SSRN-id1450260

8/13/2019 SSRN-id1450260

1/31

0

PERCEPTIONS OF POLITICAL PARTY CORRUPTION

AND VOTING BEHAVIOR IN POLAND*

Kazimierz M. Slomczynski

The Ohio State University

Departments of Sociology and Political Science

[email protected]

Goldie Shabad

The Ohio State University

Department of Political [email protected]

*This paper is based on research recently funded by the Polish Ministry of Science and

Higher Education. We would like to thank Pawel Sztabinski, Franciszek Sztabinski, andZbigniew Sawinski for their help in data preparation. We also acknowledge the

Departments of Sociology and Political Science, The Ohio State University for their

partial support of our work on this project.

Paper prepared for presentation at the Annual Meeting of the American Political Science

Association, Toronto, 2009

8/13/2019 SSRN-id1450260

2/31

1

Introduction

More so than many other political institutions, political parties are held in considerable

disrepute. This is due in large part to widespread perceptions that both party politicians and

parties as organizations are corrupt. This is especially the case in the newer democracies of post-

communist Europe, Asia, Latin America, and Africa, but also in long-lived democracies.

(Transparency International Global Corruption Barometer 2004, 2007).

There is an ever growing body of research on the causes and consequences of political

corruption on democratic politics, as well as on the attitudinal and behavioral correlates of

corruption perceptions of the mass public. Notwithstanding the use of different wording of

questions to tap both independent and dependent variables, for the most part findings of the extant

literature based on survey data as well as comparative Transparency International country-level

corruption scores suggest that, all else being equal, actual experience of corruption as well as

perceptions of corruption are associated at the mass level with disengagement from electoral

politics and/or defeat of governing parties, distrust of politicians and formal political institutions

(e.g., the legislature, the judiciary, police), and increased dissatisfaction with the way democracy

is working in their country (Anderson and Tverdova 2003; Bowler and Karp 2004; Canache and

Allison 2005; Sandholtz and Taagepera 2005; Seligson 2002; Shabad, Slomczynski and Zielinski

2008; but see Della Porta 2008; Kistner 2007; and Shabad, Slomczynski and Kistner 2008 for

contrary findings with regard to electoral participation and dissatisfaction with democracy).

For the most part, research on mass-level perceptions of corruption deals with

assessments of overall political corruption or of the extent of malfeasance in a variety of

governmental and political sectors, including political parties. But no survey we are aware of

focuses on the publics perceptions of the extent of corruption ofspecificpolitical parties. This is

regrettable since perceived corruption of politically significant parties could potentially influence

vote choice.

8/13/2019 SSRN-id1450260

3/31

2

Our paper seeks to address this gap within the context of Poland, a new democracy in

which political party corruption has been rife and a focal point of partisan competition. In

addition, electoral volatility has been higher on average than in most other new democracies of

post-communist Europe. We use data from the Polish Panel survey, POLPAN, 1988-2008, in

which respondents were asked to assess the extent to which ten parties were characterized by

corrupt practices and to report their votes in previous parliamentary elections (most recently in

2005 and 2007) as well as their vote intention in the forthcoming election.

We begin with the assumption that voters view distinct political parties for which they

may vote along a variety of dimensions, the parties perceived level of corruption being one of

them. With this assumption in mind, we pose the following research question: Is the perception of

the degree of corruption of individual political parties a significant factor in vote choice? More

specifically, we ask whether, controlling for other factors, intention to vote for aspecificparty in

the forthcoming election is influenced by individuals perceptions of corruption of that party.

Additionally, is preference for thatspecificparty affected by perception of the degree of

corruption of competing parties? Finally, to determine whether perceptions of political party

corruption matter at allfor voters preferences, we propose a party choice model in which the

voter is confronted with ten parties, one of their attributes being the voters assessment of that

partys level of corruption. The choice model allows us to estimate the influence of perceptions of

corruption of each party, net of other variables, on vote intention.

Our focus on vote intentions in the forthcoming election constitutes a hard test of our

research hypothesis that perceptions of political party corruption affect vote choice. Although

party corruption was highly salient during the 2005 parliamentary election, it was far less so in

the most recent election of 2007. Moreover, three of the four parties that won legislative seats in

the 2007 election were much less tainted by corrupt behavior than has usually been in the case in

recent Polish politics.

8/13/2019 SSRN-id1450260

4/31

3

The paper proceeds in the following manner. In the next section we describe the data

upon which our analyses are based. A discussion of the Polish context follows. We then turn to

the analyses in which we first consider the effect of overall perceptions of political party

corruption on electoral participation and, second, the impact of perceptions of corruption of

particularparties on vote choice. We conclude with a brief discussion of the contributions our

paper makes to the literature on political corruption, the limitations of our analyses, and

suggestions for further research.

Data

Data for this paper come from the Polish Panel Survey, POLPAN, a survey conducted

among the adult population in Poland in 1988, 1993, 1998, 2003, and 2008 (Slomczynski and

Marquart-Pyatt 2007). All waves were supplemented by new cohorts so that a survey for a given

year was representative of the adult population aged 21 and older. Here, we use the 2008 wave

which consists of 1216 panelists (aged 26-85) and 241 new respondents (aged 21-25). With

internal weights, the sample mirrors the population of persons aged 21-85, according to basic

demographic characteristics. Items related to perceived party corruption were assigned to a

randomly selected number of respondents, constituting about half of the total sample (N = 630).

During a pilot study of the 2008 wave it became clear that party corruption was a

sensitive topic. Respondents tried to avoid answering questions about party corruption if they felt

that their answers might be interpreted as a critique of particular sets of politicians. In recent

years in Poland several law suits have been brought against individuals who accused politicians

of corruption. For this reason, it is likely that any wording about party corruption that could be

interpreted in a personalized manner makes Polish respondents uncomfortable. This was

apparent when asking the simple question: Do you think that party A is more corrupt than other

parties, the same, less, or not at all? In our pilot study, a sizable proportion of respondents

immediately pointed out that parties are not corrupt but rather people in these parties (17.0%).

8/13/2019 SSRN-id1450260

5/31

4

The number of refusals for this formulation of the question exceeded 15%. Given this situation, in

the final study, we asked about perceived party corruption in a less direct way:

To what extent is a given party subject to corrupt practices? To a great extent? On the

average? To a low extent? Not at all?

(W jakim stopniu dana partia podlega procesom korupcji? W stopniu wysokim? Srednim?W stopniu niskim? Wcale?)

This is an impersonal formulation of the question in the sense that corruption practices

may refer to an institutional basis of corrupt behavior and not necessarily to people who engage in

such behavior. Probably for this reason, the refusal rate for this question dropped significantly to

below 5%.

Respondents were asked to evaluate the following parties:Civic Platform, PO (Platforma Obywatelska)

Law and Justice, PiS (Prawo i Sprawiedliwosc)

Democratic Left Alliance, SLD (Sojusz Lewicy Demokratycznej)

Polish Peoples Party, PSL (Polskie Stronnictwo Ludowe)

League of Polish Families, LPR (Liga Polskich Rodzin)

National Party of Retirees an Pentioners, KPEiR (Krajowa Partia Emerytow i

Rencistow)

Polish Labour Party, PPP (Polska Partia Pracy)

Womens Party, PK (Partia Kobiet)

Social Democracy of Poland, SDPL (Socjaldemokracja Polska)

Self-Defense, SO (Samoobrona)

In this paper we focus primarily on the perception of corruption of four parties: Civic

Platform, PO (Platforma Obywatelska), Law and Justice, PiS (Prawo i Sprawiedliwosc),

Democratic Left Alliance, SLD (Sojusz Lewicy Demokratycznej), and Polish Peoples Party, PSL

(Polskie Stronnictwo Ludowe). We do so for both substantive and methodological reasons.

Substantively, these four parties are the most important parties in terms of concentration of votes

and overall attention paid to them by the media and the public. Methodologically, for the

remaining parties the degree of differentiation of opinions is much less pronounced, with the

notable exception of Self-Defense, SO (Samoobrona).

However, we use information about all parties when we refer to some properties of the

distribution of perceptions of political party corruption. For example, for each of the four parties,

8/13/2019 SSRN-id1450260

6/31

5

we construct an index of a competing party in the dimension of corruption, assuming that this

would be the party with the highest score among the entire set of parties. We return to this issue

later.

The Polish Context: Elections, Political Parties and Corruption

Polands constitution can best be described as a premier-presidential system with a

bicameral legislature. Since 1991, the country has a directly elected president who serves for a

five year term, a directly elected upper house (Senat) with 100 members, and a directly elected

more powerful lower house (Sejm) with 460 members. Members of parliament serve for a four

year term, unless the Sejm is dissolved and early elections are called. The electoral system has

changed only slightly since 1993. The lower house of the parliament is chosen using the open-list

proportional representation format. Single parties must meet a five percent threshold and electoral

coalitions eight percent to gain seats. The 460 deputies of the Sejm are chosen from 41 electoral

districts whose magnitude ranges from seven to 19.

Since the holding of semi-competitive elections in 1989, effectively signaling the demise

of Polands communist regime, six fully free parliamentary elections have taken place (1991,

1993, 1997, 2001, 2005, and 2007). Until the most recent election of 2007, the party system in

Poland exhibited a considerable degree of fluidity typical of most new democracies of post-

communist Europe. New parties or electoral coalitions succeeded in gaining legislative seats or in

forming the Government (for example, Solidarity Electoral Action in 1997 and PiS in 2001).

Existing parties, including governing parties, disappeared, split, or became minor parties within

the Sejm (for instance, Solidarity Electoral Action in 2001 and Democratic Left Alliance in 2005)

or failed to gain seats whatsoever (Freedom Union in 2001, and Samoobrana and League of

Polish Families in 2007). Indeed, no governing party has managed to win reelection in

consecutive elections. Consonant with weak ties between parties and politicians, voters too have

shown weak loyalties to existing parties. Moreover, voter turnout in parliamentary elections has

8/13/2019 SSRN-id1450260

7/31

6

declined since the 1991 elections reaching a low of 40.5 percent in 2005. There was an uptick in

2007, but even then turnout only reached 53.9 percent. Not surprisingly, electoral volatility has

been considerable throughout this period, higher on average than in any other of the new

democracies of post-communist Europe.

Party competition throughout this period was multidimensional and centered on the

historic divide between the Communist and Solidarity eras, a cultural divide between

secularism and individualism versus more solidaristic and traditional values, the role of the

state in the economy, and, from the late 1990s, Polands relationship to the EU (Gwiazda 2008;

Millard 2003, 2007; Szczerbiak 2007, 2008). Mutual allegations of political corruption and

promises to wage assertive anti-corruption policies if elected also played a prominent role in

campaign politics and, in doing so, contributed to the fluidity of the party system. For example,

party corruption played a significant part in the splintering of the governing Democratic Left

Alliance and its devastating defeat in the 2005 election (Jasiewicz 2008). Law and Justices self

portrayal as a party untainted by malfeasance and promise to cleanse the state certainly

contributed to its victory in that same election (Szczerbiak 2007). A survey conducted shortly

before the 2005 election found that 38 percent of respondents said that the most important factor

determining their vote would be whether or not they felt that a party was untainted by corruption.

About one third to slightly more than half of supporters of various opposition parties stated that

this was the most important reason for their party choice (Szczerbiak 2007: 212). Apart from

campaign politics, well-publicized ongoing parliamentary commission investigations of

corruption and constant attention by the media to political scandals further fueled the sense on the

part of the mass public that political corruption was pervasive.

Quite early in the post-transition era, according to a mass survey conducted in 1995, 51

percent of respondents stated that high-level officials benefited personally from the public trough.

Sixty-one percent believed that such officials would use public funds to benefit their party

(Holmes 2006: 162). A survey conducted several years later in 2003 found that 77 percent of

8/13/2019 SSRN-id1450260

8/31

7

respondents claimed that politicians were dishonest (Tworzecki 2008). Global Corruption

Barometer surveys conducted by Transparency International in 2004 and 2007, which asked

respondents to assess the degree to which various sectors of the country were affected by

corruption on a scale ranging from (1) not at all corrupt to (5 ) extremely corrupt, showed that in

Poland political parties were thought to be the most tainted by corruption. In both 2004 and 2007,

the score for political parties was 4.2 (Global Corruption Barometer 2004, 2007). It is no surprise,

then, that significant numbers of Poles regard political parties as untrustworthy. According to

2006 data from the European Social Survey, Polands mean score on the measure of trust in

political parties was the lowest among 23 countries included in the study. The distance between

the average for Poland and the average for other countries was well over one standard deviation

(Slomczynski and Janicka 2009).

Corruption was less pivotal to the campaign leading up to the 2007 parliamentary

elections than it had been in 2005.This was so despite the ouster from power of Samoobrana, a

junior partner to the Law and Justice-led government, because of malfeasance, as well as charges

of bribe taking made during the campaign itself against a deputy of Civic Platform, Law and

Justices leading competitor (Gwiazda 2007; Szczerbiak 2008). Moreover, between the 2005 and

2007 election, stabilization of the Polish party system appeared to have taken place. The number

of parties was reduced from six to four and there were no new entrants to the parliamentary arena.

The four parties differed significantly in their ideological stances (Szczerbiak 2007,

2008). Law and Justice (PiS), the governing party between 2005 and 2007, is a conservative party

which espouses traditional values, more solidaristic policies, and close Church-state relations.

Moreover, from its inception as a party in 2001 PiS has leveled a harsh critique of the political

order following the transition as being rife with holdovers from the communist era and political

corruption. Civic Platform (PO), which won the 2007 election, is more ideologically

heterogeneous and has been categorized as a liberal-conservative or center-right party:

economically liberal and socially conservative. The post-communist turned social democratic

8/13/2019 SSRN-id1450260

9/31

8

party, Left Democratic Alliance (SLD) is a shadow of its former self when it governed between

1993-1997 and 2001-2005. It advocates more solidaristic economic policies coupled with

secular and more liberal stances with regard to social mores. Finally, the Polish Peasant Party

(PSL), the junior partner in SLD-led Governments, is primarily an agrarian party, but its message

is more broadly populist in its criticism of the existing democracy as unresponsive to the concerns

of ordinary people.

As Table 1 shows, except for the substantial increase in the reported vote for PO, there is

little change in the distribution of reported vote for the other three parties between 2005 and 2007.

This stability in reported vote mirrors the stability in actual electoral outcomes in terms of party

vote shares (Gwiazda 2008). With the exception of SLD and Samoobrana (SO), none of the

remaining parties represented in the Sejm and seeking reelection in 2007 was tainted with a

history of corruption. Does the fact that malfeasance was less salient in the most recent election

mean that perceptions of political party corruption have little impact on vote choice in the

forthcoming election? Our focus on vote intentions in the aftermath of an election in which

corruption was less pivotal an issue makes for a hard test of our central proposition that

assessments of political party malfeasance matter for voter preferences.

Table 1 about here

Overall Perceptions of Political Party Corruption and Electoral Participation

Despite the fact that Poles hold political parties in seemingly universal disrepute as being

among the most corrupt institutions in the new democracy, do Poles tend to view all parties as

being more or less equally corrupt? As was mentioned earlier, in POLPAN 2008, for each of ten

parties respondents were asked to assess the extent to which the party was subject to corrupt

practices. Possibleresponses rangedfrom (1) not at all (5) to a great extent. Table 2 presents data

on the mean value and standard deviation of perceptions of corruption of each of the four main

parties elected to the Sejm in 2007, as well as of all other parties combined. The middle category

8/13/2019 SSRN-id1450260

10/31

9

contains those who responded dont know as well as in other ways that revealed their

ambivalence or uncertainty as to how to answer the question. Experimenting with optimal scoring

of such responses indicated that (a) all such answers can be grouped together, and (b) they can be

placed in the middle of well defined answers.

As can be seen from the data presented in Table 2, there is considerable variation in the

extent to whichspecificpolitical parties are viewed as subject to corruption, and these differences

are statistically significant. Self Defense (not currently in the parliament but a junior partner for a

time in the PiS- led Government between 2005-2007) is perceived as being the most corrupt,

followed closely by the scandal-ridden Democratic Left Alliance (SLD) when it was the

governing party between 2001 and 2005. Not surprisingly, such perceptions are in accord with

extensive media coverage of incidents of corruption on the part of both parties. In contrast, Civic

Platform (now the governing party) and to a lesser extent Law and Justice (the governing party in

2005-2007) are viewed as relatively clean, in line with their self portrayals and depictions in the

media as anti-corruption forces. The Polish Peoples Party (a junior partner to SLD-led

governments) is perceived as the least tainted by malfeasance. At the same time, there is a

similarly sizable degree of variation in respondents perceptions of the level of corruption of each

of the given parties (as indicated by the standard deviation).

Table 2 about here

Generally, if respondents give relatively high scores to one party, they have a strong

tendency to give high scores to another party. As can be seen in Table 3, the correlation between

rankings for pairs of parties ranges from 0.210 to 0.494. The most similar rankings are for Civic

Platform and Polish Peoples Party, while the most different are for Law and Justice and

Democratic Left Alliance. This pattern is reflected, to great extent, in the comparisons of the

means. For example, the difference for Polish Peoples Party and Civic Platform (3.074 - 3.123 =

-0.051) is statistically insignificant. The difference for Democratic Left Alliance and Law and

Justice is much larger (3.572 3.280 = 0.292) and statistically significant. Note, however, that

8/13/2019 SSRN-id1450260

11/31

10

the greatest difference is between Democratic Left Alliance and Civic Platform: (3.572 3.123 =

0.449). This difference is equal to almost a half of the standard deviation of scores for these

parties, and therefore is substantial in terms of the distribution of scores.

Table 3 about here

Having provided descriptive statistics for each of the given parties on the perceived

corruption scale, we turn to the substantively important question of whether perceptions of overall

political party corruption have an effect on individuals willingness to participate in elections.

The prevalent view claims that it does: the more an individual views parties as being corrupt, the

more one is likely to abstain from voting (see, for example, Davis, Camp and Coleman 2004).

This is one way, among others, that political party corruption is thought to undermine the quality

of democracy. Is this the case in Poland? Previous work based on 2003 POLPAN data on

perceptions of political party corruption and electoral participation suggests a somewhat

complicated picture (Shabad, Slomczynski and Kistner 2008). The results for the 2008 survey

with regard to intention to vote in the forthcoming election are similar.

In Table 4 we present the results of a logisitic regression analysis in which intention to

vote (No = 0; Yes = 1) is regressed on overall perceptions of political party corruption (mean

value and standard deviation, respectively). Holding constant whether the respondent reported

having voted in the 2005 parliamentary election, the mean value of perceived corruption of all ten

political parties is not statistically significant.1Those who, on average, view parties as being quite

subject to corruption are no less likely to say that they will vote than those who, on average,

perceive parties as being relatively uncorrupt. But the result is quite different when we consider

_____________________________________________________________________________

1Since voting in 2007 was close in time to the survey, the correlations between reported vote in

2007 and intention to vote for a given party in the forthcoming election were too high to be

included in proposed models.

8/13/2019 SSRN-id1450260

12/31

11

the degree to which individuals vary in their perceptions of the level of corruption within all 10

parties. Controlling for both previous electoral participation and the mean value on the perceived

corruption scale, the more an individual sees divergence among parties in their level of corruption

the more likely he or she is to express an intention to vote in the next election. The result is

statistically significant. Figure 1 nicely illustrates this strong linear relationship between

differentiation of perceived corruption of political parties and the probability of voting in the

forthcoming election. For example, the probability of voting for those whose perceptions of

corruption across the parties are more or less uniform (whether high or low) is below 60%, as

compared with a probability of 85% for those who hold the most divergent views.

Table 4 and Figure 1 about here

Perceptions of Political Party Corruption and Vote Intention

We have seen that the degree to which individuals differentiatein their assessments of the

extent of corruption of all ten parties affects the probability of their voting in the next election.

We turn next to the issue of whether perception of the degree to which agivenparty is subject to

corruption affects the likelihood of voting for thatparty. Put simply, do assessments of political

party corruption matter for vote choice? In Table 5 we present the results of a logistic regression

analysis for each of the four main parties in which we regress intention to vote for a given party

(0= NO; 1 = Yes) on the respondents perception of the extent of corruption of that party. In each

analysis, in addition to the individuals score on the perceived corruption scale of a given party,

we include his or her score on the perceived corruption scale of the political party (among the

remaining nine) that is viewed by the respondent as being the most corruptcompared to the

given one. We call this variable perceived corruption of a competing party.

For all parties, with the exception of Polish Peoples Party, perceived corruption of a

given party has a statistically significant and sizable impact on the probability of voting for that

party. The higher the perceived corruption of Civic Platform or Law and Justice or Democratic

8/13/2019 SSRN-id1450260

13/31

12

Left Alliance the less likely one is to state a preference for that party in the forthcoming election.

The impact of assessments of corruption is particularly strong in the case of SLD. The results

presented in Table 5 also show that the likelihood of voting for either PiS or SLD increases

substantially when the respondent assesses a competing party as being comparatively for more

corrupt than either PiS or SLD. Thus, the likelihood of choosing PiS or SLD is affected both by

voters perceptions of the extent of corruption within that party and by their perceptions of the

extent of corruption of a competing party. In short, perceptions of political party corruption

matter for voting preferences.

Table 5 about here

Do perceptions of political party corruption of eachof the four main parties influence the

probability of voting for oneof the four parties in the forthcoming election? To answer this

question we apply a multinomial regression analysis in which our dependent variable is an

intention to vote for PO, PiS, SLD and PSL, with intention to vote for other parties as a reference

category. We include perceptions of corruption of these same four parties. Consistent with the

results presented in Table 5,holding constant perceptions of other parties, the more the

respondent views a given party as tainted by corruption, the less likely he or she is to vote for that

party in the next election. This is true for the PSL as well. As the data in Table 6 show, the

magnitude of the relationships is quite strong and similar across the four parties, but is

particularly pronounced in the case of PO and least so in the case of PSL. Corruption perceptions

of competitor parties also affect the likelihood of selecting PO or PiS. For example, the more PiS

is viewed as corrupt, the greater the probability of voting for its main rival, PO. In similar fashion,

the likelihood of voting for PiS increases the more PO and SLD are seen as corrupt. Interestingly,

the relationship is reversed for PiS in relationship to PSL: the less corrupt PSL is perceived to be

the more likely one is to prefer PiS. Thus, it seems, that, with respect to corruption at least, voters

do not see PiS and PSL as rivals parties. In contrast, corruption perceptions of other parties

have no statistically significant effect, one way or the other, on the probability of choosing SLD

8/13/2019 SSRN-id1450260

14/31

13

or PSL. It should be noted, however, that like in the cases of PiS and PSL, the negative

relationships between perceived corruption of PSL and the probability of choosing SLD and

between perceived corruption of SLD and the probability of selecting PSL, these two parties are

not competing against each other in the domain of corruption.

Table 6 about here

The results presented thus far strongly suggest that perceptions of corruption of a given

party, taking into account perceptions of competing parties, influence vote intentions. But does

this effect persist when other variables that may influence vote choice are also considered? It

should be emphasized that our purpose here is not to account as fully as possible for vote choice.

Rather, it is to determine whether corruption perceptions are related to choice of a given party

when we control for other possible factors. In Tables 7-10, we present the results of logistic

regression analyses of intention to vote for each of the four main political parties. In addition to

respondents corruption perceptions of that party and those of the party seen as the most corrupt

compared to the given one, as well as the mean value on the corruption scale as a whole, we

include three sets of control variables: (1) previous voting behavior in 2005; (2) standard

demographic characteristics (gender, age, size of residence, and educational level); and (3) an

item tapping ideological preference or attitude that best discriminates among supporters of the

four parties. In the case of Civic Platform, the questionnaire item that best differentiates between

those who express an intention to vote for PO as opposed to other parties is strong agreement

with the notion that In politics, mutual concessions are the best way to make difficult decisions.

For PiS, it is strong disagreementwith the view that the Catholic Church has too much power. In

the case of SLD, the item is strong agreementwith the statement that the Government is

responsible for decreasing income inequalities. Finally, for the populist PSL, it is strong

agreementwith the view that People have very little influence on how the country is governed.

Not surprisingly, for each of the four parties, reported vote for that party in the 2005

election has by far the greatest impact. In the cases of PO and PiS, ideological preferences also

8/13/2019 SSRN-id1450260

15/31

14

have sizable effects. Demographic characteristics have an impact as well, depending upon the

party in question. For example, size of residence affects the probability of voting for both PiS and

PSL, with those in rural areas more likely to vote for those parties. Women are far more likely

than men to choose SLD. Nonetheless, when these substantively and statistically significant

determinants of vote choice are taken into account, perceptions of corruption still play asizable

role in the probability of choosing a particular party in the forthcoming election. For PO, PiS, and

SLD, assessments of that partys level of corruption carry considerable weight, all else being

equal. Perception of corruption of a competing party or the mean value on the corruption scale

has no statistically significant effect. For PSL, however, all three corruption variables matter. The

more PSL is seen as corrupt the less likely the respondent is to choose that party; the more, on

average, the respondent views all ten parties as being corrupt, the more apt he or she is to express

a preference for PSL. This party seems to be preferred by those who are particularly dissatisfied

with party politics.

Tables 7-10 about here

Choosing a Party: Do Corruption Perceptions Matter?

Our final analysis addresses the question of whether vote choice, regardless of which

particular party is chosen,is affected by perceptions of political party corruption. Answering this

question requires setting up a party choice model that takes into account simultaneously

characteristics of both individuals and parties: each individual, with his or her own attributes, is

confronted with a set of parties from which to choose, each of which is characterized by its own

known or presumed attributes. Table 11 presents the essence of the setup of data for the party

choice model. Each respondent, A, B, C, and D, is characterized by age and gender. Note that

these characteristics are constant for a respondent vis a vis parties with which he or she is

confronted. We include also two characteristics of parties: whether they are currently in the Sejm,

and whether they are leftist/populist or not. (See Table 11 for coding of parties with regard to

these two characteristics.) These attributes of parties do not change across individuals. Finally,

8/13/2019 SSRN-id1450260

16/31

15

there are variables that change across both individuals and parties, i.e., assessment of corruption

of a given party, and intention to vote for a given party.

Table 11 about here

We solved the problem of party choice determinants by applying an Alternative Specific

Conditional Logit model (asclogit) in STATA. Gender and age have no statistically significant

effects on party choice. The significant effects of party characteristics are as one would expect.

Parties that are currently in the parliament have a higher chance of being chosen. Leftist/populist

parties have a lower probability of being chosen, due to the fact that with the exception of PSL

such parties are either quite small groups or have been greatly discredited because of their

activities in office (SLD, SO) What is important for our purpose, however, is the fact that

respondents assessments of party corruption have a considerable and statistically significant

negative effect on party choice, even when other variables are controlled. The higher the

individual ranks a party on the scale of corruption the lower the probability that he or she will

choose that party in the forthcoming election. In short, corruption perceptions matter.

Table 12 about here

DISCUSSION

Ina nutshell, all of our analyses indicate that perceptions of political party corruption

have an effect on voting behavior. This is so with regard to electoral participation per se, intention

to vote for aparticularparty as opposed to a competing party, and vote choice regardlessof

which particular party is chosen. The impact of assessments of party malfeasance holds up even

when other determinants of vote choice are taken into account. Our findings also strongly suggest

that the extent to which individuals differentiateamong parties with regard to corruption, rather

than the average value of such perceptions (mean value), is important in motivating electoral

participation. This result stands in contrast to other research that finds that the more an individual

perceives corruption to be widespread the more likely he or she is to abstain from voting (but see

8/13/2019 SSRN-id1450260

17/31

16

Kistner 2007; Shabad, Slomczynki and Kistner 2008). As we have shown, apart from their effect

on the decision whether to vote or not, differentiated perceptions of political party corruption also

affect vote choiceper se, whether with regard to the probability of voting for a given party or the

probability of voting for anyone of the several party options available. Insofar as we focus here

on the relationship between mass-level perceptions of corruption of particular parties and voting,

our findings contribute to an understanding of electoral behavior in Poland. Given that party

malfeasance has been a salient issue in electoral politics and has played a role in the waning and

rising fortunes of existing and new parties and vote switching, by implication our findings also

speak to questions related to the institutionalization of Polands party system.

Our research, however, has two types of limitation which we hope to address in future

work. Substantively, the main limitation of this study pertains to the fact that it does not

incorporate assessments of other characteristics of specific parties apart from their perceived level

of corruption. In particular, it would be advantageous to include individuals perceptions of the

parties effectiveness in promoting new legislation in a variety of policy realms. This would allow

us to assess, for example, whether voters are more likely to prefer corrupt but effective parties

over those that are clean but ineffective. A second limitation concerns our use of responses to

questionnaire items relating to vote choice and perceptions of political party corruption taken

solely from the 2008 wave of POLPAN. The validity of causal claims would be considerably

improved were we to use reported vote in 2005 (taken from the 2008 wave) as the dependent

variable and perceptions of political party corruption from the 2003 wave of the panel study.

Methodologically, assessment of corruption was performed on a scale that did not

involve pair-wise comparisons between parties. Such comparisons would allow us to discriminate

better between sets of particular parties along the dimension of perceived corruption and, thus, to

gauge the impact of more nuanced variations on vote choice. Moreover, in the survey there were

no questions that pertained to respondents actual knowledge of incidents of political party

corruption. A more comprehensive investigation of the correlates of corruption perceptions

8/13/2019 SSRN-id1450260

18/31

17

should take into account both the extent to which individuals are informed about malfeasance and

the sources from which they derive such information.

Taken together, these substantive and methodological limitations point to several fruitful

areas for further investigation on mass-level perceptions of party corruption and voting behavior.

Additionally, an extension of this type of research to other new democracies as well as to long-

lived democracies would make a significant contribution to our understanding of the ways by

which political corruption affects both voters and party systems alike.

8/13/2019 SSRN-id1450260

19/31

18

REFERENCES

Anderson, Christopher J.and Yuliya V. Tverdova. (2003). Corruption, Political Allegiances, and

Attitudes toward Government in Contemporary Democracies. American Journal of

Political Science 47 (1): 91-109.

Bowler, Shaun and Jeffrey A. Karp. (2004). Politicians, Scandals, and Trust in Government.

Political Behavior 26 (3): 271-87.

Canache, Damarys and Michael E. Allison. (2005). Perceptions of Political Corruption in Latin

American Democracies. Latin American Politics and Societies 47 (3): 91-111.

Davis, Charles L., Roderic Ai Camp, and Kenneth M. Coleman. (2004). The Influence of Party

Systems on Citizens Perceptions of Corruption and Electoral Response in Latin

America. Comparative Political Studies 37 (6): 677-703.

Della Porta, Donatella (2000). Social Capital, Beliefs in Government, and Political Corruption.

In Disaffected Democracies: Whats Troubling the Trilateral Countries? eds. Susan J.Pharr and Robert D. Putnam. Princeton: Princeton University Press, 202-38.

Gwiazda, Anna. (2008). The parliamentary election in Poland, October 2007. Electoral Studies

27: 760-4.

Holmes, Leslie. 2006. Rotten States? Corruption, Post-Communism and Neolibrealism. Durham,

N.C.: Duke University Press.

Jasiewicz, Krzysztof. (2008). The (not always sweet) uses of opportunism: Post-communist

political parties in Poland. Communist and Post-Communist Studies 41: 421-42.

Kistner, Natalie. (2007). The Sources and Consequences of Public Perceptions of Corruption,In Continuity and Change in Social Life: Structural and Psychological Adjustment in

Poland, eds. Kazimierz M. Slomczynski and Sandra T. Marquart-Pyatt. Warsaw: IFIS,

355-73.

Millard, Frances. (2003). The parliamentary elections in Poland, September 2001. Electoral

Studies 22: 367-74.

Millard, Frances. (2007). The 2005 parliamentary and presidential elections in Poland.

Electoral Studies 26: 210-15.

Sandholtz, Wayne and Rein Taagepera. (2005). Corruption, Culture, and Communism.

International Review of Sociology 15 (1): 109-31.

Seligson, Mitchell A. (2002). The Impact of Corruption on Regime Legitimacy: A Comparative

Study of Four Latin American Countries. Journal of Politics 64 (2): 408-33.

Shabad, Goldie, Kazimierz M. Slomczynski, and Natalie Kistner. (2008). Mass Perceptions of

Political Party Corruption in Poland: Patterns, Correlates, and Implications. Poster

presented at the annual meeting of the American Political Science Association, Boston.

8/13/2019 SSRN-id1450260

20/31

19

Shabad, Goldie, Kazimierz M. Slomczynski, and Jakub Zielinski. (2008). Electoral Control in a

New Democracy: Voters and Politicians in Poland. Unpublished mss.

Slomczynski, Kazimierz M. and Krystyna Janicka. 2009. Structural Determinants of Trust in

Public Institutions: Cross-National Differentiation. International Journal of Sociology 39

(1): 8-29.

Slomczynski, Kazimierz M. and Sandra T. Marquart-Pyatt., eds. (2007). Continuity and Change

In Social Life: Structural and Psychological Adjustment in Poland. Warsaw: IFIS.

Szczerbiak, Aleks. 2007. Social Poland Defeats Liberal Poland? The September-October

2005 Polish Parliamentary and Presidential Elections. Journal of Communist Studies

and Transition Politics 23(2): 203-32.

Szczerbiak, Aleks. 2008. The Birth of a Bipolar Party System or a Referendum on a Polarizing

Government? The October 2007 Polish Parliamentary Election. Journal of Communist

Studies and Transition Politics 24 (3): 415-43.

Transparency International Global Corruption Barometer Reports. ( 2004, 2007).

Tworzecki, Hubert. (2008). A disaffected new democracy? Identities, institutions and civic

engagement in post-communist Poland. Communist and Post-Communist Studies 41:

47-62.

8/13/2019 SSRN-id1450260

21/31

20

Table 1. Voting Declarations for the 2005 and 2007 Elections, and Voting Intentions for the

Forthcoming Election, POLPAN Data, 2008

2005Election

2007Election

Future

Election

VotersPercentage of the respondents who claimed to

have voted (or intend to vote) a 71.0 68.8 65.5

Declared Votes for Parties (%)Civic Platform, PO (Platforma Obywatelska)

29.7 48.7 45.0Law and Justice, PiS (Prawo i Sprawiedliwo)

32.5 26.7 19.2Democratic Left Alliance, SLD (Sojusz Lewicy

Demokratycznej) 11.1 8.0 5.7

Polish Peoples Party, PSL (PolskieStronnictwo Ludowe) 5.9 7.0 5.0Other Parties, OTH (Among them Self-

Defence, SO, Samoobrona) 20.8 (2.8) 9.6 (1.5) 2.9 (0.5)No Party Declared, NO

- - 22.2Those who claimed to have voted (or intend to

vote) N = 100% 1034 1002 955

aN = 1,457

8/13/2019 SSRN-id1450260

22/31

21

Table 2. Perceived Corruption of Political Parties in Poland, POLPAN Data, 2008

Political parties

Mean

ValueStandard

DeviationPercent

dont know

Civic Platform, PO (Platforma Obywatelska)3.123 1.098 26.9

Law and Justice, PiS (Prawo i

Sprawiedliwo) 3.280 1.117 26.5Democratic Left Alliance, SLD (Sojusz

Lewicy Demokratycznej) 3.572 1.071 25.0Polish Peoples Party, PSL (Polskie

Stronnictwo Ludowe) 3.074 1.067 33.5Other Parties, OTH (Among them Self-

Defence, Samoobrona) 3.082 (3.740) 0.716 (1.131) 29.2 (28.0)

All Parties, ALL 3.154 0.672 22.2

Note: Perceived corruption measured on the five-point scale, from 1 denoting that a party is not at allsubject to corruption to 5 denoting that a party is subject to corruption to a high degree. Dont knows are

put in the middle of the scale.

Table 3. Correlations (Above Diagonal) and Mean Differences (Below Diagonal) for Perceived

Corruption of Political Parties in Poland, POLPAN Data, 2008

Political parties

PO PiS SLD PSL OTH

Civic Platform, PO (Platforma

Obywatelska) 0.395* 0.351* 0.494* 0.485*

Law and Justice, PiS (Prawo iSprawiedliwo) 0.157* 0.210* 0.365* 0.373*Democratic Left Alliance, SLD (Sojusz

Lewicy Demokratycznej) 0.449* 0.292* 0.394* 0.409*Polish Peoples Party, PSL (Polskie

Stronnictwo Ludowe) -0.051 -0.208* -0.499* 0.526*Other Parties, OTH (Among them Self-

Defence, Samoobrona) -0.041 -0.198* -0.490* 0.009

Note: Mean differences: mean values for parties listed in rows minus mean values for parties listed incolumns. The test of the significance for the repeated measurement.

* p < 0.01

8/13/2019 SSRN-id1450260

23/31

22

Table 4. Intention to Vote in the Forthcoming Elections Regressed on Overall Perception of

Political Party Corruption, POLPAN Data, 2008

Dependent variable

V = log (p / 1 p)Independent Variables

B SE Exp(B)Mean Value of the Perceived Corruption

Scale for All Parties 0.014 0.158 1.014Standard Deviation of the Perceived

Corruption Scale Across Parties 0.194* 0.073 1.214Voting in the 2007 ElectionsYes = 1, No = 0 2.108* 0.219 8.233

Constant -1.040 0.573 0.354

Chi-square = 115.2, -2 Log-likelihood = 553.7, Nagelkerke R-square = 0.272

Figure 1. Differentiation of Perceived Corruption of Political Parties and Predicted Probability of

Voting in the Forthcoming Elections

8/13/2019 SSRN-id1450260

24/31

23

Table 5. Intention of Vote for a Given Party Regressed on Perception of Political Party

Corruption, POLPAN Data, 2008

Dependent variable

V = log (p / 1 p)

Independent Variables B SE Exp(B)

Civic Platform, PO (Platforma Obywatelska) a

Perceived corruption of PO -0.338* 0.100 0.713Perceived corruption of a competing party b

0.138 0.127 1.148

Constant 0.314 0.569 1.369

Law and Justice, PiS (Prawo i Sprawiedliwo) c

Perceived corruption of PiS -0.456* 0.120 0.634Perceived corruption of a competing party b

0.411* 0.168 1.508

Constant -1.834** 0.807 0.160

Democratic Left Alliance, SLD (Sojusz Lewicy Demokratycznej) d

Perceived corruption of SLD -0.537* 0.183 0.584Perceived corruption of a competing party b

0.431** 0.248 1.539

Constant -2.470** 1.071 0.085

Polish Peoples Party, PSL (Polskie Stronnictwo Ludowe) e

Perceived corruption of PSL -0.328 0.256 0.720Perceived corruption of a competing party b

0.195 0.263 1.215

Constant -2.920** 1.063 0.054

aChi-square = 12.0, -2 Log-likelihood = 492.5, Nagelkerke R-square = 0.043bSubjectively most corrupt party other than a chosen one.cChi-square = 19.6, -2 Log-likelihood = 331.4, Nagelkerke R-square = 0.085

dChi-square = 8.8, -2 Log-likelihood = 188.8, Nagelkerke R-square = 0.057eChi-square = 1.75, -2 Log-likelihood = 123.4, Nagelkerke R-square = 0.016

* p < 0.01 ** p < 0.05

8/13/2019 SSRN-id1450260

25/31

24

Table 6. Multinomial Regression of Intention to Vote for Civic Platform (PO), Law and Justice

(PiS), Democratic Left Alliance(SLD), and Polish Peoples Party(PSL) on Perceived Corruption

of These Parties, POLPAN Data, 2008

Voting / Perceived corruption B SE Sig. Exp(B)

Perceived corruption of PO -.362 .156 .020 .696

Perceived corruption of PiS .262 .126 .037 1.299

Perceived corruption of SLD .152 .127 .231 1.164

PO

Perceived corruption of PSL .114 .157 .468 1.120

Perceived corruption of PO .376 .185 .041 1.457

Perceived corruption of PiS -.434 .149 .004 .648

Perceived corru tion of SLD .272 .145 .061 1.313

PiS

Perceived corruption of PSL -.366 .184 .047 .693


Perceived corru tion of PiS .163 .195 .402 1.177

Perceived corruption of SLD -.424 .208 .041 .654

SLD



Perceived corru tion of PiS .003 .233 .991 1.003

Perceived corru tion of SLD -.045 .229 .844 .956

PSL


8/13/2019 SSRN-id1450260

26/31

25

Table 7. Intention to Vote for Civic Platform(PO) Regressed on Perception of Political Party

Corruption and Control Variables, POLPAN Data, 2008

Dependent variable

V = log (p / 1 p)


Perceived corruption of PO -0.402** 0.197 0.669

Perceived corruption of the rival party a 0.159 0.192 1.173

Mean value of perceived corruption b 0.081 0.336 1.085

Gender (Male = 1, female = 0) -0.020 0.322 0.981

Age (Years) -0.004 0.010 0.992

Urban (Yes = 1, no = 0) 0.354 0.370 1.425

Education (Years of schooling) -0.008 0.050 0.992Support of principle of negotiation and

compromise 0.643** 0.334 1.902

Voting for PO in 2005 election 3.302* 0.359 27.167

Constant -0.986* 0.375 0.373

Chi-square = 156.5, -2 Log-likelihood = 270.1, Nagelkerke R squared = 0.531

a Subjectively most corrupt party other than a chosen one.

b For all parties in the study.

* p < 0.01 ** p < 0.05

8/13/2019 SSRN-id1450260

27/31

26

Table 8. Intention to Vote for Law and Justice (PiS) Regressed on Perception of Political Party

Corruption and Control Variables, POLPAN Data, 2008

Dependent variable

V = log (p / 1 p)


Perceived corruption of PIS -0.579* 0.229 0.560




Age (Years) 0.004 0.014 1.004

Urban (Yes = 1, no = 0) -1.105** 0.168 0.331

Education (Years of schooling) 0.017 0.066 1.017

Support of close Church state relations 0.924*** 0.807 2.520

Voting for PiS in 2005 election 3.768* 0.477 43.286

Constant -4.344* 1.629 0.013




* p < 0.01 ** p < 0.05 *** p < 0.10

8/13/2019 SSRN-id1450260

28/31

27

Table 9. Intention to Vote for Democratic Left Alliance(SLD) Regressed on Perception of

Political Party Corruption and Control Variables, POLPAN Data, 2008

Dependent variable

V = log (p / 1 p)


Perceived corruption of SLD -0.613** 0.339 0.542



Gender (Male = 1, female = 0) -1.072*** 0.608 0.342

Age (Years) 0.016 0.018 1.016

Urban (Yes = 1, no = 0) -0.879 1.693 0.415

Education (Years of schooling) 0.020 0.091 1.020

Support of egalitarianism 0.721 0.637 2.056Voting for SLDin 2005 election 4.378* 0.657 79.665

Constant -4.376** 1.983 0.013




* p < 0.01 ** p < 0.05 *** p < 0.10

8/13/2019 SSRN-id1450260

29/31

28

Table 10. Intention to Vote for Polish Peoples Party (PSL), Regressed on Perception of Political

Party Corruption and Control Variables, POLPAN Data, 2008

Dependent variable

V = log (p / 1 p)


Perceived corruption of PSL -1.316*** 0.804 0.268

Perceived corruption of the rival party a -1.332*** 0.768 0.264

Mean value of perceived corruption b 2.677*** 1.590 14.539

Gender (Male = 1, female = 0) 0.030 0.909 1.031

Age (Years) 0.009 0.031 1.009

Urban (Yes = 1, no = 0) -1.726** 0.760 0.178

Education (Years of schooling) -0.165 0.173 0.848Agreement with idea that people have little

political influence 0.883 0.979 2.418

Voting for PSL in 2005 election 4.867* 1.064 129.928

Constant -4.123 3.299 0.016




* p < 0.01 ** p < 0.05 *** p < 0.10

8/13/2019 SSRN-id1450260

30/31

29

Table. 11. Data Setup for Party Choice Model in the Forthcoming Election

Constant for a given individual Constant for a givenparty

Vary acrossindividuals and parties

ID Gender Age

Party

name In

Sejm

Left /populist

Corruptionscore

Intendedvote

Respondent A

3 1 42 PO 1 0 2 0

3 1 42 PIS 1 0 2 0

3 1 42 SLD 1 1 4 0

3 1 42 PSL 1 1 3 1

3 1 42 LPR 0 0 1 0

3 1 42 KPEiR 0 1 2 0

3 1 42 PPP 0 1 3 0

3 1 42 PK 0 0 3 0

3 1 42 SDPL 0 1 2 0

3 1 42 SO 0 1 5 0

Respondent B

141 0 26 PO 1 0 1 1

141 0 26 PIS 1 0 2 0

141 0 26 SLD 1 1 5 0

141 0 26 PSL 1 1 2 0141 0 26 LPR 0 0 4 0

141 0 26 KPEiR 0 1 4 0

141 0 26 PPP 0 1 2 0

141 0 26 PK 0 0 3 0

141 0 26 SDPL 0 1 3 0

141 0 26 SO 0 1 4 0

Respondent C

182 0 44 PO 1 0 1 0

182 0 44 PIS 1 0 1 1

182 0 44 SLD 1 1 3 0

182 0 44 PSL 1 1 3 0

182 0 44 LPR 0 0 3 0

182 0 44 KPEiR 0 1 3 0

182 0 44 PPP 0 1 3 0182 0 44 PK 0 0 3 0

182 0 44 SDPL 0 1 3 0

182 0 44 SO 0 1 4 0

Respondent D

141 0 26 PO 1 0 3 0

141 0 26 PIS 1 0 3 0

141 0 26 SLD 1 1 5 0

141 0 26 PSL 1 1 4 0

141 0 26 LPR 0 0 2 0

141 0 26 KPEiR 0 1 2 0

141 0 26 PPP 0 1 2 0

141 0 26 PK 0 0 1 0

141 0 26 SDPL 0 1 3 0

141 0 26 SO 0 1 1 1

8/13/2019 SSRN-id1450260

31/31

Table 12. Determinants of Party Choice in the Forthcoming Election, POLPAN Data, 2008

Dependent variable

V = log (p / 1 p)


Perceived corruption of a party -0.215** 0.105 0.807

Party in Sejm -0.973* 0.274 2.645

Left / populist party -1.353*** 0.743 0.259


Age (Years) - 0.007 0.038 0.993

Number of observations = 5210, Number of cases = 521, Alternatives per case: 10 (constant)Alternative variable: party, Case variable: id, Wald Chi-squared (10) = 8.21, Log likelihood = -274.52

* p < 0.01 ** p < 0.05 *** p < 0.10

SSRN-id1450260

Documents