Political Corruption and Voter Turnout: Mobilization or Disaffection? Elena Costas-Pérez Universitat de Barcelona & Institut d’Economia de Barcelona (IEB) Work in progress. March 2014. ABSTRACT: Corruption scandals may modify voter turnout, either by mobilizing citizens to go to the polls to punish or support the malfeasant politician or by demotivating individuals to vote as a consequence of disaffection with the democratic process. We study whether these effects depend on individual’s partisan leanings and/or the timing of corruption scandals. Our database includes information on Spanish local scandals from 1999 to 2007, and survey data on individuals’ turnout at the 2007 local elections. We use a matched database to identify the corruption-free pairs for our corruption affected municipalities’ sample. Our results show that while neither past nor recent corruption scandals have impact on turnout, repeated corruption cases boost abstentionism. We also find that independent voters - those with no attachment to any political party - are the only ones that withdraw from elections as a consequence of corruption. Core supporters do not modify their electoral participation after a scandal has broken out. Those who support the incumbent do not even recognise that their party is corrupt, while both independent voters and opposition core supporters report higher levels of corruption perceptions once a scandal is revealed. Keywords: electoral turnout, accountability, corruption JEL Classification: P16, D72, D73 Contact address: Faculty of Economics and Business University of Barcelona Av. Diagonal 690, torre4, planta 2ona 08034 Barcelona Telf.: 00 34 93 403 90 12 e-mail: [email protected]
35
Embed
Political Corruption and Voter Turnout: Mobilization or ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Political Corruption and Voter Turnout:
Mobilization or Disaffection?
Elena Costas-Pérez
Universitat de Barcelona &
Institut d’Economia de Barcelona (IEB)
Work in progress. March 2014.
ABSTRACT: Corruption scandals may modify voter turnout, either by
mobilizing citizens to go to the polls to punish or support the malfeasant
politician or by demotivating individuals to vote as a consequence of
disaffection with the democratic process. We study whether these
effects depend on individual’s partisan leanings and/or the timing of
corruption scandals. Our database includes information on Spanish local
scandals from 1999 to 2007, and survey data on individuals’ turnout at
the 2007 local elections. We use a matched database to identify the
corruption-free pairs for our corruption affected municipalities’ sample.
Our results show that while neither past nor recent corruption scandals
have impact on turnout, repeated corruption cases boost abstentionism.
We also find that independent voters - those with no attachment to any
political party - are the only ones that withdraw from elections as a
consequence of corruption. Core supporters do not modify their
electoral participation after a scandal has broken out. Those who
support the incumbent do not even recognise that their party is corrupt,
while both independent voters and opposition core supporters report
higher levels of corruption perceptions once a scandal is revealed.
2. Corruption’s mobilization and disaffection effects
2.1. Previous literature
Neither theoretical nor empirical studies have arrived at unified conclusions on the
relationship between corruption and voter turnout3. An important shortcoming of
existing literature is the lack of an empirical strategy designed to identify the effects
once a scandal breaks out, i.e., an increase in electoral participation through a
mobilization of voters to the polls or a decrease of electoral participation as a result of
their disaffection with the democratic process. Kostadinova’s (2009) study on post-
communist transitional countries tries to identify both a direct –mobilization- and an
indirect –disaffection- effect of corruption on turnout. However, she considers
corruption perceptions, which may be correlated to voting decisions, casting doubt on
the model’s overall exogeneity.
It has been proven that good governance is related to citizens’ capacity to hold their
politicians accountable (Adsera, et al., 2003). Hence, if we understand elections as an
effective accountability tool, individuals who feel betrayed by a corrupt politician may
cast a ballot to bring him or her out of power. In such a case, corruption would act as a
mobilization factor. That mobilization effect would make some citizens, who would
otherwise abstain, go to the polls to punish the accused politician. After a scandal, party
3 In most cases empirical evidence confirms a negative relationship between corruption and turnout
(Domínguez and McCann, 1998; Kostadinova, 2009; Birch, 2010; Chong et al., 2012). In contrast, a
scholarly minority (Karahan et al., 2006; Escaleras et al., 2012) attributes to corruption a positive effect
on turnout. Finally, some studies do not find any relationship between corruption scandals and voter
turnout (Stockemer, 2013).
5
members and sympathisers could also be called upon to mobilize to give their vote to
the accused politician, either by thinking the allegations are false or by giving
unconditional support. An alternative case, in the context of highly corrupt situations
with clientelistic networks, is that scandals could stimulate turnout if corrupt politicians
were to buy voters to retain their power (Karahan et al., 2006).
Conversely, corruption can also be a drain on voter turnout (Putnam, 1993; Warren,
2004; Chang and Chu, 2006). Corruption harms citizens’ trust in local and national
politicians, generating cynicism and voter apathy (Solé-Ollé and Sorribas-Navarro,
2013). That disaffection effect would make individuals less likely to vote for existing
corrupt political parties (Warren, 2004; Wagner et al., 2009). If corruption is repeated,
in the long run disaffected individuals may decide to disengage from the electoral
system (Chong et al., 2012). If voters consider that corruption is widespread, replacing
the corrupt incumbent for a new one will not fix the situation. In the most extreme
cases, widespread corruption might also cast doubt on the sustainability of the
democratic system (Kostadinova, 2009). Empirical evidence has confirmed this
negative relationship between corruption and turnout (Domínguez and McCann, 1998;
Andersen and Tverdova, 2003; Kostadinova, 2009; Stockemer, 2013; and Stockemer et
al., 2013).
2.2. Who and why will be affected by corruption?
As stated above, most existing studies are unable to differentiate among the relative
strengths of the mobilization and disaffection effects that actual corruption cases may
cause. We argue that the literature on corruption and voter turnout does not point
conclusively to voters’ reactions because two crucial factors are excluded from these
analyses: the role of individual partisan leanings and the timing of corruption.
First, we consider that corruption’s mobilization and disaffection effects will differ
depending on voters’ degree of partisan affinity with the incumbent involved. Similar to
our strategy, Chong et al. (2012) take a step forward over previous studies, proving with
experimental evidence for Mexico that exposing citizens to information on corruption
not only decreases voter turnout, but also negatively affects voters’ identification with
the corrupt incumbent's party. They find that providing information about high levels of
corruption has a bigger impact on challengers' votes than on incumbents'; however, their
data does not allow them to analyse if voters with a different degree of partisanship
respond differently to corruption.
Independent voters, those who occasionally vote or do not always vote for the same
party, tend to be more independent from partisan attachments, and they can be more
affected by shocks like occasional corruption (Rundquist et al., 1977; Feddersen and
Pesendorfer, 1996; Sobbrio and Navarra, 2010; Stockemer, 2013). Thus, the
6
disaffection effect will be higher for those, so we would expect to observe independent
voters withdrawing from the elections if corruption scandals occur. Hence, the first
hypothesis regarding partisan leaning is:
H1.a: Independent voters, defined as those who do not always vote for the same party,
will be more likely to abstain if a corruption scandal is revealed.
On the contrary, core supporters have stronger partisan leanings and are less likely to
defect (Chong et al., 2012). If partisan leanings are strong, citizens may disregard
corruption as a determining factor in their decision and continue to vote for the party to
which they are ideologically aligned (Peters and Welch, 1980; and Anderson and
Tverdova, 2003). Ideology or party identification can also modify how voters evaluate
corruption, depending on the party to which the corrupt incumbent belongs. Supporting
this hypothesis, recent research of Anduiza et al. (2012) finds that individuals have a
partisan bias and are more tolerant to corruption if the politician involved belongs to
their own party. If that is true, we should find core supporters of corrupt incumbents
unaffected by corruption scandals when deciding to vote. Thus, the hypothesis to test is:
H1.b: Incumbent core supporters will not modify their electoral participation decision
after an incumbent’s corruption scandal is revealed.
Considering core supporters of opposition parties, we would expect that, even realising
that corruption exists, they will keep voting for their party options. It may be even
possible that, in the presence of corruption, some opposition core supporters who would
otherwise abstain would go to the polls to defeat the corrupt incumbent. Nevertheless,
people who identify themselves with a political party are more likely to vote (Norris,
2004), so it is improbable that we will find that mobilization effect. Regarding partisan
leanings, the last hypothesis is:
H1.c: Opposition core supporters will not modify their electoral participation decision
after an incumbent’s corruption scandal is revealed.
Second, we consider that scandals occurring at different points in time may modify the
influence of corruption’s mobilization or disaffection effect on individual participation
decision. There is evidence that voters tend to overweigh the information received
closer to elections, as recent events have greater influence on evaluations of
incumbents’ performance (Fair 1978; Kramer 1971). Another explanation for the fact
that later scandals could have greater effects on turnout is that voters are more attentive
to indicators of incumbents’ outcomes as an election approaches (Valentino and Sears
1998), being unable to easily recollect information on earlier performance (i.e., past
7
corruption cases) (Huber et al. 2012). Also, old corruption cases might be difficult for
voters to remember.
Thus, while in the short term corruption can mobilize voters to bring down corrupt
governments, that relationship can be reversed if corruption continues over time. This
hypothesis is based on the fact that the persistence of corruption can differently affect
citizens’ trust in the political system. Corruption cases repeated over several years can
make citizens cast doubt on the democratic system’s capacity to keep politicians
accountable (Kostadinova, 2009). That distrust in the political system leads to
disaffection and alienation from politics, which may result in the decision to withdraw
from the electoral process, i.e., abstention. In that scenario, repeated corruption cases
will then generate the disaffection mechanism through which corruption scandals lessen
voter turnout.
Hence, the hypothesis regarding the effects of the timing of corruption cases on voters’
turnout is:
H2: Past corruption cases will not affect turnout, while recent scandals would either
have no effect or would mobilize people to vote. Repeated corruption in time will lessen
voter turnout as a result of the disaffection effect.
In conducting our study, we analyse these hypotheses, both independently and
considered together. We predict that independent voters - those more susceptible to
defection - will be more sensitive to repeated corruption scandals, which will further
harm their trust in the system. Regarding core supporters, it is difficult to speculate how
they will react to corruption occurring at different times, since we predict that they will
be less likely to modify their electoral participation as a consequence of corruption.
3. Data and Empirical Analysis
3.1. Data and typology of corruption scandals
In order to carry out our analysis, we use a novel database that includes information on
local corruption scandals in Spain and data on a survey of voting behaviour in Spanish
municipalities. We define a local corruption scandal as the “public allegation of
corruption brought to light by a newspaper”. Our data about corruption cases is based
on a report compiled by the Fundación Alternativas (2007). After a wave of local
corruption scandals starting in the first years of the 2000s, that Spanish think-tank hired
several journalists to gather all corruption-related stories published in national, regional
and local media between January 2000 and January 2007. However, the time period we
are interested in goes from the local elections in July 1999 to the ones in May 2007. For
that reason, we completed the Fundación Alternativas’s information with an internet-
8
guided search4 on news on corruption scandals. We ended up with a total of 565
municipalities affected by corruption scandals during that period5. We also checked for
the non-partisan bias of our news, comparing our data with other corruption maps
compiled by media outlets from different political ideologies. The percentage of
corruption cases by political party was not statistically different in all databases,
verifying that our compilation of cases was not ideologically biased.
It is important to bear in mind that during the first Spanish democratic governments
(1979-1999) there were no exceedingly significant local corruption cases reflected in
Spanish media (Jiménez and Caínzos, 2006). The corruption cases we are studying are
related to land use regulations, one of the most important local corruption’s typologies
in Spain during the peak of the housing boom. Scandals involved local politicians
receiving bribes in return to changes over the land use plans already approved (i.e.,
reclassifying public land). Municipal governments in Spain are responsible for the land
use regulation, making easier to identify the effect of those scandals on electoral
outcomes. In that specific case, voters can clearly identify the incumbent as guilty for
the land-use related corruption.
The wave of corruption cases rose significantly in the late nineties, when Spanish media
started highlighting numerous corruption scandals and the judiciary also began the
investigation of some of them. The number of corruption cases shot up after 1999
(Costas-Pérez et al., 2012), peaking before the 2007 local elections. That distribution of
scandals makes the Spanish situation the optimal context to test our hypothesis that
corruption occurring at different points in time will have a different effect on citizens’
voting behaviour.
Our database accounts for 122 municipalities affected by corruption in the period from
June 1999 to May 2007. We have classified them among three different sub-categories
regarding corruption persistence that we will use throughout our analysis. First, 32
municipalities of our database experienced at least one corruption scandal in the term
1999-2003, but not corruption cases broke out afterwards. We have considered them as
‘past corruption cases’, since in the 2007 elections voters may have a faint memory
about these scandals. Second, 58 municipalities had at least one corruption scandal in
the term 2003-2007, but not corruption cases had broke out in the previous term. These
are classified as ‘recent corruption cases’ from the perspective of an individual facing
the 2007 local election. Finally, we have also considered those places that have
4 We used a paid digital information management service covering all national and many of the regional
newspapers, MyNews, until November 2009. Thus, we have an additional sample of corruption cases
occurring between the local elections of 2007 and November 2009 that we save to perform a placebo test
explained in section 4.6. 5 See Costas-Pérez et al. (2012) for more information on the construction of the corruption database.
9
experience repeated corruption cases, both in the 1999-2003 and the 2003-2007 terms.
32 municipalities are classified in the category of ’repeated corruption cases’.
Hence, our corruption database indicates whether there has been at least one corruption
scandal between June 1999 and May 2007 (the two terms of office analysed). Since our
objective is to measure the impact of corruption, we need a sample of individuals from
corruption-free municipalities to be compared with those from localities affected by
scandals. The fact of using a matched set of municipalities allows us to balance the
distribution of covariates between corruption-ridden and corruption-free municipalities,
to avoid biased estimations.
3.2. The matching strategy
In order to construct our analysis’ sample we use a matched database that identifies for
the selected municipalities affected by corruption a valid control group. Budget
limitations implied a restriction for the number of municipalities analysed in the survey.
Hence, a matching procedure was followed to select those corruption-free
municipalities to be compared with the corruption-ridden ones (our control and
treatment groups, respectively). Over the 565 municipalities that experienced at least
one corruption scandal between 1999 and 2007, a sample of 122 municipalities was
randomly select – stratified according to population size and term of corruption -. Thus,
old and new corruption cases are equally represented in the final database.
The municipalities that potentially belong to the control group are those with similar
characteristics to the corruption-ridden ones but where no scandals broke out. In order
to identify our control group a Propensity Score Matching approach was used6.
In order to construct the ‘propensity score’ a Logit model7 was estimated, using as a
dependent variable a dummy equal to one if a corruption scandal broke out in the
municipality, and zero otherwise. The Logit equation was estimated with data from all
the corruption-ridden and corruption-free municipalities. The municipal-level variables
used to implement the matching strategy were8: historical average levels of aggregate
turnout, population size, unemployment, ethnic diversity, income, school attendance,
divorce rate, and historical average of right voters9.
6 The full procedure and estimations results of Solé-Ollé and Sorribas-Navarro’s (2013) matching
strategy, as well as additional data checks, are available upon request. 7 Because of data limitations to carry out the matching we were limited to municipalities larger than 1.000
habitants. 8 Descriptive statistics of these data can be found in Table A.1. in the Appendix.
9 Two alternative estimations were also considered, including some additional covariates with low
explanatory power (i.e., percentage of elderly population, historical average of electoral margin of
victory, coalition governments, percentage of residents born in that municipality, newspapers per capita
or associations per capita), and also higher order terms of the main covariates and interactions amongst
them. Nevertheless, the balance did not improve in either case, so the first option was maintained.
10
With the Logit estimation the ‘propensity score’ was computed and control
municipalities were matched to the treatment ones based on having a similar value of
the ‘propensity score’. As the matching algorithm the ‘Nearest Neighbour with
replacement’10
was use. This method allows for more than one given control unit
matching more than one treatment unit, which increases the average quality of matching
and reduces the bias11
.
The matching strategy’s implementation allows to balance the covariates in the two
subsamples. We ended up with a control group of 97 municipalities that did not
experience any corruption scandal between 1999 and 2007. These, plus the 122 treated
municipalities where at least one corruption scandal broke out during the same period,
constitute the 219 municipalities included in our database12
.
To confirm that the sample of treatment and control municipalities used in this paper
was a good matching we conducted different tests. We first analysed the percentage
reduction in the standardised bias as the result of the matching procedure, finding a
considerable decrease that showed a statistically significant bias before the matching
(i.e., school attendance (66% drop) or divorce rate (99% drop)). Second, we performed
a comparison of means between treated and control units in the unmatched and matched
samples (see Rosenbaum and Rubin, 1985). Table A.2. in the Appendix shows the
means of each groups for all variables considered to perform our matching. The last
column of the table reports the test and p-values of the differences in means between the
treated and the control group. Once our sample is matched those difference are not
statistically significant anymore. Third, we re-calculated the propensity score on the
matched sample and compare the pseudo-R2 before and after matching
13.
We also performed a difference in means test for the individual level variables used in
our analysis, using the survey observations as the treatment and control groups. The
results of this test verify that interviewees from our treatment and control groups not
only live in very similar municipalities, but also share the same individual traits. Table
A.3. shows that, using our sample individual data, we arrive to the same conclusions
about our matching quality. For that reason we consider that matching at the individual
10
Other matching options were considered (e.g., ‘without replacement’). None of them worked that well
for some of the biggest municipalities, so they were not used. 11
The Nearest Neighbour matching may generate bad matches (i.e., the distance to the nearest neighbour
is too large). However, 95% of the matches had an absolute distance in the ‘propensity score’ lower than
0.01, so it can be considered that a considerably good matching was achieved. 12
Solé-Ollé and Sorribas-Navarro’s (2013) originally selected a sample of 160 treatment municipalities
and 130 controls, including scandals between 1999 and 2009. For the specific propose of this paper we
restricted the corruption cases to those occurring before the 2007 local elections, and that is the matched
sub-sample data that we use. All tests to verify that we have achieved a good matching have been done to
both samples. 13
They were 0.237 and 0.002, respectively. LR tests of joint significance of the regressors before and
after the matching have values of 1871.77 and 2.32, with p-values of 0.000 and 0.941.
11
level is not necessarily in our case since the citizens interviewed in the treatment and
control municipalities are already very similar.
After performing all these tests we can confirm the matching strategy successfully
created balance in baseline characteristics across our treatment and control groups, both
for the municipalities and the individuals analysed. An additional advantage of the
matching procedure is that it assures a complete transparency and predetermination of
our research design. Since the matching algorithm must be applied before the estimation
of the treatment, the decisions taken at this stage are no influenced by any information
on the estimation results (Ho et al., 2007).
3.3. Data on individual turnout and corruption perceptions
This paper makes use of a special survey designed to be conducted in the selected
matched municipalities14
. The survey was undertaken in November of 2009 and to
obtain an indicator of individual electoral turnout interviewees were asked if they voted
in the 2007 local elections or not15
.
As Table A.1 shows, our sample average turnout is a bit higher than the actual one16
.
Previous papers have also suffered from this "overreporting" bias, explained both by the
vote misreporting of actual non-voters among survey respondents and the
overrepresentation of actual voters (Traugott, 1989). However, several studies have
proved that the overreporting problem has no real effect on the actual implications of
the models’ estimations that try to understand the factors that may influence voting and
abstention (Hillygus, 2003). Also, recent research demonstrates that participation in
surveys does not increase vote likelihood (Mann, 2005). Thus, we are confident about
the implications drawn from our estimations’ results.
The survey also included a question on individuals’ corruption perceptions: ‘Which do
you consider is the degree of corruption in your local government?’. The interviewees
could answer one of the following five categories: 5 “very high corruption”, 4 “high”, 3
“medium”, 2 “low”, and 1 “none”17
.
Among other socioeconomic characteristics, the survey included questions regarding
political preferences (e.g., partisan attachment and ideology), and information on a
series of socio-economic controls (e.g., unemployed, type of job, marital status, etc.).
14
The questionnaire used by Solé-Ollé and Sorribas-Navarro’s (2013) is available upon request. 15
Individuals who did not vote in the 2007 local elections because they were too young or they were not
registered in that municipality are excluded from our analysis. 16
The actual voter turnout in the 2007 Spanish local elections for the municipalities analysed was 68.9%.
More information on Spanish electoral outcomes can be found in: http://www.infoelectoral.mir.es/min/ 17
There were also two additional categories: 98 “do not know” and 99 “do not answer” that we do not use
Independent Dummy variable coded 1 for people who do not always
vote for the same party 0,512 0,500 0 1
Incumbent core
Supporter
Dummy variable coded 1 for people who always vote for
the same party, and their ideology is the same as the
incumbent's
0,215 0,411 0 1
Opposition
core Supporter
Dummy variable coded 1 for people who always vote for
the same party, and their ideology is the same as the
opposition' party
0,272 0,445 0 1
Contextual-level variables
Corruption Dummy variable coded 1 for municipalities with at least
one corruption scandal in the period 1999-2007 0,482 0,500 0 1
Past
corruption
cases
Dummy variable coded 1 for municipalities with at least
one corruption scandal in the period 1999-2003, but were
not corruption has broke out afterwards
0,110 0,313 0 1
Recent
corruption
cases
Dummy variable coded 1 for municipalities with at least
one corruption scandal in the period 2003-2007, but were
not corruption has broke out before
0,188 0,390 0 1
Repeated
corruption
cases
Dummy variable coded 1 for municipalities with at
corruption scandal in both periods: 1999-2003 and 2007-
2009.
0,185 0,388 0 1
Voter turnout Average voter turnout at the 1987, 1991 and 1995 local
elections 0,687 0,084 0,508 0,922
Income p.c.
Average socio-economic condition. Arithmetic average of
the socio-economic condition according to their
employment status
0,951 0,118 0,610 1,200
Divorced Percentage of divorced and separated among all
population 0,029 0,012 0,002 0,074
Graduate
Percentage of population with third level studies (diploma,
degree and doctorate) among population 16 years and
older
0,128 0,069 0,016 0,434
Unemployment Percentage of unemployed among individuals aged 20-59 0,157 0,079 0,048 0,675
Ethnic diversity
1- Σk(Popk/Population)2 where Pop_contk is population
whose nationality is from continent k, and k refers to
Europe, Africa, America and others
0,058 0,058 0 0,279
Right voters Average historical vote share that the right wing parties
obtained in 1979, 1982, 1986 and 1989 local elections 0,226 0,093 0,037 0,483
log(Population) Log of the registered population 10,545 1,778 6,973 14,957 Notes: (1) Source of the individual-level variables: own-designed survey (see Box A.1). (2) the contextual-level variables: (i)
2001 Census of Population (National Institute of www.ine.es), for Income p.c., % Divorced, % Graduate, % Unemployed,
population by continent used to construct the Ethnic diversity index, and Population. (ii) Database on corruption scandals,
construct (see section 3 for more
details). (iii) Voting data from the Ministry of the Interior, used for the construction of the % Right voters and % Vote turnout variables.
34
Table A.2: Differences in means between Treated and Control groups.
Mean t-test
Treated Control [p-value]
Unmatched sample
% Vote turnout 0.541 0.654 4.25 [0.000]
Income p.c. 0.947 0.939 1.09 [0.282]
% Divorced 0.026 0.018 14.09 [0.000]
% Graduate 0.106 0.077 12.57 [0.000]
% Unemployment 0.147 0.143 0.88 [0.381]
Ethnic diversity 0.060 0.035 10.83 [0.002]
% Right voters 0.507 0.505 0.36 [0.724]
log(Population) 9.610 8.182 27.31 [0.003]
Matched sample
% Vote turnout 0.708 0.692 -1.28 [0.202]
Income p.c. 0.947 0.934 -0.79 [0.428]
% Divorced 0.026 0.026 0.03 [0.976]
% Graduate 0.114 0.104 -1.09 [0.278]
% Unemployment 0.151 0.166 1.23 [0.220]
Ethnic diversity 0.053 0.054 0.08 [0.932]
% Right voters 0.219 0.218 -0.12 [0.908]
log(Population) 9.788 9.678 -0.51 [0.610]
Observations 122 97
Note: (1) Treated group = municipalities where at least one corruption scandal
broke out during the period 1999-2007; Control group=municipalities where
no corruption scandal broke out during the same period.
Table A.3.: Differences in means (survey observations).
Mean t-test
Treated Control [p-value]
Turnout 0.833 0.846 -1.01 [0.220]
Income 2.702 2.721 0.02 [0.479]
Schooling 3.184 3.159 0.03 [0.700]
Age 47.968 47.243 0.72 [0.260]
Female 0.522 0.504 0.02 [0.225]
Divorced 0.044 0.039 0.00 [0.424]
Unemployed 0.130 0.140 -0.01 [0.327]
Students 0.045 0.052 -0.01 [0.222]
Retired 0.241 0.226 0.01 [0.283]
Immigrants 0.039 0.031 0.01 [0.168]
Ideology 1.841 1.817 0.02 [0.324]
Independent 0.510 0.520 -0.01 [0.463]
Incumbent core Supporter 0.224 0.203 1.28 [0.200]
Opposition core Supporter 0.265 0.275 -0.01 [0.507]
Interviewees per municipality 34.779 38.016 -1.18 [0.241]
Number of municipalities 122 97
Note: (1) Treated group = municipalities where at least one corruption scandal
broke out during the period 1999-2007; Control group=municipalities where
no corruption scandal broke out during the same period.