Content

4 The Bigger Picture of Corruption in Europe in:

Ina Kubbe

Corruption in Europe, page 160 - 184

Is it all about Democracy?

1. Edition 2015, ISBN print: 978-3-8487-2347-8, ISBN online: 978-3-8452-6451-6, https://doi.org/10.5771/9783845264516-160

Series: Comparative Politics - Vergleichende Politikwissenschaft, vol. 6

Bibliographic information
The Bigger Picture of Corruption in Europe To discover the factors that affect the extent of (perceived) corruption in European states and to answer the research questions: “What affects (perceived) corruption in European states over time and across and within countries?” and “is it all about democracy?”, I have designed a heuristic model at the micro and macro level that allows for panel-analyses as well as cross- and within-national comparison, the so called bathtub model of corruption. The conducted panel and multilevel analyses consider both country and individual level. The model takes several theoretical approaches, namely economic and sociological ones such as rational-choice, sociological and historical institutionalism, as well as cultural approaches into account. Following this model, the extent of corruption in Europe can be explained by a number of national and individual characteristics. The indicators for explaining the causes and variations of corruption at the country level are categorized into economic, political, socio-cultural and historical factors. Besides a number of country-specific attributes such as a country’s economic development, degree and duration of democracy or women’s inclusion in parliaments, personal characteristics of individuals also have an impact on the extent of corruption – although to a lesser extent than country characteristics. At the micro level, the indicators that are included in the analysis are categorized into socio-demographic characteristics such as gender and age, values and norms such as interpersonal trust, and attitudes such as an individual’s satisfaction with financial situation and justification of bribery. In the following chapter, these results are summarized, closely examined and critically discussed. Macro Model of Corruption in Europe Macro Model of Corruption The empirical analysis performed at the macro level follow a panel-data research design that enables researchers to run regression analyses that regard both the spatial and temporal dimension of data. The indicators for the explanation of the extent of corruption at the country level are catego- 4 4.1 160 rized into economic, political, socio-cultural and historical factors. Economic factors include a country’s economic development and EU-membership. Political factors include a country’s degree of democracy and percentage of female members in parliaments. The socio-cultural factors capture a country’s dominant religion (Catholicism, Orthodox, Protestantism, Islam). Assuming that corruption has historical roots and is therefore pathdependent, I have also included historical variables such as the durability of democracy and the communist past of a society. To measure corruption at the country level, I used the (transformed) Corruption Perceptions Index from Transparency International that has been recognized as one of the most reliable indicators of corruption, as well the most widely-used (Rose-Ackerman, 1999; Treisman, 2000; Sandholtz and Koetzle, 2000). The CPI-scales are adjusted to a range of 0 to 10, where 0 indicates the lowest and 10 the highest level of corruption. I am aware of the limitations of using these macro-indices and have discussed them comprehensively in chapter 2.3. On the basis of data availability, my investigation includes 37 European countries over a period of 19 years (1995-2013). Through descriptive analyses, I have demonstrated that the extent of corruption varies widely across European countries and time and that there are significant differences between certain states. For instance, countries such as Denmark, Finland and Sweden received very low degrees of corruption by Transparency International, while states such as the Ukraine or Albania are still exposed with high scores of corrupt activities. Taking a closer look at the patterns of the development of corruption in recent years, it is striking that both new and established democracies in Europe show varying levels of corruption. Even though the extent of corruption is still higher in Central, Eastern and Southern Europe than in the Northern countries, corruption is not a problem that only young democracies are confronted with. Examining the patterns of corruption development over time, it is also striking that the extent of corruption in European states is continuously rising. While in 1995 the (transformed) average score was 2.91 among European states, the level deteriorated to 4.0 in 2013. As an expectation to this trend, corruption levels have significantly worsened in Western European countries such as Spain, Greece and Portugal. However, although corruption still seems to be part of social norms and traditions in much of Central and Eastern Europe, there are some countries in these regions such as Albania, Estonia, Latvia or Lithuania, where the degree of corruption has improved in recent years. As my study reveals, this can especially be attributed to their economic development, 4.1 Macro Model of Corruption in Europe 161 integration in the European Union, women in parliaments, Protestantism and improvements in their democracy scores. I was able to illustrate that a variety of these factors appear as causes of corruption, obviously with varying emphases. In terms of the relationship between a country’s economic development and EU-membership, I have provided evidence that a higher economic development and membership in the EU is associated with lower levels of corruption in Europe. For instance, panel estimates of the macro model show that as one moves up one unit on the variable economic development, the extent of corruption is expected to decrease by around 0.42 points. Based on this result, I can conclude that by improving their economic conditions countries are very likely to improve their corruption scores. Furthermore, my findings indicate, that especially after becoming a member in the European Union, the corruption scores of several countries such as Cyprus, the Czech Republic, Estonia, Latvia, Lithuania, Poland, Slovakia and Slovenia have significantly improved. A similar relationship is observed in a previous study by Kubbe (2013) between the extent of corruption and a country’s WTO- and OECD-membership. Since their admission to the EU, most of the post-communist countries such as Albania, Croatia, Estonia, Georgia, Latvia, Lithuania and the Ukraine seem to experience fewer incidences of corruption as evidenced by lower corruption scores. Thereby, international integration has been shown to tend to prevent corruption. These findings confirm my assumed hypothesis and further support the work of such researchers as Sandholtz and Gray (2003) and Kostadinova (2012). Following rational-choice and cultural approaches, I have revealed that on the one hand, a country’s involvement in international organizations affects the extent of corruption in that country by offering economic incentives. On the other hand, international integration seems to affect a country’s level of corruption in a normative way, by creating norms and values that stigmatize corrupt conducts. This also corroborates cultural approaches that assume that culture interacts with corruption through formal and informal institutions. More specifically, from a cultural perspective, corruption norms are conceived of as a specific form of social norms that dictate the extent to which individuals engage in, and expect others to engage in, corruption (Elster, 1989; Banuri and Eckel, 2012). In particular, countries that are more integrated into Western international networks are more exposed to both economic and normative pressures against corruption. They are more tied into international networks of exchange, communication, and organization. Thus, to reduce a society’s 4 The Bigger Picture of Corruption in Europe 162 level of corruption, the international community should promote its integration into international organizations such as the EU, OECD or WTO. This also implies that rules such as the Copenhagen Criteria, which define whether a country is eligible to join the European Union or not, seem to have a significant influence on a country’s level of corruption. In this context, the establishment of the rule of law in a country is not compatible with widespread levels of corruption, and if countries attempt to join the EU, they have to minimize corrupt activities as far as possible. Therefore, the admission of countries into organizations with high anti-corruption standards such as the European Union seems to be an overall efficient anti-corruption instrument because international pressure tends to produce behavioral changes in countries regarding their corruption levels (Kostadinova, 2012). This indicates that international integration generally plays a more important role in the reduction of corruption. In terms of the political variables, the degree of democracy and the percentage of women in parliaments are influential in explaining the levels of corruption in European states. In the macro model, the degree of democracy is one of the most important contributors to the reduction of corruption levels. The panel estimates of the macro model show that as one moves up one unit on the variable degree of democracy, the extent of corruption is expected to decrease by around 0.12 points. In other words, more advanced democratic structures and institutions lead to lower levels of corruption. This is attributed to the fact that in democratic states, principles such as equality, fairness, transparency, checks and balances, and accountability are more strongly fostered than in authoritarian regimes that are characterized in particular by strong hierarchies. This corruption-democracy-nexus confirms economic and institutional considerations that attribute the emergence of corruption to a lack of competition in institutions that provide public officials with incentives and opportunities to nurture bribes (Klitgaard, 1988; Shleifer and Vishny, 1993; Rose-Ackerman, 1997). Accordingly, corrupt actions do not arise due to low morale, but rather by a bad arrangement of rules and institutions that cannot prevent illegal and harmful human actions. Therefore, specific components of democracies and democratization processes include necessary conditions for honest governments because their institutions tend to include corruption-restraining mechanisms. My results again confirm cultural approaches that claim democratic societies and polities are often committed to norms and values of justice and equal opportunities that are in opposition to corruption norms (Uslaner, 2006). Persson et al. (2012, pp. 16–17) have also indicated 4.1 Macro Model of Corruption in Europe 163 that “increased transparency will risk increasing the level of corruption since this will make people even more aware of the problem, encouraging even former noncorrupt actors to take part in the corrupt game.” My findings also reveal that a higher percentage of women in parliaments alleviates corrupt actions in European states. This result is in accordance with the findings of Swamy et al. (2001) who have illustrated that higher female labour participation leads to less corruption in general. This could be explained by two assumptions: On the one hand, women are less involved in corrupt transactions than men because male politicians tend to possess more senior positions than female politicians (Agerberg, 2014). Moreover, women are less integrated in well-established networks in which corruption often flourishes (Alemann, 2004). According to this rationalization, however, it can be assumed that if women were to ascend to even higher positions and become part of these networks, or build new ones, their actions could also change to adopt more corrupt practices. On the other hand, I assume that this negative relationship is caused by other aspects of liberal democracy that accompany the protection and equality of women’s political rights (e.g. Norris et al., 2002; Sung, 2003; Agerberg, 2014). Accordingly, it might be the fairer systems, which are created within democratic systems, and not women’s greater degree of integrity that explain why corruption is lower in countries where more women are in parliaments. As a result, my findings support previous studies demonstrating that higher gender equality in the political arena and in the labor market decreases Europe’s corruption level. This assumption reflects the interplay between a country’s level of democracy and the number of women in parliaments, where more democratic systems have a higher percentage of women in parliaments. In other words, gender equality may be conducive to democracy by promoting a less hierarchical cultural milieu for decision making (Norris et al., 2002). A third explanation of the relationship between the extent of corruption and the percentage of women in parliament is argued by Gottfredson and Hirschi (1990) and Paternoster and Sally (1996) claiming that women seem to be more honest or are more moral and risk-averse than men by nature. However, this claim should be carefully considered further by behavioral or psychological studies. This finding could also be related to cultural rather than gender-based explanations suggesting that gender differences are less universal and more culture-specific (Alatas et al., 2009; Alhassan-Alolo, 2007). 4 The Bigger Picture of Corruption in Europe 164 The analysis of the socio-cultural variables illustrate that religion is a strong predictor of corruption levels in Europe.74 The variables a society’s percentage of Catholics, Orthodox, Protestants and Muslims have a significant influence on the extent of corruption in European states. As Dreher et al. (2007) has theorized, religion may shape social attitudes towards social hierarchy and family values and thus, in turn, affect corrupt practices. My findings confirm this and indicate that societies with a higher percentage of individuals of Catholics, Orthodox and Islamic faith show higher levels of corruption, while the relationship in Protestant societies such as Denmark, Sweden or Norway seems to be opposite. In terms of cultural approaches, I attribute this relationship to the fact that compared to Orthodox, Catholic or Muslim societies, Protestant societies show less hierarchy and are strongly characterized by egalitarian norms and values. Therefore, they are less prone to tolerance towards power abuses and corrupt transactions (Paldam, 2001). Furthermore, the Protestant church has traditionally been separated from the state and has played a role of opposition to the abuses of the government (Treisman, 2000). The Puritan aspects of this religious tradition could also have a corruption-preventing effect on both providers and receivers of bribery (Skaaning, 2009). Moreover, Protestants tend to be less embedded in social networks that seem to be a breeding ground for corruption (Lambsdorff, 2002). As a result, societies that indicate more egalitarian and individualistic features are more likely to show lower levels of corruption. The analysis of the historical variables of corruption reveals that a country’s years of democracy and communist past are significant in explaining corruption in European states. As assumed, there is a strong negative relationship between the durability of democratic systems and the extent of corruption. The panel estimates show that as one moves up one unit on the variable years of democracy, the extent of corruption is expected to decrease by around 0.23 points. This complies with the empirical analyses of Treisman (2000), Blake and Martin (2006) and Pellegrini and Gerlagh (2008). For instance, Treisman (2000) has observed a significant 74 However, data on religion has to be cautiously interpreted. For instance, the membership in religious communities does not necessarily imply that people also take part in the activities and institutions of their churches. In this context, studies reveal that Catholics, Orthodox people and Muslims go more frequently to churches than Protestants. For instance, in Germany a lot of people are official church members, but do not actively participate in church events (e.g. Pickel, 2009). 4.1 Macro Model of Corruption in Europe 165 impact of the distant past on the extent of corruption. He has illustrated that a long duration of democracies seems to be necessary to significantly reduce corruption levels. This result implies that democratic structures do not only decrease levels of corruption, but that this effect is also strengthened by the duration of democratic principles. In other words, the longer a democracy lasts, the less corrupt it is. In terms of sociological and historical institutionalism, this demonstrates that individuals are expected to get used to formal and informal democratic institutions and adapt their actions to these structures (Harrison and Huntington, 2000).75 This also reconfirms that the relationship between institutions and actors is reciprocal and cyclical (Groenendijk, 1997; Scharpf, 2006). Thus, institutions explicitly affect the preferences and actions of individuals through rules, norms or other social frameworks and vice versa (March and Olsen, 1989; Hall and Taylor, 1996).76 To conclude, democracies are not per se free of corruption and do not necessarily exhibit honest governments and politicians, but they have fewer problems with corruption reflecting the duration of democratic rule. With regard to a country’s communist past, the hypothesis that corruption levels are higher, if the country has a communist past is confirmed. This indicates that a country’s communist past fosters the growth of corruption levels and that post-communist countries seem to be still susceptible to corrupt practices. With regard to cultural approaches, this could particularly be attributed to the heritage of economic and political decisionmaking under communist rule that created structural incentives for engaging in corrupt actions and became such a widespread fact of life that they became rooted in the culture and social norms (Sandholtz and Taagepera, 2005). It can also be hypothesized that the higher average level of corruption in post-communist countries are not related with post-communism per 75 Moreover, the role of a society’s previous democratic experience in the process of political consolidation is discussed by certain scholars such as Huntington (2002). It has been indicated that longer and more recent democratic experiences are more favorable for the consolidation of a political system than shorter and remotes ones. Putnam (1993) suggests that democratic traditions serve as a reservoir for the next generations for the right patterns of beliefs and behaviors. 76 This is also consistent with an “institutional learning” approach (e.g. Rustow, 1970) that implies that people embrace the values into which institutions socialize them. Following this assumption, institutions need time to exert a socialization effect. Accordingly, Welzel (2013, p. 126) suggests that ”democracy should shape values through its long-time endurance, not its momentary presence.”. 4 The Bigger Picture of Corruption in Europe 166 se, but that these countries, as Treisman (2003, p. 22) suggested, have “bad governments, largely because they are poor and lack a post-war history of democracy.” In sum, culture changes slowly and new bureaucracies and institutions are not created from scratch. People in general had internalized certain practices that after more than 20 years still exist (Skaaning, 2009). Furthermore, Paldam (2002) has found little evidence to bolster the belief that corruption is a persistent phenomenon and so deeply embedded in the culture of the society as to be unchangeable. This is also consistent with assumptions of historical institutionalism implying that corruption evolves over time and is determined by certain lasting traditions and values (Thelen, 1999). The findings also corroborate the arguments of Herzfeld and Weiss (2003), Kostadinova (2012) and Skaaning (2009) who have described corruption as cultural heritage and suggested that corrupt actions are path-dependent. Moreover, it demonstrates that in societies with high levels of corruption, people have greater expectations and a higher estimated probability that, for instance, a given public official will engage in corrupt acts (Fisman and Miguel, 2007). These facts also indicate a cultural transmission of corruption, implying that individuals from societies in which corrupt transactions are quite common are more likely to engage in and expect others to engage in corrupt acts (Hauk and Saez-Marti, 2002; Barr and Serra, 2010). This is in line with the insights offered by Persson et al. (2012) who describe corruption as a collective action problem. They argue that people behave in a corrupt manner because they understand the situation as a collective action where it makes little sense to be “the only one” that refrains from using or accepting bribes. In the words of Persson et al. (2012, p. 16) this means that “the important thing will be to change actors’ belief about what ‘all’ other actors are likely to do so that most actors expect most other actors to play fairly.” Consequently, reducing levels of corruption would imply a change of specific practices and habits that are deeply embedded in a society’s culture and its institutions. Following rational-choice approaches, the implementation of certain institutional control mechanisms at places that are known for corrupt practices such as registration offices or passport controls present a useful instrument to prevent corruption. For instance, the integration of commissioners into public institutions that help to minimize and manage misconducts is a very successful to strategy for handling corruption (Shleifer and Vishny, 1993; Bobkova and Egbert, 2012). Moreover, whistleblowing, including reports of alleged dishonest or illegal activities occurring in public or private institutions, should generally be socially and 4.1 Macro Model of Corruption in Europe 167 ethically more established in a society. Individuals reporting illegal activities need stronger protections and do not have to follow bureaucratic chains of command to make reports (Collier, 2002). Micro Models of Corruption in Europe To explain perceived corruption at the individual level, I ran several multilevel models that indicated level specific relations when intercepts and slopes varied for the level analysis. Multilevel models present an appropriate analytical procedure for analyzing corruption at certain levels. By allowing for residual components at each level, multilevel modeling takes the existence of hierarchical data structure into account (Hox, 2002). To measure corruption at the individual level, I used the item “How widespread do you think bribe taking and corruption is in this country?” from three waves of the World Values Survey (1994-1999; 1999-2004 and 2005-2009). The fourth wave of the WVS does not include the item for European states. The item is scaled on a four-point scale from 1 (“no public officials engaged in it”) to 4 (“almost all public officials are engaged in it”) (World Values Survey, 2015). Combined together, these polls provide data on 25 European societies. However, in the analysis only 20 countries were finally included because of missing data availability of the independent variables. On average, the European models, i.e. the random intercept, random intercept and slope, and the final multilevel model, include 20 European countries and approximately 22.000 observations. Through descriptive analyses, I have first demonstrated that the extent of perceived corruption vary widely across European societies. It is striking, that the findings are very similar to the extent of corruption at the country level. Similar to the country level, there are also significant differences in the perception of corruption between West and East European states. The highest extent of perceived corruption are found in post-communist countries such as Macedonia (3.39), Lithuania (3.33), and the Ukraine (3.30), whereas the countries with the lowest extent of perceived corruption turn out to be again the Scandinavian countries such as Norway (2.01) and Finland (2.18). Additionally, levels of perceived corruption are not exceptionally lower in Southern Europe than in post-communist societies. On the whole, European societies are characterized by diverging values of perceived corruption extent. 4.2 4 The Bigger Picture of Corruption in Europe 168 To explain the differences among the respondents in their estimation I ran several multilevel models. According to economic and cultural approaches, I assumed that corruption is likely experienced differently depending on socio-demographic factors, social values, norms, and attitudes. The analysis has demonstrated that an individual’s employment status, level of interpersonal trust, satisfaction with the financial situation, and the justification of bribery are significant in the explanation of the extent of perceived corruption. While an individual’s and employment status and justification of bribery show a positive relationship with the extent of perceived corruption, the variables level of interpersonal trust and satisfaction with the financial situation are negatively related to the dependent variable. Yet, socio-demographic characteristics such as an individual’s gender, age and level of income do not show a significant relationship with the perception of corruption. I have demonstrated that interpersonal trust is constantly the strongest predictor of the extent of perceived corruption. The estimates of the micro model show that as one moves up one unit on the variable interpersonal trust, the perceived extent of corruption is expected to decrease by around 0.24 points (Model 1). This implies that people who have high levels of interpersonal trust show lower levels in the perception of corruption. This may be due to the fact that societies with high levels of interpersonal trust such as the Scandinavian countries have lower levels of corruption as well. Consequently, interpersonal trust is a significant predictor of the perception and the extent of corruption. This finding confirms the work of scholars of sociological approaches who argue that decisions about whether to engage in corrupt transactions are particularly influenced by social norms and cultural values (Husted, 1999; Getz and Volkema, 2001; Banuri and Eckel, 2012). Based on this result, my analysis indicates that trust seems to be a good control mechanism of corruption within a society. Generally, trust is a central component of social capital and a value that expresses the belief that others are part of your moral community (Uslaner, 2006; Putnam, 1993). Therefore, if corruption is perceived as an unethical and immoral behavior that is harmful for the community, then people are more likely to refrain from corrupt acts. Societies such as those in Sweden or Norway, for example, have high levels of interpersonal trust and are characterized by egalitarian values. Therefore, corrupt actions would imply that some individuals try to take advantages by using illegal instruments that do harm to other people in their society. However, from a rational-choice point of view, trust is simply based on the expectation that 4.2 Micro Models of Corruption in Europe 169 others behave predictably (Hardin, 2002). Therefore, trust is also the basis for cooperation with people who are not like yourself (Putnam, 1993; Uslaner, 2006). As a result, for Europe this finding is in line with the results of Paldam and Svendsen (2001), Uslaner (2006) and You (2004) who have found a strong negative relationship between corruption and interpersonal trust, implying that trusting societies have fewer people behaving corruptly. In terms of the variable gender, the assumption that there is a significant relationship between perceived corruption and an individual’s gender could not be confirmed. This contradicts studies of Swamy et al. (2001) and Dollar et al. (2001). Both of these have demonstrated that women are less involved in corrupt transactions and are less likely to condone bribetaking than men. Nevertheless, this result does not contradict the findings on the macro level that indicates a linkage between low levels of corruption and a high percentage of women in parliaments. Instead, it confirms the assumption that the impact of gender on the extent of corruption seems to be overestimated. Thereby, it supports the analyses of Sung (2003) and Alatas et al. (2009) who have argued that differences in corrupt actions seem to be more culture-specific rather than caused by gender differences. In terms of the variable an individual’s age and the perception of corruption, my findings of the European sample have not confirmed previous research by Hirschi and Gottfredson (2000) and Torgler and Valev (2006) suggesting that age is an important indicator of illegal activities. According to my expectations based on rational-choice approaches, the relationships between the extent of perceived corruption and an individual’s (un)employment status is significant. I assumed that an individual’s employment status influences the extent of perceived corruption. This implies that unemployed people tend to perceive higher levels of corruption than individuals having a job. In all models an individual’s satisfaction with their financial situation have indicated a negative relationship with the extent of perceived corruption. This implies that people who are unsatisfied with their financial situation perceive a higher extent of corruption of public officials. As a result, my findings suggest that financial dissatisfaction leads to a more critical or even more attentive perception of the extent of corruption. In this context, I assumed that on the one hand, these people are more critical towards public officials and politicians because they are unsatisfied with their politics. Maybe they blame these officials for their personal situation or they feel unjustly treated by them. On the other hand, I hypothesize that 4 The Bigger Picture of Corruption in Europe 170 people who are unsatisfied with their own financial situation have already engaged in corrupt transactions to strive for higher incomes and are also prepared to accept illegal payments. This would confirm the research of Torgler and Valev (2006) who have provided evidence that people who are dissatisfied with their financial situation tend to be more willing to act illegally. Consequently, people are more experienced and are therefore more “qualified” to estimate corruption levels. These assumptions corroborate rational-choice approaches, implying that actors follow a rationalchoice logic and are often motivated by material interests and commit or refrain from corrupt acts for tangible goods. Nevertheless, this again demonstrates that both the extent of corruption and its perception are culturally influenced and determine individuals’ behavior. It seems that people have greater expectations and a higher estimated probability that, for instance, a given public official will engage in corrupt acts in societies with high levels of corruption (Fisman and Miguel, 2007). These results also clearly demonstrate the cultural transmission of corruption, which implies that individuals from societies in which corrupt transactions are quite common are more likely to engage in corruption and expect others to engage in it as well (Hauk and Saez-Marti, 2002). This is again in line with the arguments offered by Persson et al. (2012) who delineate corruption as a collective action problem. Moreover, the results of the European multilevel model that includes the significant variables of the macro level have demonstrated that the three cross-level terms (an individual’s interpersonal trust and a country’s degree of democracy; the percentage of women in parliaments and a country’s degree of democracy; and a country’s duration and degree of democracy) also common explanatory variables in terms of the extent of perceived corruption. Summing up, the analysis further illustrates the strong linkage between the micro and the macro level because countries in which people perceive high corruption levels (measured at the individual level) are often ranked by high levels of the Corruption Perceptions Index (measured at the macro level). A correlation coefficient of .81 between both indices strongly supports this linkage. In conclusion, cross-national differences in the extent of perceived corruption are not necessarily misinterpreted when they are understood as culture-biased extent of corruption. 4.2 Micro Models of Corruption in Europe 171 Europe’s Bathtub Model of Corruption In this study, the bathtub model of corruption has served as a framework for analyzing the causes of (perceived) corruption on different levels as well as across and within certain countries and across time. It combines the macro and micro level and considers corruption as individual behavior, taking country and personal characteristics in particular into account. Moreover, this model combines economic (e.g. rational-choice theory) and sociological approaches (e.g. cultural theories such as sociological and historical institutionalism) into an interdisciplinary framework and offers an integration and analysis of certain variables at different levels that influence the extent of (perceived) corruption. For the purpose of identifying the factors that influence corruption, specific situations at each level of the model were filled by empirical data. In this analysis, rational-choice assumptions are especially examined by contextual conditions on the macro level (e.g. a country’s economic development) and micro level (e.g. an individual’s level of income, satisfaction with the financial situation) referring particularly to various resources. Cultural and institutional approaches are taken into account by variables such as international integration, a country’s degree and duration of democracy, religion, a country’s communist past, or an individual’s level of interpersonal trust. European-Specific Bathtub Model Transferring the findings of the macro and micro models to the bathtub model of corruption, I have been able to illustrate that the various factors influence the extent and perception of corruption in European states in both longitudinal as well as in cross-national sections and at specific levels (see Figure 36). At the macro level (situation 1 of the model) these factors include the degree of economic development, EU-membership, a country’s degree of democracy, the percentage of women in parliaments, Catholicism, Orthodoxy, Protestantism, Islam, years of democracy, and a society’s communist past. While a high percentage of Catholics, Orthodox, Muslims and a country’s communist past are more likely to enhance corruption over time, a country’s economic development, EU-membership, degree of democracy, the percentage of women in parliaments and a high number of democratic years significantly decrease corruption levels for the period between 1995 and 2013. In terms of the bathtub model, the- 4.3 4 The Bigger Picture of Corruption in Europe 172 se characteristics of European countries shape a certain “situational logic” (Esser, 1993a) by creating specific social frameworks for the actions of individuals. This implies that individuals, in turn, act under these particular social structural and institutional conditions and behave according to a specific social logic at the micro level. Thus, an individual’s action is a consequence of a particular “logic of selection” (Esser, 1993b, p. 8). At the micro level, an actor’s decision was analyzed on the basis of rational-choice and cultural theories that describe an individual’s perception of corrupt actions either as rational or as decisions that are culturally dependent and influenced by certain norms and values. Therefore, individuals follow a certain logic of selection and decide to engage in corrupt activities either by virtue of rationally- or socially-based motives. At the individual level, corruption was measured by the World Values Surveys item “How widespread do you think bribe taking and corruption is in this country?”. According to the results of the micro models, the variables an individual’s (un)employment status, level of interpersonal trust, satisfaction with the financial situation and the justification of bribery are significant in the explanation of the extent of perceived corruption. While an individual’s employment status and justification of bribery indicate a positive relationship to the extent of perceived corruption, an individual’s level of interpersonal trust and an individual’s satisfaction with the financial situation show negative relationships. However, this does necessarily not mean that people with high degrees of justification of bribery are more likely to act in a corrupt manner. Rather, it implies that these people perceive the extent of corruption as more widespread than others. In this sense, they might be more critical and sensitive when it comes to the extent of corruption in a given country. However, a correlation between the Corruption Perceptions Index (transformed) and the aggregated item World Values Survey has shown a high coefficient of 0.82, indicating the linkage between the country and the individual level offered by the bathtub model of corruption. This also suggests that both indices seem to measure the same phenomenon: the extent of corruption. Moreover, the multilevel analysis reveals that the micro and macro level are linked by the following lagged variables: the degree and duration of democracy, women in parliaments and religion. All these variables are expected to affect (perceived) corruption scores at both country and individual level. Following the bathtub model, these variables represent the causal influences for macro factors that work through disaggregated effects at the micro level. With regard to the micro level, I have identified that corruption is likely experien- 4.3 Europe’s Bathtub Model of Corruption 173 ced differently depending on certain values and attitudes. However, sociodemographic features such as an individual’s gender, age and level of income do not show any influence in European countries. Yet, the multilevel analysis demonstrates that an individual’s level of interpersonal trust, satisfaction with the financial situation and justification of bribery affect the individual’s perception of corruption. While interpersonal trust and an individual’s satisfaction with the financial situation decrease the extent of perceived corruption, an individual’s (un)employment status and the justification of bribery increase its perception. Moreover, the multilevel analysis has indicated that an individual’s interpersonal trust and a country’s degree of democracy; the percentage of women in parliaments and a country’s degree of democracy; and a country’s duration and degree of democracy have cross-level effects on the extent of perceived corruption. Finally, the aggregation of the individual actor’s (corrupt) actions leads to situation 2 on the macro level (“collective effects”), where corruption is measured by the transformed Corruption Perceptions Index (“logic of aggregation”). According to the model, the aggregation of the individual actor’s actions can either reproduce social structures and behavioral patterns or modify them. In other words, individual (corrupt) actions reproduce certain structures and patterns that, in turn, affect corruption levels in the future. Thus, levels of corruption can be explained at the macro level via the aggregated individual actions at the micro level. This includes that phenomena of the macro level influence system outcomes through their effect on individuals’ orientations and behavior at the micro level (actor’s action). Thereby, macro-sociological phenomena such as corruption are reconstructed as unintended consequences of individual behavior in respective situations (e.g. Coleman, 1990). This illustrates again the strong path-dependence of corruption. 4 The Bigger Picture of Corruption in Europe 174 The Bathtub Model of Corruption for European Countries Consequently, I conclude, that there are area-specific factors that are unique to European countries affect the extent of corruption. This model can be attributed to historical factors such as the democratic experience of each country, the post-communist past of certain European countries, cultural norms such as interpersonal trust, and religious aspects. Addressing corruption more effectively means to understand these factors in a better way. The findings suggest that the reduction of corruption in the European context seems to be “all about democracy”. The included variables of the analysis are strongly linked to a society’s development of democracy such as a society’s economic development, degree and duration of democracy, the EU-membership and interpersonal trust. For this reason, a “democratic culture” creating anti-corruption norms lead to a lower extent of (perceived) corruption and might be the key to reduce corruption in Europe. Figure 36: 4.3 Europe’s Bathtub Model of Corruption 175 Conclusion and Research Prospects Generally, previous corruption research has consisted of two branches. On the one hand, researchers have focused on the effects of corruption and have indicated that it is detrimental to economic, social and political development (Mauro, 1997; Rose-Ackerman, 1999; Richey, 2010; Uslaner, 2012). On the other hand, corruption studies have concentrated on the causes of corruption that particularly hinder or foster its occurrence (Seldadyo and Haan, 2006). This study is linked to the latter branch of corruption research. With this study, I make several contributions to the current research on corruption. First, a literature review on the causes of corruption has demonstrated that an overall theory for the analysis of corruption does not yet exist and is still needed. Therefore, I have integrated certain theoretical perspectives from different disciplines such as the principal-agent-approaches, and sociological and historical institutionalism to build a solid foundation for analyzing the causes of corruption. I have demonstrated that all of these approaches have great power in explaining corruption in European states. Secondly, it became apparent that empirical analyses of corruption seem to have generally developed separately from theoretical approaches and either neglect important explanations such as cultural considerations or fail to combine them. This inquiry has attempted to build a bridge between these certain perspectives of corruption as well as theoretical and empirical research developments. Thirdly, corruption has often been framed and measured by certain macro level indices, although it takes place between individuals at the micro level. To shed light on the linkage between macro and micro levels, I have considered corruption as a multilevel phenomenon that takes place on different levels of analysis in an interdisciplinary manner through different perspectives and methods. In this context, the bathtub model of corruption has served as a heuristic framework for analyzing the extent of (perceived) corruption on different levels as well as across and within certain countries and across time and provides a contribution to current theoretical and empirical discussions and developments. It has been demonstrated that, on the one hand, the prevailing research focus has been dominated by highly aggregated large-n analyses (Goel and Nelson, 2010; Littvay and Donica, 2011) that tend to neglect significant region-specific characteristics and differences and variations within 4.4 4 The Bigger Picture of Corruption in Europe 176 countries and therefore could lead to biased analyses of corruption from a comparative perspective. On the other hand, researchers have concentrated on certain case-studies that particularly investigated individual cases of corruption and rarely provide generalizable results (Pujas and Rohdes, 2009; Miller et al., 2009; You, 2015). In order to overcome these limitations and find a middle ground, it is of great importance to investigate areaspecific factors of corruption that influence its extent and to focus on individual regions such as Europe rather than on global samples. Due to the detected intra- and inter-European variation and continuously increasing corruption scores in certain countries, European countries presented excellent cases to examine corruption. Contrary to other regions in the world such as Africa, Europe has remained largely unexplored. This study is an extension of previous empirical studies such as Littvay and Donica (2011) or O'Connor and Fischer (2012). When analyzing corruption, I also contribute inter alia to the current research by including new variables that have not yet been taken into account in analyses of corruption. The variables extended the list of indicators and include micro level variables such as an individual’s level of income, the individual’s satisfaction with the financial situation and the justification of bribery. The major findings of my investigations at the macro level can be summarized as follows: As demonstrated, the factors that affect the extent of corruption over time and across European countries include a wide range of variables: Whereas a country’s percentage of Orthodox, Catholics, Muslims and communist past enhance corruption over time, a high economic development, the integration in the European Union, a country’s degree of democracy, the percentage of women in parliaments, and a high number of democratic years, in contrast, significantly decrease corruption levels for the period of 1995-2013. These results confirm previous research from other parts of the world and demonstrate their robustness also for Europe. Yet, I can conclude that there is a model of context factors that influence the extent and perception of corruption that applies to European countries. While previous corruption research has focused predominantly on macro variables, this study also takes micro variables into consideration. In this context, I have identified that corruption is likely experienced differently depending on certain socio-demographic aspects, values, norms, and attitudes. With regard to the micro level, I have identified that corruption is likely experienced differently depending on certain values and attitudes. However, socio-demographic features such as an individual’s gen- 4.4 Conclusion and Research Prospects 177 der, age and level of income do not show any influence in European countries. Yet, the multilevel analysis demonstrates that an individual’s level of interpersonal trust, satisfaction with the financial situation and justification of bribery affect the individual’s perception of corruption. While interpersonal trust and an individual’s satisfaction with the financial situation decrease the extent of perceived corruption, an individual’s (un)employment status and the justification of bribery increase its perception. Moreover, the multilevel analysis has indicated that an individual’s interpersonal trust and a country’s degree of democracy; the percentage of women in parliaments and a country’s degree of democracy; and a country’s duration and degree of democracy show cross-level effects on the extent of perceived corruption. This displays that not even structural components matter, but that corruption exists, persists, and varies significantly by culture. The channels through which culture and corruption interact are manifold. As suggested, these channels include formal and informal institutions such as interpersonal trust (Elster, 1989; Banuri and Eckel, 2012). Together these findings reveal that participating is not just about profit and people do not act for purely material reasons such as money or immaterial resources such as power and prestige. It is also about culture and traditions that generally determine an individual’s decision to commit or refrain from corrupt acts. Furthermore, the multilevel models reveal that between 80-90% of the variance of (perceived) corruption is attributed to differences between countries and 10-20% to differences between individuals. This again confirms that the roots of corruption lie on the country or cultural level rather on the individual level. With regard to the measurement of corruption, my study also point out that there is a high correlation between the dependent variables of the country level, measured by the CPI, and individual level, measured by WVS survey data. Comparing rankings that are produced by experts at the aggregate level with perceptions of citizens at the individual level, my study shows that both perceptions are highly correlated. Although both variables are measured differently, this demonstrates that they are strongly linked and suggests a strong validity of these measures. In fact, both indices seem to measure the same phenomenon: the extent of perceived corruption. Prospectively, researchers analyzing corruption can additionally rely on the perceptions of citizens as well, however, always be considered in connection with survey data by experts (see also Charron 2015, p. 21-22). 4 The Bigger Picture of Corruption in Europe 178 Although it is not directly intended to give policy recommendations with this work, some implications for policy do follow from my results. Particularly, my analysis indicates that enhancing an individual’s economic situation is an adequate way of reducing and preventing corruption, or as Uslaner (2010, p. 246) suggests: “The key to reducing corruption seems to involve making people less dependent upon it.” Only a society where relatively few people live in poverty offers the requisites for equal economic, political and social participation and therefore equality. More precisely, a highly unequal distribution of key values such as income, wealth, education and knowledge are equivalent to inequality in the distribution of key political resources and hence unfavorable to competitive politics (Lipset, 1953; Dahl, 1991). Therefore, in addition to the level of economic development of a country, the distribution of material resources is regarded as important for the prospects and benefits of democracy that in turn hinders corruption development (Diamond, 1992). However, as my study has indicated, democracy does not necessarily guarantee honest governments and corruption-free societies. In contrast, the duration of democracy is the decisive element that improves corruption levels in the long-term. In sum, by providing the institutional component of people power, democracy leads to a higher levels of transparency and enables the civil society and social engagement of groups including non-governmental organizations, media and the press, to call attention to corruption, sensitize the population and act as watch-dogs (Ahrend, 2002). As studies have indicated, in mature democracies these social organizations and movements have become a constant source of influence on government, keeping elected officials under permanent pressure in terms of accountability and responsiveness (McAdam et al., 2001; Welzel, 2013). In this context, Collier (2002, p. 27) has also revealed that “an empowered civil society playing a vital role in elite accountability emerges as the foundation to building commitment rules.” Social empowerment77, especially through mass citizen participation, is therefore essential in the fight of corruption. A strong civil society cultivates anti-corruption commitment rules that, in turn, lead to self-enforcing mechanisms where the ruling elites make it their duty not to behave corruptly and civil society takes this promise as their corresponding right. Ci- 77 Social empowerment means “strengthening civil society in order to enhance its political and economic vitality, providing more orderly paths of access and rules of interaction between state and society, and balancing economic and political opportunities” (Johnston, 1998, p. 85). 4.4 Conclusion and Research Prospects 179 vil society is therefore the chief manifestation of a vital democracy and a main source of governmental accountability and responsiveness (Putnam, 1993; Collier, 2002). However, Johnston (2012, p. 342) points out that “any nation relying on democratic processes to check corruption must face the possibility that the new order will create corruption risks all its own, and that voters will be all too willing to re-elect leaders of dubious integrity who nonetheless ‘deliver the goods’. The key is not the formal hardware of democracy, but rather fairness, loyalty, legitimacy, and credible accountability – the values that make democracy worth pursuing, and corruption worth worrying about, in the first place.” In this context, “values-based corruption control“ is required which is based on the notion that values, social sanctions, and widely shared conceptions of right and wrong should play an important role, alongside laws and punishments, in guiding the uses of public power and resources. Banuri and Eckel (2012, p. 7) suggests that “For a government that seeks to inhibit corruption, the goal is to devise formal institutions that can reinforce existing social norms.” Consequently, fighting corruption consists of the combination of formal democratic institutions that provide transparency and accountability, but also include monitoring and sanctions mechanism against corrupt actors. Preventing corruption implies fostering informal institutions such as interpersonal trust. Consequently, a specific “democratic culture” including certain norms and values such as interpersonal trust and social and economic equality is the most important contributor in the fight of corruption in Europe. This democratic culture goes along with the degree and especially the duration of democracy, economic development and EU-membership which overall represent a system of democratic values or anti-corruption norms. In other words, a democratic culture as specified above hinders the growth of corruption and acts as remedy by generating strong democratic institutions and norms and values as well. As a result, according to the reduction of corruption in Europe, yes, it is all about democracy. For this reason, future research should strongly focus on the theoretical and empirical linkage between corruption and democracy and develop adapted theories and analytical framework based on this relationship. Turning the tables, good governance and government performance can, in turn, help to improve the democratic situation and enhance citizens’ trust in political institutions because citizens who perceive clean and honest governments, higher levels of fairness, satisfaction, and brightening 4 The Bigger Picture of Corruption in Europe 180 economic prospects develop higher trust in institutions (Moreno, 2002; Manzetti and Wilson, 2007; Tavits, 2008). Overall, my analysis provides a diagnostic as well as broad, yet coherent framework of the factors that affect corruption in Europe. This framework can be used as a template for future analyses such as case studies with stronger focus, for instance, on the underlying mechanisms of corruption. However, one should keep in mind, that “Corruption can only be contained, but never eliminated”, as Kreuzer (1996, p. 110) suggests Limitations and Implications for Future Research This study is not without weaknesses that should also be addressed in future research. First, when interpreting the results of this study, it is important to keep in mind the limitations of current corruption research. The evidence shows that corruption is still a shadowy and highly complex field of research whose definitions, methods and measurement are limited. It is of particular importance to be aware of potential restrictions to analyses, especially of statistical methods. In this context, future research on corruption has to find alternative, more precise definitions and measurements of corruption that take into account cultural specificities which, as my findings indicated, play an important role in the extent of perceived corruption. Improving measurement accuracy is essential for advancing research on the causes of corruption across time. A step in this direction is the sharper differentiation between an individual’s perception and the actual level of corruption in a country. Another weakness is the generalizability of my results. European countries are unique in many norms and values that are influenced and driven by institutions and organizations such as the European Union that create common understandings of how to define and recognize corruption and attitudes towards corrupt actions. Yet, it is very important to show the factors that affect corruption in certain parts and regions of the world such as Europe (see also Charron, 2015). Fourth, my analysis underlines the importance of interdisciplinary studies. As shown, the combination of different approaches and qualitative and quantitative methods and several theories and perspectives (e.g. economic and sociological theories) presents a further step in analyzing and explaining corruption. Although rational-choice approaches or cultural and institutional theories can significantly explain the extent of corruption, 4.4 Conclusion and Research Prospects 181 further corruption research needs a stronger focus on anthropological and psychological considerations (Rothstein and Torsello, 2013). Our knowledge regarding the relationship between gender, honesty, risk aversion and interpersonal trust has to be deepened. With regard to this, experimental approaches have a great potential and serve a useful purpose to elicit causes of corruption. As a complementary mode of investigation, they can help to identify specific factors that foster or hinder corrupt actions of individuals (Armantier and Boly, 2010; Alatas et al., 2009; Banuri and Eckel, 2012). Fifth, prospectively, it is also important to have a closer look at the individual countries and specific areas of societies. Quantitative research on corruption should take multilevel approaches more closely into account and focus strongly on the individual level. In fact, corruption occurs between individuals and does not take place at the macro level. Moreover, further research is needed regarding corruption at the institutional and organizational level of analysis in public organizations and institutions that still remained uncovered. Scientists have to focus on discovering the underlying mechanisms of corruption. This particularly implies a concentration of research on country-specific cultures, values and attitudes. Consequently, future research should strongly focus on the research of the third sector that has emerged as a central player in the battle against corruption by taking active positions against officials’ ethical violations and by keeping the anticorruption discourse alive (Kostadinova, 2012). Thanks to greater availability of data, there are a range of further variables that could be analyzed in the research of corruption. In addition to quantitative analyses, a qualitative focus should be applied in order to create a more in-depth approach to discover causal relationships with corruption. Although all included independent variables are time-lagged, further research should have closer looks at causations. In particular, nested analyses are very qualified for further research. Case studies that enable researchers to elicit the mechanism behind corrupt actions within the broader European sample should be conducted. In particular, case comparison is vital to theory development, and qualitative approaches to comparison are well suited for the task. In this way, almost as a by-product of contextualized comparisons, new variables can be discovered and introduced as parts of new explanatory variables and hypotheses. Although statistical approaches achieve systematic comparisons, they cannot discover new variables (George and Bennett, 2005). That is why quantitative analyses have limitations and should not be overstrained – especially concerning 4 The Bigger Picture of Corruption in Europe 182 phenomena such as corruption. Furthermore, scientists in the field of corruption should focus on specific areas and regions and avoid using worldsamples, that, as my findings have illustrated, neglect or bias certain country-characteristics such as traditions and values as central features in the explanation of corruption. In sum, research on corruption remains a challenging enterprise that needs to be pursued further. 4.4 Conclusion and Research Prospects 183

Chapter Preview

References

Abstract

International studies often point to Europe for low levels of corruption. However, recent scandals in nearly all European states illustrate that corruption continues to be on the rise. The author investigates the causes of corruption in Europe. The analysis indicates that a country’s contextual conditions such as the economic development, the degree and duration of democracy or historical factors like the post-communist past strongly influence Europe’s level of corruption. Furthermore, corruption is likely experienced differently depending on interpersonal trust and the justification of bribery. The findings reveal that a bundle of factors adding up to a specific “democratic culture” hinders the growth of corruption by generating strong democratic institutions and fostering citizen norms and values aimed at monitoring and sanctioning corrupt actors. As a result, democracy promotion is the best remedy against corruption spread in Europe.

Zusammenfassung

Auch wenn europäische Staaten vergleichsweise geringe Korruptionswerte aufzeigen, verdeutlichen Skandale immer wieder, dass Korruption ein großes Problem darstellt, mit dem auch Europa stark zu kämpfen hat. Die Autorin untersucht daher die Ursachen von Korruption auf dem europäischen Kontinent. Verschiedene Analysen zeigen, dass Kontextfaktoren eines Landes wie dessen ökonomischer Entwicklungsstand, der Demokratisierungsgrad und die jeweilige Dauer oder historische Faktoren wie die kommunistische Vergangenheit das Auftreten von Korruption stark beeinflussen.

Darüber hinaus spielen interpersonales Vertrauen und die Rechtfertigung von Bestechungszahlungen eine erhebliche Rolle in der Wahrnehmung von Korruption. Insgesamt zeigen die Befunde, dass letztendlich eine „demokratische Kultur“ der Schlüssel im Kampf gegen Korruption in Europa ist. Diese fördert demokratische Institutionen sowie Normen und Werte, die darauf abzielen, korrupte Akteure zu kontrollieren und sanktionieren.