Science), Ian G. Anson and Timothy
Hellwig
Electoral Authoritarianism (Political Science), Andreas Schedler
Presidential Power (Political Science), William G. Howell
Political Inequality (Sociology), Jeff Manza
Money in Politics (Political Science), Jeffrey Milyo

k

k

k
PSYHB etrds0444.tex

V2 - 10/22/2018

Corruption and Electoral Accountability: Avenues for Future Research

6:14 P.M. Page 13

13

Information Politics in Dictatorships (Political Science), Jeremy L. Wallace
Lawmaking (Political Science), Jamie L. Carson and Mark E. Owens
News Framing Effects and Emotions (Political Science), Andreas R. T. Schuck
and Alina Feinholdt
Political Advertising (Political Science), Erika Franklin Fowler
The Politics of Disaster Relief (Political Science), Alexander J. Oliver and
Andrew Reeves
Racial Disenfranchisement (Political Science), Vincent L. Hutchings and
Davin L. Phoenix
Perceptions of the Legitimacy of Collective Decisions (Political Science),John
W. Patty
Partisan News: A Perspective from Economics (Economics), Daniel F. Stone

k

k

k



k
PSYHB etrds0444.tex

V2 - 10/22/2018

6:14 P.M.

Page 1

Corruption and Electoral
Accountability: Avenues for Future
Research
CATHERINE E. DE VRIES and HECTOR SOLAZ

Abstract

k

The vast majority of people across the globe lives in countries characterized by
high levels of corruption, commonly defined as the public misuse of private gains.
Although the exact costs of corruption are difficult to estimate, research suggests
that corruption is bad for economic and social development. Not only has corruption
been shown to have a detrimental effect on a country’s economy, the ability to
generate tax revenue and social equality, the political effects of corruption are also
considerable. Owing to its association with weak state capacity, corruption may
damage the ability of governments to craft and implement policies in areas in which
continued intervention and investment is needed and their capacity to respond
quickly and effectively to sudden shocks. Owing to these undesirable outcomes,
elections are supposed to curb corruption because voters will throw the rascals
out. Recent research, however, suggests that more often than not corruption goes
unpunished at the ballot box. This essay sets out possible research avenues to find
out why this is the case.

k

INTRODUCTION
Many academics, policy makers, and pundits assume that fair and periodic
elections will curb political malfeasance because of voters’ disdain of
corruption. Much of the recent empirical literature, however, contradicts this
view. On average, the effective electoral sanctioning of corruption is rare (De
Vries & Solaz, 2017; Fisman & Golden, 2017). Even if there is a clear incentive
for a society at large to root out corrupt practices, as this would increase
societal welfare, political elites, and voters, especially in high-corruption
contexts, face enormous coordination challenges that favor the status quo
(Klašnja, Little, & Tucker, 2016). An important question then arises: why
voters reelect politicians who steal from them?
Emerging Trends in the Social and Behavioral Sciences.
Robert A. Scott and Marlis Buchmann (General Editors) with Stephen Kosslyn (Consulting Editor).
© 2018 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.

1

k

k
PSYHB etrds0444.tex

2

k

V2 - 10/22/2018

6:14 P.M.

Page 2

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

In previous work, we have highlighted three stages that are important
in voters’ decision-making about how to deal with candidate corruption:
an information acquisition, causal attribution, and behavioral response stage
(De Vries & Solaz, 2017). Information acquisition refers to the ways in
which voters learn about corruption. Causal attribution is the process by
which voters assign blame to a politician (or bureaucrat) for corruption.
Finally, a behavioral response denotes that voters choose to act on their
corruption evaluations and blame attribution by throwing the rascals out.
A breakdown of any one or more of these stages can result in corruption
not being punished at the ballot box. Finding out how each of these stages
matter is important for the design of behavioral or institutional solutions to
increase accountability for corruption and by doing so increasing societal
welfare.
Much of the existing literature deals with the first and last stage. When
it comes to information acquisition, empirical work demonstrates that a
lack of punishment of corruption is the result of the quantity and quality of
information that voters receive. Information dissemination could be imperfect due to a lack of media reporting, political control of the media, high
levels of illiteracy, or the lack of credible sources (Botero, Cornejo, Gamboa,
Pavão, & Nickerson, 2015; Ferraz & Finan, 2008; Winters & Weitz-Shapiro,
2013). When it comes to the behavioral response stage, a wealth of evidence
suggests that voters trade-off corruption information against other benefits
when they decide how to cast their ballot. These benefits could be based on
side-payments (Fernández-Vázquez et al., 2016), performance in other areas,
or policy pledges (Klašnja & Tucker, 2013; Rundquist, Strom, & Peters, 1977).
Hence, voters may tolerate corrupt practices, or even approve of them,
because they value other things more.
The way in which causal attribution matters has received far less scholarly attention. Could it be the case that voters simply might not update their
beliefs about candidate quality even when presented with concrete evidence
of malfeasance? Some recent field and laboratory experimental evidence suggests that this is indeed the case. So far, this is based on only a handful of
studies and we do not know if these findings are generalizable. Moreover, to
date, we do not yet have a comprehensive understanding of the factors that
explain why voters fail to update their beliefs about candidate quality even
if they are confronted with unambiguous corruption information. Against
this backdrop, we suggest that the main challenge for future research is to
develop a coherent understanding of how causal attribution matters for the
electoral punishment of corruption.
In the ensuing sections, we first elaborate the three stages that are important in voters’ decision-making about candidate corruption and discuss existing work in terms of the stages they examine. Second, we highlight several

k

k

k
PSYHB etrds0444.tex

V2 - 10/22/2018

Corruption and Electoral Accountability: Avenues for Future Research

6:14 P.M.

Page 3

3

important avenues for future research regarding the role of causal attribution.
Finally, we conclude by highlighting why these avenues of future research
are so important for our understanding of corruption more broadly.
SHORT OVERVIEW OF EXISTING RESEARCH: THE IMPORTANCE
OF INFORMATION AND OTHER BENEFITS

k

The idea that voters hold politicians to account for corruption fits classical
models of retrospective voting. These models maintain that voters use elections to sanction low-performing politicians by selecting high-performing
ones (Besley, 2006). In order to win elections, candidates need to convince
voters that they are competent and thus able to promote and secure their
well-being. Voters form beliefs about candidate competence by paying attention to quality signals, such as developments in their living standards, the
current state of the economy or corruption scandals among other things (Fiorina, 1981). When voters indeed cast their ballots based on these beliefs about
candidate quality, we can expect that periodic and competitive elections help
to ensure that high-quality candidates get elected and are being reelected.
Information about candidate corruption is important, because it acts as a signal for voters that candidates are of low quality. As a result, we would expect
corrupt politicians to be punished at the ballot box, and the extent to which
this happens to vary by the quality of the signal. However persuasive this theoretical account may be, the real world turns out to be more complex. While
some studies find that corrupt activities indeed take a considerable electoral
toll on incumbents, especially when the information signal is clear or people’s
ability to distill it is high (Ferraz & Finan, 2008; Klašnja, 2016; Weitz-Shapiro &
Winters, 2017), other work suggests that voters might reelect politicians they
know to be corrupt even when confronted with strong evidence of malfeasance, because they prefer other benefits or due to group loyalty (Anduiza,
Gallego, & Munoz, 2013; Rundquist et al., 1977; Solaz, De Vries, & De Geus,
2018).
These conflicting findings might not be entirely surprising when we
consider that several conditions should be met in order for retrospective
voting to exist (Healy & Malhotra, 2013). In our previous work, we outlined three stages that are important for holding candidates accountable
for political malfeasance, an information acquisition, causal attribution, and
behavioral response stage. Information acquisition relates to voters needing
to be aware of corruption. In the first stage of information acquisition, it
follows that in order for voters to be able to punish corruption, they need
to know about it. Becoming aware of corruption is not always easy, because
elected public officials engaged in corrupt activities have an incentive to
hide information, and may try to influence the media or the judiciary not to

k

k

k
PSYHB etrds0444.tex

4

k

V2 - 10/22/2018

6:14 P.M.

Page 4

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

report on or prosecute their malfeasance. The literature suggests that people
learn about corruption either through personal experience or through
indirect perception by ways of media reporting for example (Klašnja,
Tucker, & Deegan-Krause, 2016). The second stage of causal attribution
refers to the fact that when voters are informed about candidate corruption,
they need to attribute responsibility for it and adjust their evaluations of
the candidate quality accordingly. Research has demonstrated that blame
attribution is far from straightforward, and often plagued by group-serving
biases. Whereas voters tend to attribute positive events or outcomes to their
preferred in-group, they absolve the in-group of blame for negative events
or outcomes (Anduiza et al., 2013; Solaz et al., 2018). The last stage refers
to behavioral responses. These blame attributions should also lead voters
to adjust their voting decisions by either voting for the opposition or by
refraining to cast a ballot in the first place. Voters have to consider a large
array of factors when deciding for whom to cast their ballot, and corruption
may simply not be a salient issue in a specific election campaign (Chang,
Golden, & Hill, 2010; Klašnja, Tucker, et al., 2016). Moreover, even if voters
find corruption important, they might be willing to trade it off against some
other benefit they care about more (Rundquist et al., 1977).
The majority of work to date has focused on the information acquisition
and behavioral response stage. When it comes to the former, studies suggest that the lack of electoral punishment of corruption is largely a function of the quantity of information available to voters (Chang et al., 2010;
Ferraz & Finan, 2008), and the quality of that information (Botero et al., 2015;
Winters & Weitz-Shapiro, 2013). A seminal study on the role of information conducted by Ferraz and Finan (2008) studies the effects of publicly
released audit reports about corruption practices in Brazilian municipalities.
The authors show that information dissemination increases the electoral punishment of corruption, and that these effects were most pronounced in those
municipalities where locally based radio stations disclosed this information.
Recent work in a similar vein suggests that not only the quantity of information matters but also the perceived quality of the source of information
(Botero et al., 2015; Winters & Weitz-Shapiro, 2013). The source deemed most
credible by voters may differ greatly from context to context. For instance,
a study from Brazil shows that respondents were more likely to punish a
candidate for corruption when the information stems from a federal audit
rather than from the opposition party (Winters & Weitz-Shapiro, 2013), while
a study conducted in Columbia found that corruption allegations made by
the leading national newspaper rather than a nongovernmental organization
or the judiciary engenders the strongest effects on punishment (Botero et al.,
2015).

k

k

k
PSYHB etrds0444.tex

V2 - 10/22/2018

Corruption and Electoral Accountability: Avenues for Future Research

6:14 P.M.

Page 5

5

Another set of studies suggests that corruption information might not lead
voters to punish malfeasance, because they care about other benefits more.
One of the first empirical papers on the topic by Rundquist and colleagues
featuring a set of survey experiments suggested that voters often engage in
a trade. Although they may generally disapprove of corrupt conduct, they
need to weigh their corruption evaluations against policy benefits, and this
may result in an overall positive balance in favor of reelection. As a result,
they might prefer a corrupt candidate who represents their policy positions to
a clean candidate who does not. Voters may trade off corruption against some
material gains (Fernández-Vázquez, Barberá, & Rivero, 2016), or turn a blind
eye in favor of other benefits such as ideological proximity (Rundquist et al.,
1977) or strong economic performance (Klašnja & Tucker, 2013). Overall, this
evidence suggests that because voters care about a multitude of things when
deciding to cast their ballot, oftentimes corruption becomes a secondary concern compared to other things like the state of the economy or a candidate’s
policy positions. If voters fail to punish corruption at the ballot box due to
instrumental calculations, such behavior might not be irrational, but still normatively undesirable, as it does not incentivize politicians to be clean.

k

AVENUES FOR FUTURE WORK: TOWARD A BETTER
UNDERSTANDING OF THE ROLE OF CAUSAL ATTRIBUTION
AND PRIOR BELIEFS
Notwithstanding the compelling evidence outlining the important role that
both information acquisition and voters’ decisions to trade off corruption
against other benefits play, to date we know much less about the intermediate stage of causal attribution. This is unfortunate, as we know that people
do not process information objectively. Social psychologists, for example,
have shown that people tend to make internal attributions (to their own
in-group) for positive events or outcomes and make external attributions
(to an out-group) for negative events or outcomes, something coined the
“ultimate attribution error” (Pettrigrew, 1979). While studies on retrospective
voting in political science have over the years paid considerable attention
to causal attribution (for an overview see Healy & Malhotra, 2013), to our
knowledge no study to date has explicitly examined causal attributions for
corruption as a dependent variable. In the following sections, we suggest
that understanding (i) how people attribute blame when exposed to information about corruption, and (ii) which factors hamper blame attribution
are two important areas that deserve more scholarly attention.
As highlighted in the previous section, existing work centers on the study
of how voters get informed about candidate corruption, and how they trade
this information off against other benefits they may care about. Yet, could

k

k

k
PSYHB etrds0444.tex

6

k

V2 - 10/22/2018

6:14 P.M.

Page 6

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

it be the case that voters simply do not update their beliefs about candidate
quality even when presented with unequivocal evidence of malfeasance?
When group loyalties, based on partisanship, religion, ethnicity, and so on,
are strong, corruption information of in-group members might not have
any effect on voter’s decision-making. Owing to the fact that corruption
allegations may pose a serious threat to in-group status and the in-group’s
hold on power, people might absolve in-group members of blame. The
most straightforward channel through which this would work is through
people’s in-group serving bias (Solaz et al., 2018). While voters tend to
reward the in-group for positive outcomes, they absolve it of blame for
negative outcomes. This would constitute a form of motivated reasoning
in which people, because of their automatic affective ties to their in-group,
selectively process information. Another channel through which this may
work is through in-group status. Evidence from behavioral economics shows
that people may gain some utility from the utility of others, in particular
from members of their own in-group (Charness & Rabin, 2002). In his work
on redistribution, for example, Shayo (2009) shows that individuals derive
utility from having high-status members in their group. Because of this,
poorer groups might not support redistribution from the rich to the poor
even though it would benefit them personally. If we extend this insight to
corruption-generated wealth, voters could gain some utility from the fact
that a candidate from their in-group stands to gain from corruption even if
they would personally lose out. In this case, voters might actually reward
candidates from their own group for being corrupt.
In a recent study, we started to examine these possible channels by exploiting a real-world corruption scandal in Spain that fell in the middle of a survey
fieldwork period and complementing it with laboratory experiments from
the United Kingdom (Solaz et al., 2018). By comparing Spanish respondents
before and after a corruption scandal was uncovered in a national newspaper,
we show that in-group loyalty based on partisanship decreased the likelihood of withdrawing support for the ruling party due to corruption, while
out-group identification increased it. In a next step, we replicated our results
in a laboratory setting that allowed us to isolate in-group effects more carefully by controlling the precise setting and information voters they received
as well as randomizing the group identity of the candidates that voters faced.
The laboratory results support our findings from the Spanish case: voters fail
to punish in-group candidates they know to be corrupt but punish out-group
candidates for the same indiscretions. These findings are important and lead
to a host of new questions. Which factors could condition the magnitude
of the effects of in-group serving biases? And what kind of group identities
might trigger these kinds of in-group effects, and which ones do not? These
are crucially important questions for future work to address.

k

k

k
PSYHB etrds0444.tex

V2 - 10/22/2018

Corruption and Electoral Accountability: Avenues for Future Research

k

Page 7

7

Another important avenue of future research is the importance of nature of
prior beliefs for understanding the degree to which voters, in fact, update
their evaluations of candidate quality in light of malfeasance revelations.
A recent study by Arias, Balan, Larreguy, Marshall, and Querubin (2017)
suggests that electoral punishment of corruption is rare when voters already
believe that candidates are of low quality. Owing to their negative priors, a
signal that the incumbent is corrupt will not be very informative. In fact, their
evidence stemming from a field experiment conducted in Mexico, where voters were informed about malfeasant municipal spending or not, suggests that
when priors were already negative, support for the incumbent party actually
increased after corruption was revealed because voters positively updated
their priors from information that was better than expected. These findings
suggest that remedying information asymmetries may not always lead to the
results that scholars or practitioners hope for. The evidence also leads to a
host of new questions, for example, which prior beliefs matter, those relating
to the incumbent as such or those relating to the prevalence of corruption
in society at large? And under which conditions do these different types of
priors matter most?
Concerning the latter, a study by Klašnja and Tucker (2013), comparing survey experimental evidence from Sweden and Moldova, for example, shows
that voters in Sweden, a low-corruption country, responded more to information about corruption than voters in Moldova, a high-corruption country,
do. What is more, voters in Moldova only responded to corruption information when economic conditions were poor. This perhaps suggests that while
on average due to their priors about the prevalence of corruption, voters in
high-corruption countries may not respond to one additional case of corruption, some heterogeneity in reactions most likely exists. This heterogeneity
could be the result of the overall state of the economy as the study by Klašnja
and Tucker suggests or individuals’ prior beliefs about candidate quality as
highlighted by Arias et al. (2017), but it could also be due to other prior beliefs
people might hold.
We think that people’s social priors might be especially important. Social
priors refer to people’s beliefs about society at large. The role of three sets
of social priors seem especially important in our view, people’s beliefs about
(i) the prevalence of corruption within society, (ii) the way other voters faced with
corruption allegations will act and (iii) the quality of the pool of future political
entrants. These social priors are likely going to condition the extent to which
voters will actually update their beliefs about candidate quality and blame
the incumbent for corrupt activities. The pervasiveness of these prior beliefs,
especially in high-corruption countries, may make it less likely that candidate
corruption will be punished at the ballot box. Let us address each of them
in turn.

k

6:14 P.M.

k

k
PSYHB etrds0444.tex

8

k

V2 - 10/22/2018

6:14 P.M.

Page 8

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

When people already think that corruption is widespread, they do not
punish corruption at the ballot box. This could be due to different reasons.
It might be due to the fact that when corruption is believed to be prevalent,
corruption allegations add little new information to voters’ decision-making
and therefore might carry very little weight in voters’ minds. This type
of reasoning is consistent with arguments that citizens residing in countries with higher quality of governance might display greater sensitivity to
accountability failures (Ashworth, Bueno de Mesquita, & Friedenberg, 2017).
It could also be the case that when corruption is more widespread, engaging
in corrupt practices might be deemed more acceptable by voters (Fisman
& Golden, 2017). Yet, it could also be that if voters think that corruption is
widespread, they also believe that it takes a corrupt politician to navigate the
system and to get things done (Bauhr & Charron, 2018). A recent study based
on conjoint experiments that was embedded in surveys in Argentina, Chile,
and Uruguay shows no effect of reminding people about the prevalence
of corruption in their country (Klašnja, Lupu, & Tucker, 2018). Yet, as the
authors themselves admit, empirically establishing the effect of people’s
perceptions about the pervasiveness of corruption is difficult, in part because
people’s understandings of what corruption might actually entail are likely
going to differ substantially between contexts in which corruption is more
or less widespread, and among individuals. Studying how people’s beliefs
about the prevalence of corruption matter is an important area for future
research.
Second, not only people’s perceptions about the extent of corruption in
society will likely affect their willingness to hold candidates to account
for corruption, but also their expectations about the behavior of others.
People’s prior beliefs about the behavior of other voters relate to the
importance of voter coordination in the punishment of corruption. When
people think that others are unlikely to punish existing politicians for
corruption, they are unlikely to do so as well (especially when there is
any form of side-benefit that they could lose by withdrawing support
from a corrupt incumbent). In order for voter coordination to occur, two
conditions need to be met: first, voters have to agree that corruption is
important for their vote, and second, they need to be able to agree to vote
for another clean candidate. A study by Chang et al. (2010) highlights the
importance of voter coordination. It is based on legislators in Italy’s lower
house between 1948 and 1994 and candidates of the two largest parties
(Christian Democracy and the Italian Communist Party). The authors show
that only in the 1992–1994 legislature did corrupt legislators face a serious
electoral penalty. They suggest that this structural break in the response to
corruption by Italian voters was due to the media coverage that grew out of
the Clean Hands operation in the early 1990s. The authors argue that a wider

k

k

k
PSYHB etrds0444.tex

V2 - 10/22/2018

Corruption and Electoral Accountability: Avenues for Future Research

k

Page 9

9

dissemination of information about corruption in the media not only makes
corruption more salient in voters’ minds but also provides a coordination
signal to voters, who realize that others will most likely punish corruption
as well.
Several recent studies examined the role of coordination signals through
different types of experimental treatments and found mixed results (Adida,
Gottlieb, Kramon, & McClendon, 2017; Arias, Larreguy, Marshall, &
Querubin, 2018; Yap, 2017). Citing evidence from a field experiment conducted in the context of legislative elections in Benin, Adida and colleagues
show that voters only punished bad performance when information was
accompanied with a message highlighting the importance of legislative performance for voter welfare and the fact that other villages in the constituency
received the same information. Laboratory evidence from Australia, Singapore, and the United States also highlights the importance of positive
coordination signals (Yap, 2017). Yet, evidence from a field experiment
revealing mayoral malfeasance before the 2015 Mexican municipal elections
suggests that accompanying leaflets detailing results of audit reports by
loudspeakers in order to signal to voters that others received the malfeasance
information as well did not significantly alter the effects of information on
incumbent vote shares (Arias et al., 2018). All in all, these results highlight
that understanding (i) which coordination signals matter and (ii) which
conditions may hamper their effects are two important areas for future
research.
Finally, voters’ beliefs about the quality of the pool of future political
entrants is likely important for their willingness to hold candidates to
account for corruption. When people think that the pool of possible political
entrants will also be corrupt, there is little reason to punish an existing
politician for corruption. In their theoretical work on corruption traps,
Klašnja, Little, et al. (2016) highlighted that overcoming corruption involves
coordination not only among voters and existing candidates but also among
possible entrants. Voters’ perceptions about the quality of the pool of
possible entrants will inform their behavior. When voters’ expect an alternative pool of candidates for office to be likely as corrupt as current ones,
why would they cast a ballot based on corruption? The pool of potential
candidates has been found to be important in other aspects of political
behavior as well, such as protest behavior (Meirowitz and Tucker, 2013).
When voters become informed about widespread corruption of incumbents,
this might influence their beliefs about the overall average quality of the
pool of potential governments, especially if they have little experience with
elections to begin with. While in established democracies, learning about
corruption may lead people to think that they should root out this one bad
apple, in newly established democracies this may not spark off pessimism

k

6:14 P.M.

k

k
PSYHB etrds0444.tex

10

V2 - 10/22/2018

6:14 P.M. Page 10

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

about all possible political candidates. Empirical examinations regarding
(i) the way in which people’s expectations about the quality of the pool of
possible alternatives affect the likelihood of punishing corruption, and (ii)
how these relationships differ based on people’s experience with democracy
constitute important avenues for future research.
CONCLUSION

k

In recent years, we have experienced somewhat of a boom in examining the
extent to which people hold politicians to account for corruption, not least
due to the widespread use of field, laboratory, and survey experiments. These
experimental tools help researchers to overcome some of the core problems
associated with studying corruption, such as the fact that politicians try to
hide malfeasance and people may not want to admit that they support corrupt politicians. Much of this recent work has centered on the importance of
corruption information and the willingness of voters, once they are informed
about corruption, to trade it off against other benefits. In this essay, we have
highlighted that the role of causal attribution has thus far received far less
scholarly attention. Based on a handful of recent studies, we suggested that
corruption might go unpunished at the ballot box simply because voters do
not update their beliefs about candidate quality in light of unambiguous
information about malfeasance. Recent work suggests that this is largely the
result of the weight voters attach to in-group status and their prior beliefs
about candidate quality. Yet, we do not know under which conditions these
factors are important, and if other beliefs matter as well. Moreover, most of
this early experimental work stems from one country context at one point in
time, so we do not know if we can generalize existing findings.
Overall, we have highlighted two key avenues for future research. First,
what is the role of role of attribution errors? To what extent do people fail to
blame their in-group candidate for corruption, and punish out-group candidates for the same indiscretion? Which factors may condition the magnitude
of these in-group effects? What is more, which kinds of group identities may
trigger these effects, and which ones do not? These are crucially important
questions for future work to address. Second, what is the role of people’s
beliefs about society at large? To what extent does the electoral punishment
of corruption hinge on people’s priors about the prevalence of corruption
within society, the way they expect other voters to act, and the quality of the
pool of future politicians? We believe that these social priors are crucial for
understanding the conditions under which corruption gets punished at the
ballot box. The extent to which they actually do, is an important topic for
future research.

k

k

k
PSYHB etrds0444.tex

V2 - 10/22/2018

Corruption and Electoral Accountability: Avenues for Future Research

6:14 P.M. Page 11

11

REFERENCES

k

Adida, C., Gottlieb, J., Kramon, E., & McClendon, G. (2017, November). Breaking the
clientelistic voting equilibrium: The joint importance of salience and coordination (Working Paper 48). Washington, DC: AIDDATA.
Anduiza, E., Gallego, A., & Munoz, J. (2013). Turning a blind eye: Experimental evidence of partisan bias in attitudes toward corruption. Comparative Political Studies,
46(12), 1664–1692.
Arias, E., Balan, P., Larreguy, H., Marshall, J., & Querubin, P. (2017). Priors rule: When
malfeasance revelations help and hurt incumbent parties? (Working Paper). Cambridge,
MA: Harvard University.
Arias, E., Larreguy, H., Marshall, J., & Querubin, P. (2018). Does the content and mode
of delivery of information matter for electoral accountability? Evidence from a field experiment in Mexico (Working Paper). Cambridge, MA: Harvard University.
Ashworth, S., Bueno de Mesquita, E., & Friedenberg, A. (2017). Accountability and
information in elections. American Economic Journal: Microeconomics, 9(2), 95–138.
Bauhr, M., & Charron, N. (2018). Insider or outsider? Grand corruption and electoral
accountability. Comparative Political Studies, 51(4), 415–446.
Besley, T. (2006). Principled agents? The political economy of good government. Oxford,
UK: Oxford University Press.
Botero, S., Cornejo, R. C., Gamboa, L., Pavão, N., & Nickerson, D. W. (2015). Says
who? An experiment on allegations of corruption and credibility of sources. Political Research Quarterly, 68, 493–504.
Chang, E. C., Golden, M. A., & Hill, S. J. (2010). Legislative malfeasance and political
accountability. World Politics, 62(2), 177–220.
Charness, G., & Rabin, M. (2002). Understanding social preferences with simple tests.
Quarterly Journal of Economics, 117(3), 817–869.
De Vries, C. E., & Solaz, H. (2017). The electoral consequences of corruption. Annual
Review of Political Science, 20, 391–408.
Fernández-Vázquez, P., Barberá, P., & Rivero, G. (2016). Rooting out corruption or
rooting for corruption? The heterogeneous electoral consequences of scandals.
Political Science Research and Methods, 4(2), 379–397.
Ferraz, C., & Finan, F. (2008). Exposing corrupt politicians: The effects of Brazil’s publicly released audits on electoral outcomes. Quarterly Journal of Economics, 123(2),
703–745.
Fiorina, M. P. (1981). Retrospective voting in American national elections. New Haven,
CT: Yale University Press.
Fisman, R., & Golden, M. (2017). Corruption: What everyone needs to know. Oxford:
Oxford University Press.
Healy, A., & Malhotra, N. (2013). Retrospective voting reconsidered. Annual Review
of Political Science, 16, 285–306.
Klašnja, M. (2016). Increasing rents and incumbency disadvantage. Journal of Theoretical Politics, 28(2), 225–265.
Klašnja, M., Little, A., & Tucker, J. A. (2016). Political corruption traps. Political Science
Research and Methods, 1–16.

k

k

k
PSYHB etrds0444.tex

12

k

V2 - 10/22/2018

6:14 P.M. Page 12

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Klašnja, M., Lupu, N., & Tucker, J. A. (2018). When do voters sanction corrupt politicians?
(Working Paper). Washington, D.C: University of Georgetown.
Klašnja, M., & Tucker, J. A. (2013). The economy, corruption, and the vote: Evidence
from experiments in Sweden and Moldova. Electoral Studies, 32(3), 536–543.
Klašnja, M., Tucker, J. A., & Deegan-Krause, K. (2016). Pocketbook vs. sociotropic
corruption voting. British Journal of Political Science, 46(1), 67–94.
Meirowitz, A., & Tucker, J. A. (2013). People power or a one-shot deal? A dynamic
model of protest. American Journal of Political Science, 57(2), 478–490.
Pettrigrew, T. F. (1979). The ultimate attribution error: Extending allport’s cognitive
analysis of prejudice. Personality and Social Psychology Bulletin, 5(4), 461–476.
Rundquist, B. S., Strom, G. S., & Peters, J. G. (1977). Corrupt politicians and their electoral support: Some experimental observations. American Political Science Review,
71(3), 954–963.
Shayo, M. (2009). A model of social identity with an application to political economy:
Nation, class, and redistribution. American Political Science Review, 103(2), 147–174.
Solaz, H., De Vries, C. E., & De Geus, R. (2018). In-group loyalty and the punishment
of corruption (Forthcoming). Comparative Political Studies.
Weitz-Shapiro, R., & Winters, M. S. (2017). Can citizens discern? Information credibility, political sophistication, and the punishment of corruption in Brazil. The Journal
of Politics, 79(1), 60–74.
Winters, M. S., & Weitz-Shapiro, R. (2013). Lacking information or condoning corruption: When do voters support corrupt politicians? Comparative Politics, 45(4),
418–436.
Yap, O. F. (2017). When do citizens take costly action against government corruption?
Evidence from experiments in Australia, Singapore, and the United States Journal
of East Asian Studies, 17(1), 119–136.

Catherine E. De Vries is a Professor of Political Behavior in Europe in the
Department of Political Science and Public Administration at the Vrije Universiteit Amsterdam. Her research spans a range of topics concerning public
opinion, political behavior and political corruption in the European context.
Hector Solaz is a lecturer at the John Stuart Mill College of the Vrije Universiteit Amsterdam. His research interests lie within behavioral science and
political economy. He uses lab experimental methods to study the foundations and implications of political and economic decision-making.
RELATED ESSAYS
Economic Models of Voting (Political Science), Ian G. Anson and Timothy
Hellwig
Electoral Authoritarianism (Political Science), Andreas Schedler
Presidential Power (Political Science), William G. Howell
Political Inequality (Sociology), Jeff Manza
Money in Politics (Political Science), Jeffrey Milyo

k

k

k
PSYHB etrds0444.tex

V2 - 10/22/2018

Corruption and Electoral Accountability: Avenues for Future Research

6:14 P.M. Page 13

13

Information Politics in Dictatorships (Political Science), Jeremy L. Wallace
Lawmaking (Political Science), Jamie L. Carson and Mark E. Owens
News Framing Effects and Emotions (Political Science), Andreas R. T. Schuck
and Alina Feinholdt
Political Advertising (Political Science), Erika Franklin Fowler
The Politics of Disaster Relief (Political Science), Alexander J. Oliver and
Andrew Reeves
Racial Disenfranchisement (Political Science), Vincent L. Hutchings and
Davin L. Phoenix
Perceptions of the Legitimacy of Collective Decisions (Political Science),John
W. Patty
Partisan News: A Perspective from Economics (Economics), Daniel F. Stone

k

k

k


k
PSYHB etrds0444.tex

V2 - 10/22/2018

6:14 P.M.

Page 1

Corruption and Electoral
Accountability: Avenues for Future
Research
CATHERINE E. DE VRIES and HECTOR SOLAZ

Abstract

k

The vast majority of people across the globe lives in countries characterized by
high levels of corruption, commonly defined as the public misuse of private gains.
Although the exact costs of corruption are difficult to estimate, research suggests
that corruption is bad for economic and social development. Not only has corruption
been shown to have a detrimental effect on a country’s economy, the ability to
generate tax revenue and social equality, the political effects of corruption are also
considerable. Owing to its association with weak state capacity, corruption may
damage the ability of governments to craft and implement policies in areas in which
continued intervention and investment is needed and their capacity to respond
quickly and effectively to sudden shocks. Owing to these undesirable outcomes,
elections are supposed to curb corruption because voters will throw the rascals
out. Recent research, however, suggests that more often than not corruption goes
unpunished at the ballot box. This essay sets out possible research avenues to find
out why this is the case.

k

INTRODUCTION
Many academics, policy makers, and pundits assume that fair and periodic
elections will curb political malfeasance because of voters’ disdain of
corruption. Much of the recent empirical literature, however, contradicts this
view. On average, the effective electoral sanctioning of corruption is rare (De
Vries & Solaz, 2017; Fisman & Golden, 2017). Even if there is a clear incentive
for a society at large to root out corrupt practices, as this would increase
societal welfare, political elites, and voters, especially in high-corruption
contexts, face enormous coordination challenges that favor the status quo
(Klašnja, Little, & Tucker, 2016). An important question then arises: why
voters reelect politicians who steal from them?
Emerging Trends in the Social and Behavioral Sciences.
Robert A. Scott and Marlis Buchmann (General Editors) with Stephen Kosslyn (Consulting Editor).
© 2018 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.

1

k

k
PSYHB etrds0444.tex

2

k

V2 - 10/22/2018

6:14 P.M.

Page 2

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

In previous work, we have highlighted three stages that are important
in voters’ decision-making about how to deal with candidate corruption:
an information acquisition, causal attribution, and behavioral response stage
(De Vries & Solaz, 2017). Information acquisition refers to the ways in
which voters learn about corruption. Causal attribution is the process by
which voters assign blame to a politician (or bureaucrat) for corruption.
Finally, a behavioral response denotes that voters choose to act on their
corruption evaluations and blame attribution by throwing the rascals out.
A breakdown of any one or more of these stages can result in corruption
not being punished at the ballot box. Finding out how each of these stages
matter is important for the design of behavioral or institutional solutions to
increase accountability for corruption and by doing so increasing societal
welfare.
Much of the existing literature deals with the first and last stage. When
it comes to information acquisition, empirical work demonstrates that a
lack of punishment of corruption is the result of the quantity and quality of
information that voters receive. Information dissemination could be imperfect due to a lack of media reporting, political control of the media, high
levels of illiteracy, or the lack of credible sources (Botero, Cornejo, Gamboa,
Pavão, & Nickerson, 2015; Ferraz & Finan, 2008; Winters & Weitz-Shapiro,
2013). When it comes to the behavioral response stage, a wealth of evidence
suggests that voters trade-off corruption information against other benefits
when they decide how to cast their ballot. These benefits could be based on
side-payments (Fernández-Vázquez et al., 2016), performance in other areas,
or policy pledges (Klašnja & Tucker, 2013; Rundquist, Strom, & Peters, 1977).
Hence, voters may tolerate corrupt practices, or even approve of them,
because they value other things more.
The way in which causal attribution matters has received far less scholarly attention. Could it be the case that voters simply might not update their
beliefs about candidate quality even when presented with concrete evidence
of malfeasance? Some recent field and laboratory experimental evidence suggests that this is indeed the case. So far, this is based on only a handful of
studies and we do not know if these findings are generalizable. Moreover, to
date, we do not yet have a comprehensive understanding of the factors that
explain why voters fail to update their beliefs about candidate quality even
if they are confronted with unambiguous corruption information. Against
this backdrop, we suggest that the main challenge for future research is to
develop a coherent understanding of how causal attribution matters for the
electoral punishment of corruption.
In the ensuing sections, we first elaborate the three stages that are important in voters’ decision-making about candidate corruption and discuss existing work in terms of the stages they examine. Second, we highlight several

k

k

k
PSYHB etrds0444.tex

V2 - 10/22/2018

Corruption and Electoral Accountability: Avenues for Future Research

6:14 P.M.

Page 3

3

important avenues for future research regarding the role of causal attribution.
Finally, we conclude by highlighting why these avenues of future research
are so important for our understanding of corruption more broadly.
SHORT OVERVIEW OF EXISTING RESEARCH: THE IMPORTANCE
OF INFORMATION AND OTHER BENEFITS

k

The idea that voters hold politicians to account for corruption fits classical
models of retrospective voting. These models maintain that voters use elections to sanction low-performing politicians by selecting high-performing
ones (Besley, 2006). In order to win elections, candidates need to convince
voters that they are competent and thus able to promote and secure their
well-being. Voters form beliefs about candidate competence by paying attention to quality signals, such as developments in their living standards, the
current state of the economy or corruption scandals among other things (Fiorina, 1981). When voters indeed cast their ballots based on these beliefs about
candidate quality, we can expect that periodic and competitive elections help
to ensure that high-quality candidates get elected and are being reelected.
Information about candidate corruption is important, because it acts as a signal for voters that candidates are of low quality. As a result, we would expect
corrupt politicians to be punished at the ballot box, and the extent to which
this happens to vary by the quality of the signal. However persuasive this theoretical account may be, the real world turns out to be more complex. While
some studies find that corrupt activities indeed take a considerable electoral
toll on incumbents, especially when the information signal is clear or people’s
ability to distill it is high (Ferraz & Finan, 2008; Klašnja, 2016; Weitz-Shapiro &
Winters, 2017), other work suggests that voters might reelect politicians they
know to be corrupt even when confronted with strong evidence of malfeasance, because they prefer other benefits or due to group loyalty (Anduiza,
Gallego, & Munoz, 2013; Rundquist et al., 1977; Solaz, De Vries, & De Geus,
2018).
These conflicting findings might not be entirely surprising when we
consider that several conditions should be met in order for retrospective
voting to exist (Healy & Malhotra, 2013). In our previous work, we outlined three stages that are important for holding candidates accountable
for political malfeasance, an information acquisition, causal attribution, and
behavioral response stage. Information acquisition relates to voters needing
to be aware of corruption. In the first stage of information acquisition, it
follows that in order for voters to be able to punish corruption, they need
to know about it. Becoming aware of corruption is not always easy, because
elected public officials engaged in corrupt activities have an incentive to
hide information, and may try to influence the media or the judiciary not to

k

k

k
PSYHB etrds0444.tex

4

k

V2 - 10/22/2018

6:14 P.M.

Page 4

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

report on or prosecute their malfeasance. The literature suggests that people
learn about corruption either through personal experience or through
indirect perception by ways of media reporting for example (Klašnja,
Tucker, & Deegan-Krause, 2016). The second stage of causal attribution
refers to the fact that when voters are informed about candidate corruption,
they need to attribute responsibility for it and adjust their evaluations of
the candidate quality accordingly. Research has demonstrated that blame
attribution is far from straightforward, and often plagued by group-serving
biases. Whereas voters tend to attribute positive events or outcomes to their
preferred in-group, they absolve the in-group of blame for negative events
or outcomes (Anduiza et al., 2013; Solaz et al., 2018). The last stage refers
to behavioral responses. These blame attributions should also lead voters
to adjust their voting decisions by either voting for the opposition or by
refraining to cast a ballot in the first place. Voters have to consider a large
array of factors when deciding for whom to cast their ballot, and corruption
may simply not be a salient issue in a specific election campaign (Chang,
Golden, & Hill, 2010; Klašnja, Tucker, et al., 2016). Moreover, even if voters
find corruption important, they might be willing to trade it off against some
other benefit they care about more (Rundquist et al., 1977).
The majority of work to date has focused on the information acquisition
and behavioral response stage. When it comes to the former, studies suggest that the lack of electoral punishment of corruption is largely a function of the quantity of information available to voters (Chang et al., 2010;
Ferraz & Finan, 2008), and the quality of that information (Botero et al., 2015;
Winters & Weitz-Shapiro, 2013). A seminal study on the role of information conducted by Ferraz and Finan (2008) studies the effects of publicly
released audit reports about corruption practices in Brazilian municipalities.
The authors show that information dissemination increases the electoral punishment of corruption, and that these effects were most pronounced in those
municipalities where locally based radio stations disclosed this information.
Recent work in a similar vein suggests that not only the quantity of information matters but also the perceived quality of the source of information
(Botero et al., 2015; Winters & Weitz-Shapiro, 2013). The source deemed most
credible by voters may differ greatly from context to context. For instance,
a study from Brazil shows that respondents were more likely to punish a
candidate for corruption when the information stems from a federal audit
rather than from the opposition party (Winters & Weitz-Shapiro, 2013), while
a study conducted in Columbia found that corruption allegations made by
the leading national newspaper rather than a nongovernmental organization
or the judiciary engenders the strongest effects on punishment (Botero et al.,
2015).

k

k

k
PSYHB etrds0444.tex

V2 - 10/22/2018

Corruption and Electoral Accountability: Avenues for Future Research

6:14 P.M.

Page 5

5

Another set of studies suggests that corruption information might not lead
voters to punish malfeasance, because they care about other benefits more.
One of the first empirical papers on the topic by Rundquist and colleagues
featuring a set of survey experiments suggested that voters often engage in
a trade. Although they may generally disapprove of corrupt conduct, they
need to weigh their corruption evaluations against policy benefits, and this
may result in an overall positive balance in favor of reelection. As a result,
they might prefer a corrupt candidate who represents their policy positions to
a clean candidate who does not. Voters may trade off corruption against some
material gains (Fernández-Vázquez, Barberá, & Rivero, 2016), or turn a blind
eye in favor of other benefits such as ideological proximity (Rundquist et al.,
1977) or strong economic performance (Klašnja & Tucker, 2013). Overall, this
evidence suggests that because voters care about a multitude of things when
deciding to cast their ballot, oftentimes corruption becomes a secondary concern compared to other things like the state of the economy or a candidate’s
policy positions. If voters fail to punish corruption at the ballot box due to
instrumental calculations, such behavior might not be irrational, but still normatively undesirable, as it does not incentivize politicians to be clean.

k

AVENUES FOR FUTURE WORK: TOWARD A BETTER
UNDERSTANDING OF THE ROLE OF CAUSAL ATTRIBUTION
AND PRIOR BELIEFS
Notwithstanding the compelling evidence outlining the important role that
both information acquisition and voters’ decisions to trade off corruption
against other benefits play, to date we know much less about the intermediate stage of causal attribution. This is unfortunate, as we know that people
do not process information objectively. Social psychologists, for example,
have shown that people tend to make internal attributions (to their own
in-group) for positive events or outcomes and make external attributions
(to an out-group) for negative events or outcomes, something coined the
“ultimate attribution error” (Pettrigrew, 1979). While studies on retrospective
voting in political science have over the years paid considerable attention
to causal attribution (for an overview see Healy & Malhotra, 2013), to our
knowledge no study to date has explicitly examined causal attributions for
corruption as a dependent variable. In the following sections, we suggest
that understanding (i) how people attribute blame when exposed to information about corruption, and (ii) which factors hamper blame attribution
are two important areas that deserve more scholarly attention.
As highlighted in the previous section, existing work centers on the study
of how voters get informed about candidate corruption, and how they trade
this information off against other benefits they may care about. Yet, could

k

k

k
PSYHB etrds0444.tex

6

k

V2 - 10/22/2018

6:14 P.M.

Page 6

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

it be the case that voters simply do not update their beliefs about candidate
quality even when presented with unequivocal evidence of malfeasance?
When group loyalties, based on partisanship, religion, ethnicity, and so on,
are strong, corruption information of in-group members might not have
any effect on voter’s decision-making. Owing to the fact that corruption
allegations may pose a serious threat to in-group status and the in-group’s
hold on power, people might absolve in-group members of blame. The
most straightforward channel through which this would work is through
people’s in-group serving bias (Solaz et al., 2018). While voters tend to
reward the in-group for positive outcomes, they absolve it of blame for
negative outcomes. This would constitute a form of motivated reasoning
in which people, because of their automatic affective ties to their in-group,
selectively process information. Another channel through which this may
work is through in-group status. Evidence from behavioral economics shows
that people may gain some utility from the utility of others, in particular
from members of their own in-group (Charness & Rabin, 2002). In his work
on redistribution, for example, Shayo (2009) shows that individuals derive
utility from having high-status members in their group. Because of this,
poorer groups might not support redistribution from the rich to the poor
even though it would benefit them personally. If we extend this insight to
corruption-generated wealth, voters could gain some utility from the fact
that a candidate from their in-group stands to gain from corruption even if
they would personally lose out. In this case, voters might actually reward
candidates from their own group for being corrupt.
In a recent study, we started to examine these possible channels by exploiting a real-world corruption scandal in Spain that fell in the middle of a survey
fieldwork period and complementing it with laboratory experiments from
the United Kingdom (Solaz et al., 2018). By comparing Spanish respondents
before and after a corruption scandal was uncovered in a national newspaper,
we show that in-group loyalty based on partisanship decreased the likelihood of withdrawing support for the ruling party due to corruption, while
out-group identification increased it. In a next step, we replicated our results
in a laboratory setting that allowed us to isolate in-group effects more carefully by controlling the precise setting and information voters they received
as well as randomizing the group identity of the candidates that voters faced.
The laboratory results support our findings from the Spanish case: voters fail
to punish in-group candidates they know to be corrupt but punish out-group
candidates for the same indiscretions. These findings are important and lead
to a host of new questions. Which factors could condition the magnitude
of the effects of in-group serving biases? And what kind of group identities
might trigger these kinds of in-group effects, and which ones do not? These
are crucially important questions for future work to address.

k

k

k
PSYHB etrds0444.tex

V2 - 10/22/2018

Corruption and Electoral Accountability: Avenues for Future Research

k

Page 7

7

Another important avenue of future research is the importance of nature of
prior beliefs for understanding the degree to which voters, in fact, update
their evaluations of candidate quality in light of malfeasance revelations.
A recent study by Arias, Balan, Larreguy, Marshall, and Querubin (2017)
suggests that electoral punishment of corruption is rare when voters already
believe that candidates are of low quality. Owing to their negative priors, a
signal that the incumbent is corrupt will not be very informative. In fact, their
evidence stemming from a field experiment conducted in Mexico, where voters were informed about malfeasant municipal spending or not, suggests that
when priors were already negative, support for the incumbent party actually
increased after corruption was revealed because voters positively updated
their priors from information that was better than expected. These findings
suggest that remedying information asymmetries may not always lead to the
results that scholars or practitioners hope for. The evidence also leads to a
host of new questions, for example, which prior beliefs matter, those relating
to the incumbent as such or those relating to the prevalence of corruption
in society at large? And under which conditions do these different types of
priors matter most?
Concerning the latter, a study by Klašnja and Tucker (2013), comparing survey experimental evidence from Sweden and Moldova, for example, shows
that voters in Sweden, a low-corruption country, responded more to information about corruption than voters in Moldova, a high-corruption country,
do. What is more, voters in Moldova only responded to corruption information when economic conditions were poor. This perhaps suggests that while
on average due to their priors about the prevalence of corruption, voters in
high-corruption countries may not respond to one additional case of corruption, some heterogeneity in reactions most likely exists. This heterogeneity
could be the result of the overall state of the economy as the study by Klašnja
and Tucker suggests or individuals’ prior beliefs about candidate quality as
highlighted by Arias et al. (2017), but it could also be due to other prior beliefs
people might hold.
We think that people’s social priors might be especially important. Social
priors refer to people’s beliefs about society at large. The role of three sets
of social priors seem especially important in our view, people’s beliefs about
(i) the prevalence of corruption within society, (ii) the way other voters faced with
corruption allegations will act and (iii) the quality of the pool of future political
entrants. These social priors are likely going to condition the extent to which
voters will actually update their beliefs about candidate quality and blame
the incumbent for corrupt activities. The pervasiveness of these prior beliefs,
especially in high-corruption countries, may make it less likely that candidate
corruption will be punished at the ballot box. Let us address each of them
in turn.

k

6:14 P.M.

k

k
PSYHB etrds0444.tex

8

k

V2 - 10/22/2018

6:14 P.M.

Page 8

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

When people already think that corruption is widespread, they do not
punish corruption at the ballot box. This could be due to different reasons.
It might be due to the fact that when corruption is believed to be prevalent,
corruption allegations add little new information to voters’ decision-making
and therefore might carry very little weight in voters’ minds. This type
of reasoning is consistent with arguments that citizens residing in countries with higher quality of governance might display greater sensitivity to
accountability failures (Ashworth, Bueno de Mesquita, & Friedenberg, 2017).
It could also be the case that when corruption is more widespread, engaging
in corrupt practices might be deemed more acceptable by voters (Fisman
& Golden, 2017). Yet, it could also be that if voters think that corruption is
widespread, they also believe that it takes a corrupt politician to navigate the
system and to get things done (Bauhr & Charron, 2018). A recent study based
on conjoint experiments that was embedded in surveys in Argentina, Chile,
and Uruguay shows no effect of reminding people about the prevalence
of corruption in their country (Klašnja, Lupu, & Tucker, 2018). Yet, as the
authors themselves admit, empirically establishing the effect of people’s
perceptions about the pervasiveness of corruption is difficult, in part because
people’s understandings of what corruption might actually entail are likely
going to differ substantially between contexts in which corruption is more
or less widespread, and among individuals. Studying how people’s beliefs
about the prevalence of corruption matter is an important area for future
research.
Second, not only people’s perceptions about the extent of corruption in
society will likely affect their willingness to hold candidates to account
for corruption, but also their expectations about the behavior of others.
People’s prior beliefs about the behavior of other voters relate to the
importance of voter coordination in the punishment of corruption. When
people think that others are unlikely to punish existing politicians for
corruption, they are unlikely to do so as well (especially when there is
any form of side-benefit that they could lose by withdrawing support
from a corrupt incumbent). In order for voter coordination to occur, two
conditions need to be met: first, voters have to agree that corruption is
important for their vote, and second, they need to be able to agree to vote
for another clean candidate. A study by Chang et al. (2010) highlights the
importance of voter coordination. It is based on legislators in Italy’s lower
house between 1948 and 1994 and candidates of the two largest parties
(Christian Democracy and the Italian Communist Party). The authors show
that only in the 1992–1994 legislature did corrupt legislators face a serious
electoral penalty. They suggest that this structural break in the response to
corruption by Italian voters was due to the media coverage that grew out of
the Clean Hands operation in the early 1990s. The authors argue that a wider

k

k

k
PSYHB etrds0444.tex

V2 - 10/22/2018

Corruption and Electoral Accountability: Avenues for Future Research

k

Page 9

9

dissemination of information about corruption in the media not only makes
corruption more salient in voters’ minds but also provides a coordination
signal to voters, who realize that others will most likely punish corruption
as well.
Several recent studies examined the role of coordination signals through
different types of experimental treatments and found mixed results (Adida,
Gottlieb, Kramon, & McClendon, 2017; Arias, Larreguy, Marshall, &
Querubin, 2018; Yap, 2017). Citing evidence from a field experiment conducted in the context of legislative elections in Benin, Adida and colleagues
show that voters only punished bad performance when information was
accompanied with a message highlighting the importance of legislative performance for voter welfare and the fact that other villages in the constituency
received the same information. Laboratory evidence from Australia, Singapore, and the United States also highlights the importance of positive
coordination signals (Yap, 2017). Yet, evidence from a field experiment
revealing mayoral malfeasance before the 2015 Mexican municipal elections
suggests that accompanying leaflets detailing results of audit reports by
loudspeakers in order to signal to voters that others received the malfeasance
information as well did not significantly alter the effects of information on
incumbent vote shares (Arias et al., 2018). All in all, these results highlight
that understanding (i) which coordination signals matter and (ii) which
conditions may hamper their effects are two important areas for future
research.
Finally, voters’ beliefs about the quality of the pool of future political
entrants is likely important for their willingness to hold candidates to
account for corruption. When people think that the pool of possible political
entrants will also be corrupt, there is little reason to punish an existing
politician for corruption. In their theoretical work on corruption traps,
Klašnja, Little, et al. (2016) highlighted that overcoming corruption involves
coordination not only among voters and existing candidates but also among
possible entrants. Voters’ perceptions about the quality of the pool of
possible entrants will inform their behavior. When voters’ expect an alternative pool of candidates for office to be likely as corrupt as current ones,
why would they cast a ballot based on corruption? The pool of potential
candidates has been found to be important in other aspects of political
behavior as well, such as protest behavior (Meirowitz and Tucker, 2013).
When voters become informed about widespread corruption of incumbents,
this might influence their beliefs about the overall average quality of the
pool of potential governments, especially if they have little experience with
elections to begin with. While in established democracies, learning about
corruption may lead people to think that they should root out this one bad
apple, in newly established democracies this may not spark off pessimism

k

6:14 P.M.

k

k
PSYHB etrds0444.tex

10

V2 - 10/22/2018

6:14 P.M. Page 10

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

about all possible political candidates. Empirical examinations regarding
(i) the way in which people’s expectations about the quality of the pool of
possible alternatives affect the likelihood of punishing corruption, and (ii)
how these relationships differ based on people’s experience with democracy
constitute important avenues for future research.
CONCLUSION

k

In recent years, we have experienced somewhat of a boom in examining the
extent to which people hold politicians to account for corruption, not least
due to the widespread use of field, laboratory, and survey experiments. These
experimental tools help researchers to overcome some of the core problems
associated with studying corruption, such as the fact that politicians try to
hide malfeasance and people may not want to admit that they support corrupt politicians. Much of this recent work has centered on the importance of
corruption information and the willingness of voters, once they are informed
about corruption, to trade it off against other benefits. In this essay, we have
highlighted that the role of causal attribution has thus far received far less
scholarly attention. Based on a handful of recent studies, we suggested that
corruption might go unpunished at the ballot box simply because voters do
not update their beliefs about candidate quality in light of unambiguous
information about malfeasance. Recent work suggests that this is largely the
result of the weight voters attach to in-group status and their prior beliefs
about candidate quality. Yet, we do not know under which conditions these
factors are important, and if other beliefs matter as well. Moreover, most of
this early experimental work stems from one country context at one point in
time, so we do not know if we can generalize existing findings.
Overall, we have highlighted two key avenues for future research. First,
what is the role of role of attribution errors? To what extent do people fail to
blame their in-group candidate for corruption, and punish out-group candidates for the same indiscretion? Which factors may condition the magnitude
of these in-group effects? What is more, which kinds of group identities may
trigger these effects, and which ones do not? These are crucially important
questions for future work to address. Second, what is the role of people’s
beliefs about society at large? To what extent does the electoral punishment
of corruption hinge on people’s priors about the prevalence of corruption
within society, the way they expect other voters to act, and the quality of the
pool of future politicians? We believe that these social priors are crucial for
understanding the conditions under which corruption gets punished at the
ballot box. The extent to which they actually do, is an important topic for
future research.

k

k

k
PSYHB etrds0444.tex

V2 - 10/22/2018

Corruption and Electoral Accountability: Avenues for Future Research

6:14 P.M. Page 11

11

REFERENCES

k

Adida, C., Gottlieb, J., Kramon, E., & McClendon, G. (2017, November). Breaking the
clientelistic voting equilibrium: The joint importance of salience and coordination (Working Paper 48). Washington, DC: AIDDATA.
Anduiza, E., Gallego, A., & Munoz, J. (2013). Turning a blind eye: Experimental evidence of partisan bias in attitudes toward corruption. Comparative Political Studies,
46(12), 1664–1692.
Arias, E., Balan, P., Larreguy, H., Marshall, J., & Querubin, P. (2017). Priors rule: When
malfeasance revelations help and hurt incumbent parties? (Working Paper). Cambridge,
MA: Harvard University.
Arias, E., Larreguy, H., Marshall, J., & Querubin, P. (2018). Does the content and mode
of delivery of information matter for electoral accountability? Evidence from a field experiment in Mexico (Working Paper). Cambridge, MA: Harvard University.
Ashworth, S., Bueno de Mesquita, E., & Friedenberg, A. (2017). Accountability and
information in elections. American Economic Journal: Microeconomics, 9(2), 95–138.
Bauhr, M., & Charron, N. (2018). Insider or outsider? Grand corruption and electoral
accountability. Comparative Political Studies, 51(4), 415–446.
Besley, T. (2006). Principled agents? The political economy of good government. Oxford,
UK: Oxford University Press.
Botero, S., Cornejo, R. C., Gamboa, L., Pavão, N., & Nickerson, D. W. (2015). Says
who? An experiment on allegations of corruption and credibility of sources. Political Research Quarterly, 68, 493–504.
Chang, E. C., Golden, M. A., & Hill, S. J. (2010). Legislative malfeasance and political
accountability. World Politics, 62(2), 177–220.
Charness, G., & Rabin, M. (2002). Understanding social preferences with simple tests.
Quarterly Journal of Economics, 117(3), 817–869.
De Vries, C. E., & Solaz, H. (2017). The electoral consequences of corruption. Annual
Review of Political Science, 20, 391–408.
Fernández-Vázquez, P., Barberá, P., & Rivero, G. (2016). Rooting out corruption or
rooting for corruption? The heterogeneous electoral consequences of scandals.
Political Science Research and Methods, 4(2), 379–397.
Ferraz, C., & Finan, F. (2008). Exposing corrupt politicians: The effects of Brazil’s publicly released audits on electoral outcomes. Quarterly Journal of Economics, 123(2),
703–745.
Fiorina, M. P. (1981). Retrospective voting in American national elections. New Haven,
CT: Yale University Press.
Fisman, R., & Golden, M. (2017). Corruption: What everyone needs to know. Oxford:
Oxford University Press.
Healy, A., & Malhotra, N. (2013). Retrospective voting reconsidered. Annual Review
of Political Science, 16, 285–306.
Klašnja, M. (2016). Increasing rents and incumbency disadvantage. Journal of Theoretical Politics, 28(2), 225–265.
Klašnja, M., Little, A., & Tucker, J. A. (2016). Political corruption traps. Political Science
Research and Methods, 1–16.

k

k

k
PSYHB etrds0444.tex

12

k

V2 - 10/22/2018

6:14 P.M. Page 12

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Klašnja, M., Lupu, N., & Tucker, J. A. (2018). When do voters sanction corrupt politicians?
(Working Paper). Washington, D.C: University of Georgetown.
Klašnja, M., & Tucker, J. A. (2013). The economy, corruption, and the vote: Evidence
from experiments in Sweden and Moldova. Electoral Studies, 32(3), 536–543.
Klašnja, M., Tucker, J. A., & Deegan-Krause, K. (2016). Pocketbook vs. sociotropic
corruption voting. British Journal of Political Science, 46(1), 67–94.
Meirowitz, A., & Tucker, J. A. (2013). People power or a one-shot deal? A dynamic
model of protest. American Journal of Political Science, 57(2), 478–490.
Pettrigrew, T. F. (1979). The ultimate attribution error: Extending allport’s cognitive
analysis of prejudice. Personality and Social Psychology Bulletin, 5(4), 461–476.
Rundquist, B. S., Strom, G. S., & Peters, J. G. (1977). Corrupt politicians and their electoral support: Some experimental observations. American Political Science Review,
71(3), 954–963.
Shayo, M. (2009). A model of social identity with an application to political economy:
Nation, class, and redistribution. American Political Science Review, 103(2), 147–174.
Solaz, H., De Vries, C. E., & De Geus, R. (2018). In-group loyalty and the punishment
of corruption (Forthcoming). Comparative Political Studies.
Weitz-Shapiro, R., & Winters, M. S. (2017). Can citizens discern? Information credibility, political sophistication, and the punishment of corruption in Brazil. The Journal
of Politics, 79(1), 60–74.
Winters, M. S., & Weitz-Shapiro, R. (2013). Lacking information or condoning corruption: When do voters support corrupt politicians? Comparative Politics, 45(4),
418–436.
Yap, O. F. (2017). When do citizens take costly action against government corruption?
Evidence from experiments in Australia, Singapore, and the United States Journal
of East Asian Studies, 17(1), 119–136.

Catherine E. De Vries is a Professor of Political Behavior in Europe in the
Department of Political Science and Public Administration at the Vrije Universiteit Amsterdam. Her research spans a range of topics concerning public
opinion, political behavior and political corruption in the European context.
Hector Solaz is a lecturer at the John Stuart Mill College of the Vrije Universiteit Amsterdam. His research interests lie within behavioral science and
political economy. He uses lab experimental methods to study the foundations and implications of political and economic decision-making.
RELATED ESSAYS
Economic Models of Voting (Political Science), Ian G. Anson and Timothy
Hellwig
Electoral Authoritarianism (Political Science), Andreas Schedler
Presidential Power (Political Science), William G. Howell
Political Inequality (Sociology), Jeff Manza
Money in Politics (Political Science), Jeffrey Milyo

k

k

k
PSYHB etrds0444.tex

V2 - 10/22/2018

Corruption and Electoral Accountability: Avenues for Future Research

6:14 P.M. Page 13

13

Information Politics in Dictatorships (Political Science), Jeremy L. Wallace
Lawmaking (Political Science), Jamie L. Carson and Mark E. Owens
News Framing Effects and Emotions (Political Science), Andreas R. T. Schuck
and Alina Feinholdt
Political Advertising (Political Science), Erika Franklin Fowler
The Politics of Disaster Relief (Political Science), Alexander J. Oliver and
Andrew Reeves
Racial Disenfranchisement (Political Science), Vincent L. Hutchings and
Davin L. Phoenix
Perceptions of the Legitimacy of Collective Decisions (Political Science),John
W. Patty
Partisan News: A Perspective from Economics (Economics), Daniel F. Stone

k

k

k



Media

Corruption and Electoral Accountability: Avenues for Future Research