Religion and Democratisation: Eastern Europe in Comparative Perspective

University of Mannheim Chair for Political Science and International Comparative Social Research Dr. Katarzyna Lasinska lasinska@uni-mannheim.de Tel.: +49 621 181 2618 Religion and Democratisation: Eastern Europe in Comparative Perspective Paper prepared for delivery at the ECPR General Conference Science Po, Bordeaux 4-7 September 2013 Abstract: This paper examines the relationship between religion and social participation in Central and Eastern Europe after the liberalisation of political systems in the 1990s. The central question is whether the religious revival can contribute to more participatory citizenship in this region, and indirectly support democratisation processes. The empirical analysis of World Value Survey data shows, that religious revival in this region concerns mostly Orthodox religious tradition. Overall, we can draw three main conclusions. First, the cultural aspect of religiosity (adherence to a religious tradition) is a better predictor for secular social participation than the structural aspect of religiosity (churchgoing). Secondly, in contrast to Western Europe, Protestants are not more socially engaged in secular sphere in this part of Europe. Thirdly, Orthodoxy is the most favourable religious context for secular social participation in post-communist societies. In sum, religion can support democratisation processes in Central and Eastern Europe, and especially religious revival in Orthodox societies is a precondition for the development of civil society. Key words: social participation, religiosity, religious revival, Catholicism, Orthodoxy, Protestantism, Eastern Europe, democratisation

1 Introduction The third wave of democratisation reached Central Eastern European (CEE) countries in the 1990s. One of the triggers of systemic transformations in this part of Europe was religion in particular the Catholic Church, which supported anti-communist opposition in Middle European societies, for instance in Poland and Czechoslovakia. The communist regimes collapsed one by one and were successively replaced by new political systems either liberal democracies (e.g., Poland, Czech Republic and Hungary) or defective democracies like in Russia and Ukraine (Merkel 2004, 51). Irrespective of their current brand of democracy, the process brought people a prominent gain: freedom (of speech, association, religion, etc.). The release of restrictions encouraged people to return to religious practices: a major religious revival occurred in Russia and many of the other successor states of the Soviet Union, as well as in Bulgaria, Croatia and Slovakia (Müller 2009, 75). Religion still seems to play an important societal role in a number of countries in this part of Europe. Considering the positive impact of religiosity on social participation (Putnam 2000, 69; Smidt 2003, 216ff.; de Hart/Dekker 2005, 176), it would be interesting to examine the impact of this religious revival on civil society and, consequently, on the consolidation of democracy in the region. This paper explores the relationship between religiosity and social participation in CEE after the liberalisation of the region s political systems. The first part reviews the literature on the role of religion in social participation, including a discussion of the impact of different denominations. The second section presents the results of the empirical analyses of World Value Survey (WVS) data. The third part concludes that religion indeed plays a positive role in enhancing social participation in CEE, and in contrast to Western Europe Protestantism seems to be far less relevant for the development of participatory citizenship. In addition, the structural aspects of religiosity are less important for explaining secular participation. In summary, the religious revival in this part of Europe can be seen as a positive factor that stabilises democracy, though under specific conditions. 2 Religion and social participation 2.1. Why is religiosity positively related to social participation? Despite progressive secularisation, especially in Western democracies, religion still plays an important role in integrating modern societies. US society is a prominent example of the relevant relationship between social participation and religiosity, and social church-based engagement is one of the most popular kinds of interactions among American citizens (Cnaan et al. 2003, 21). Although social engagement is more concentrated in the secular sphere in European societies (Roßteutscher 2

2009, 412-413), religious people are more likely to be socially engaged in (mostly secularised) Western Europe than non-religious people. How do scholars explain the positive relationship between religiosity and social engagement? In general there are two streams of argumentation: structural and cultural. First, people who attend religious services on a regular basis are included in church-based networks. This provides them with more opportunities to be recruited for social activities, because they develop personal contacts with active churchgoers in their religious community. Churches and church-based organisations (particularly institutional and state churches) also have established structures and procedures involving social activities, such as charity or social projects. Therefore, newly recruited people can join existing institutional structures. In other words, attending religious services gives an individual access to ready-made structures for social participation. Secondly, a cultural line of argumentation postulates that religious persons are supposed to be intrinsically motivated to engage socially. Religions provide people with a moral, normative foundation that stresses the importance of compassion and love for one s neighbour, which serves as a good motivation for involvement in social activities. Few authors explore in detail this religiondriven motivation for social engagement among religious people: religious worldviews and practices depending on their orientations have the capacity to diminish the need for material incentives when group cooperation is anchored in religious beliefs (Harris 2003, 122). In sum, religion meets at least two out of three of the motivations for participation pointed by Verba, Scholzman and Brady in their Civic Voluntarism Model: recruitment through church-based networks and religiously driven motivation 1. In other words, religious people are more likely to be socially engaged, because somebody asked and because they want (Verba et al. 1995, 269). These claims concerning the positive role of religion on social participation find empirical evidence in cross-national studies. Irrespective of their religious denomination, those who attend religious services are more likely to be engaged socially (van Deth 2004, 307-308; Norris/Inglehart 2005, 184ff.). A more detailed study of social participation shows that organisations with a cultural and social involvement are more interesting for churchgoers in several European countries and the United States (Gabriel et al. 2002, 116ff.), while participation in other types of organisations (such as interest groups or sport and recreation organisations) are less interesting for church attendants. Most cross-national studies of participation concentrate on the structural aspect of religion and consider the frequency of churchgoing as a measure of religiosity, ignoring the cultural aspect religiously driven motivation for social participation. Yet religiously driven motivation has a stronger 1 The third motivation according to Civic Voluntary Model are resources. 3

impact on social participation than churchgoing, especially in Catholic societies (Lasinska 2013, 200ff.). However, religious motivation is much more complicated to measure empirically than churchgoing. Therefore in this paper I use the usual indicator of religiosity church attendance while cultural aspects are captured by adherence to a particular religious denomination. 2.2. Why are Protestants more socially engaged? Three religious denominations traditionally dominate Eastern Europe: Catholicism, Protestantism and Eastern Orthodoxy. Though they have the same Christian roots, all three differ in their religious rituals, tenets and the structure of their churches. But do these differences affect the religious impact on social participation? Empirical studies provide general evidence that Protestants are more engaged than other Christians. In Western Europe, for instance, traditionally Protestant Scandinavian societies are more socially engaged than southern Catholic countries (Gabriel et al. 2002, 51-52; van Deth 2004, 306; European Commission 2005). Several studies on social participation also provide empirical evidence of similar patterns at the individual level (Gabriel et al. 2002; van Deth 2004, 308-309). Yet how are these differences explained? The most commonly used explanation points to differences in the structure of the churches: Protestant churches are more egalitarian and their religious communities are smaller, so it is easier for the members of religious communities to get engaged (Verba et al. 1995, 245). On the other hand, in sizeable Catholic parishes that are embedded in the hierarchical structure of the Catholic Church, lay participation is less encouraged by the church (Verba et al. 1995, 245). A more detailed European study on relationships between religious engagement and social capital provides a new perspective on this simple argument (Roßteutscher 2009, 89). Roßteuscher argues that Protestant denominations in Europe such as Evangelical, Baptist, Pentecostal or Calvinist differ substantially in their structures. And in Europe, certain Protestant churches have hierarchical structures that are similar to the Catholic Church. So the differences in social participation among denominations should instead be explained by religious tenets and the values they espouse. In general, Protestant religious traditions promoted much earlier an individualistic worldview than Catholicism: the seeds of individualism were manifest much earlier in Protestantism. In contrast to Catholics, Protestants are personally responsible for God in religious matters (Jagodzinski/Dobbelaere 1995, 81). And this emphasis on individualism gives Protestant believers a stronger motivation to be socially active. The impact of Eastern European Orthodox traditions on social participation is a poorly explored topic. Since we can assume that Orthodoxy does not stress individualistic values in a special way, in this 4

respect, Orthodox and Catholic traditions are close to each other (and differ from Protestantism together). A study on religion and social capital in Europe that includes Orthodox Greece provides empirical evidence of a lower impact of Orthodox tradition on social participation compared to Protestantism (Traunmüller 2011, 148-152). So Protestantism probably has a stronger positive impact than Catholicism and Orthodox traditions on social participation in CEE. 2.3. Why is social participation important for democratisation? Political culture theory postulates that social participation is relevant as a precondition for successful democratisation in the societal realm. This line of thinking claims that, in order to be stable and effective, a political system requires a certain pattern of supportive and participative orientations (Almond/Verba 1963). The most desirable and most advanced political culture is a so-called civic culture, in which a substantial portion of citizens supports the system and its institutions and participates in the political decision-making processes. Civic culture can be seen as an ideal form of political culture, and a goal toward which societies should strive. As such, this goal should be especially relevant for young democracies, where democratic processes need time and practice to be fully applied as the only game in town. Social participation, understood as citizen participation in free organisations and associations, therefore plays a crucial role for young democracies supporting democratisation. Following the Tocquevillian argumentation, all kinds of free associations are schools of democracy in which citizens learn civil skills through practise. Tocqueville and Putnam argue that the content of organisation s activity is less important than the fact that, through their involvement, people learn to cooperate and act together in order to achieve a common goal (Putnam 1993, 90 ff.). This is how they learn to be better citizens. Yet social organisations are expected to contribute to the social integration of society, for instance through their heterogeneity. A heterogeneous association offers its members the opportunity to meet people of a different social status, religious denomination or race (Wuthnow 2002, 669ff.; Roßteutscher 2009, 446). In other words, the more heterogeneous the members of an organisation are, the more integrative and inclusive it is, because it helps to build bridges among individuals in a society. These particular types of organisations are usually called bridging, and they are supposed to support stronger social cohesion than bonding groups (Putnam 2000, 22 ff.). And social cohesion is definitely desirable for the stability of democracy. In contrast, bonding organisations are more homogenous and isolated than bridging organisations. They support in-group connectedness, but are not open to individuals from outside. Scholars claim 5

that bridging and bonding organisations have different impacts, especially as regards democratic stability. Three kinds of associations in particular are said to have a bonding character: sport associations, religious organisations and labour unions. Their members are, on average, less likely to join other types of organisations. Therefore these three types of organisations seem to contribute to a lesser extent to the stabilisation of democracy (Paxton 2002, 271). However, other profound empirical studies provide a more complex picture concerning the bonding and bridging functions of religious organisations. Some scholars claim that, under some circumstances, religious organisations can build bridges in a society, depending on several factors such as size, structure and connection to a highly organised church structure (Wuthnow 2002, 681; Roßteutscher 2009). In light of these findings, this analysis focuses on bridging organisations; religious organisations are analysed separately. Religion and church-related engagement played an important role in sowing the seeds of democracy in CEE even before the beginning of liberalisation. The Catholic Church contributed to the collapse of communist regimes, especially in Central Europe (Luers 1990, 86ff.; Bernhard 1993, 319; Wnuk- Lipiński 2007, 35ff.), as mentioned earlier. In the course of liberalisations across CEE, the restrictions on religion and church activities introduced by communist regimes were lifted (Krizek 2003, 153-154; Žaloudková 2003, 249; Kamiński 2008, 15). Churches regained their importance in the social and political spheres, especially the Catholic Church in Central Europe and the Orthodox Church in Eastern European countries. In several countries, churches regained their assets and properties, which allowed them to redevelop their social activities, severely limited under communism (Borowik 2000, 109; Krizek 2003, 191-197, 220; Žaloudková 2003, 273-280; Kamiński 2008, 17). These regulative and institutional changes coincided with changes in the religious behaviours and beliefs of people in CEE. Several studies report a rise in religiosity in certain post-communist societies after liberalisation (Tomka 2002; Meulemann 2004, 59-60; Müller 2009, 80). Therefore if religion has a positive influence on social participation, we could expect a rise in social participation in the CEE countries that experienced a religious revival. The objective of this paper is to examine whether religious revival contributes to the rise in social participation in CEE societies. It also analyses whether religious denomination matters in particular whether Protestantism is a more favourable factor than other denominations in enhancing social participation, as is claimed in the literature. Starting the analyses at the country level, we examine the relationship between the religious revival and changes in levels of social participation. We expect an increase in levels of participation in societies that have experienced a religious revival. In the second step, we analyse the relationships 6

between religiosity and social participation at the individual level. A number of factors that are relevant for explaining social participation (such as socio-demographic characteristics) are included in order to isolate the effects of the religious factors in which we are interested. If the effects of religiosity are still positive and significant after the inclusion of these control variables, it means that the religious revival probably contributed to the rise in social participation. Furthermore, including denominations at the individual level (Protestant, Catholic and Orthodox) allows us to find out whether Protestantism has a stronger positive influence on social participation compared to other religious traditions. We also include the denomination at the country level to see whether a society s religious context affects social participation. 3. Religious revival and social participation in CEE The WVS contains a range of questions on social participation and religious behaviour in several CEE countries. A cumulative dataset containing data from all five waves of fieldwork, from the earliest in 1981 until the latest finished in 2007 allows us to follow changes in religiosity and social participation across post-communist societies after liberalisation. Fourteen countries from CEE are included in the dataset: Bosnia and Herzegovina, Bulgaria, Croatia, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Russia, Slovakia, Slovenia and Ukraine. For most of these societies, the cumulative dataset of WVS contains data gathered in two waves: the third (1994 99) and fifth (2005 07). The exceptions are Hungary (first and third waves) and Slovakia and the Czech Republic (second and third waves). In Croatia and all three Baltic societies, the fieldwork was pursued only during the third wave. Appendix Table A1 presents information on the dataset structure concerning the country and wave during which the fieldwork was pursued. 3.1. Religious revival in CEE Religious revival can be captured in various aspects of religiosity: subjective religiosity (the feeling of belonging to a certain religious tradition), belief in God and practising religious rites (e.g., prayer, attending religious services, etc.). In this paper we are particularly interested in two aspects of religious revival that can be related to social participation: embeddedness in church structures (for instance through churchgoing) and belonging to a certain religious tradition (see: Section 2.1). Beliefs and individual religious practices are not expected to be directly relevant to social participation. Before we analyse the religious revival in CEE, let us first figure out how important religion is to people in particular societies of the region. If religion is not an important aspect of social life, we 7

Percentage of religious people should not expect it to have a considerable influence on social behaviour. The relevance of religion can be measured by the level of subjective religiosity 2 in a society. In countries with higher levels of subjective religiosity, we can expect religious tenets to play an important role in people s decisions and integration into religious networks, which can affect social participation. So, how religious feel people in CEE? Figure 1 displays the distribution of subjective religiosity in CEE societies. 3 Figure 1 Subjective religiosity in CEE societies 100 75 50 25 0 RO PL CR SK LT HU SN BH BU UK LV RU CZ ES Source: WORLD VALUES SURVEY 1981-2008 OFFICIAL AGGREGATE v.20090901, 2009. World Values Survey Association Subjective religiosity in post-communist societies is quite unevenly distributed. There are two absolute leaders that have high levels of religiosity: Poland and Romania, where almost everybody says they are religious. The other extreme is Estonia, where less than one-third of people declare themselves as religious. Interestingly, the majority of Catholic societies (five out of six) are placed to the left-hand side of Figure 1, while the majority of Orthodox societies (four out of five) are placed at the opposite side. The exceptions are Orthodox Romanians, who have high subjective religiosity, and Catholic Czechs, who have extremely low subjective religiosity. Estonia and Latvia, historically, Protestant, have less religious populations. In most of the countries, more than half of the population claims to be religious, which suggests that religion is an important aspect of social life in CEE. As we are interested in the religious revival in this region, we analyse changes in subjective religiosity over time after the collapse of communism. Unfortunately we can follow these changes in only four 2 The question wording of the item is following: Independently of whether you attend religious services or not, would you say you are: A religious person (1), Not a religious person (2), An atheist (3). Subjective religiosity at the macro level is measured by percentage of people in a society that declare being religious. 3 The following country name abbreviations are used in the graphs: BH Bosnia and Herzegovina, BU Bulgaria, CR Croatia, CZ Czech Republic, ES Estonia, HU Hungary, LV Latvia, LT Lithuania, PL Poland, RO Romania, RU Russia, SK Slovakia, SN Slovenia, UK Ukraine. 8

Changes in subjective religiosity (percentage points) societies, 4 since in the other countries this question was not asked in two waves of questioning. Figure 2 presents the changes in subjective religiosity in percentage points for these four countries. Figure 2 Changes in subjective religiosity 20 10 0-10 -20 Source: WORLD VALUES SURVEY 1981-2008 OFFICIAL AGGREGATE v.20090901, 2009. World Values Survey Association Subjective religiosity rose in three of the four analysed countries, though at different rates. While Bosnia and Herzegovina experienced a rise of only 5 percentage points, in Russia 20 per cent more people declared they were religious. Only in Hungary did fewer people declare they were religious in the third wave than in the first wave. The increase in Slovakia is modest compared to that of Bosnia and Russia, which is not surprising due to the ceiling effects: if about 80 per cent of the Slovak population is already religious, it is very difficult to experience a large increase in religiosity. We can cautiously conclude that religion has become more important for people in CEE, with the exception of Hungary. Yet the number of analysed countries is insufficient to draw more generalised conclusions for the whole region. Therefore in the next step we examine religious revival by analysing the changes in adherence to particular denominations in single societies (the data is available for 10 countries from CEE). SK HU BH RU In general, three denominations are the most common in the region: Catholic, Orthodox and Protestant, 5 and the proportions of adherents are different across CEE societies. The changes in religious adherence 6 to the three main denominations are presented in Figure 3. 4 The changes in subjective religiosity can be measured at different points of time in the following societies: in Slovakia and Russia (Waves 3 and 2), in Hungary (Waves 3 and 1) and in Bosnia and Herzegovina (Waves 5 and 3). 5 Bosnia and Herzegovina is an exception, with Islam as the dominant religious tradition (54 per cent of the population is Muslim, 28 per cent Orthodox and 17 Roman Catholic). 6 The question is worded as follows: Do you belong to a religion or religious denomination? If yes, which one? No: do not belong to a denomination (0), Roman Catholic (1), Protestant (2), Orthodox (Russian/Greek/etc.) (3), Jew (4), Muslim (5), Hindu (6), Buddhist (7), Other. 9

Changes in religious adherence (percentage points) Figure 3 Changes in religious adherence to three main religious traditions 10 5 0-5 -10 RU SK UK BU CZ BH PL RO SN HU Source: WORLD VALUES SURVEY 1981-2008 OFFICIAL AGGREGATE v.20090901, 2009. World Values Survey Association There are three general trends concerning changes in religious adherence in post-communist societies: (1) a shift from one (or more) religious traditions to another (in Czech Republic, Romania, Hungary, and Bosnia and Herzegovina); (2) a general rise in adherence to all denominations at once (Russia and Bulgaria) and (3) little change in religious adherence (Ukraine, Poland and Slovenia). The only society with a general drop in religious adherence is Slovakia, which experienced a withdrawal from the Catholic Church and quite a stable adherence to Protestantism and Orthodoxy. We can conclude from this analysis that there is evidence of a religious revival in the sense of growing adherence to a specific religious tradition in two Orthodox societies: Russia and Bulgaria. Both are relatively secularised, with around 60 per cent of people considering themselves to be religious (see Figure 1), so a religious revival still has the potential for growth in these countries. In other relatively secularised societies, such as the Czech Republic and the Ukraine, we find no convincing evidence of a religious revival. On the other hand, there is no decline in religious adherence except from Slovakia. In sum, Orthodoxy in particular seems to be awakening from the enforced secularism under communism, in contrast to other religious traditions in the region, which have more or less stable levels of adherence. Protestants Orthodox Catholics Let us turn now to the structural aspect of religiosity and analyse whether the religious revival is related to church attendance. 7 Figure 4 presents levels of churchgoing across CEE societies. In all analysed countries, at least 50 per cent of people declared they attended religious services at least once a year. We have the same two outliers: Poland and Romania, where nearly the entire society declares itself to be churchgoers. The lowest level of church attendance is in the Czech Republic at 7 Churchgoing is operationalised as attendance of church services. The question is: Apart from weddings, funerals and christenings how often do you attend religious services these days?. The answers are recoded from 1 ( more than once a week ) to 5 ( never ). 10

Chaneges in percentage points % of people attending religious services at least once a year nearly 50 per cent. Five out of ten societies analysed have experienced a substantial increase in churchgoing (see Figure 5). Figure 4 Churchgoing in CEE societies 100 75 50 25 0 PL RO BH LT SK CR UK BU LV SN RU ES HU CZ Source: WORLD VALUES SURVEY 1981-2008 OFFICIAL AGGREGATE v.20090901, 2009. World Values Survey Association A considerable outlier is Hungary, where churchgoing decreased by 15 percentage points (Figure 5). 8 Furthermore, in Poland and Romania there is almost no change in churchgoing rates, which is not surprising since virtually everybody in these societies declares they attend church at least once a year. The highest increase in churchgoing occurs in Russia, with 15 per cent point more in the fifth wave, followed by Slovakia, Ukraine and Bulgaria. There are virtually no changes in Bosnia and Herzegovina and Slovenia approximately 1 percentage point so churchgoing is rather stable over time in both countries. Wave 1 Wave 2 Wave 3 Wave 5 Figure 5 Changes in churchgoing rates in CEE societies 16 8 0-8 -16 RU SK UK BU CZ BH PL RO SN HU Source: WORLD VALUES SURVEY 1981-2008 OFFICIAL AGGREGATE v.20090901, 2009. World Values Survey Association Thus several societies in CEE experienced a religious revival in terms of a rise in churchgoing rates. This revival occurs mostly in Orthodox societies, and to a lesser extent in Catholic countries. 8 The problem with the Hungarian result is the different times of measurement: in most of the remaining societies we compare the churchgoing rates between the third and the fifth waves of the WVS, while in Hungary we only have data gathered in the first and third waves. 11

Unfortunately we cannot assess the religious revival in Protestant countries using WVS data, because churchgoing is addressed in only one wave in both societies (Latvia and Estonia). Summarising the analyses of religiosity, the picture seems to be consistent: Orthodoxy underwent a religious revival it has more adherence and more believers taking part in religious services. This applies to a lesser extent to Catholic societies. Yet Figure 1 shows that Catholics are, in general, more religious than Orthodox believers, so perhaps the Orthodox churches just have more social space to develop, while the Catholic Church has already reached quite a high level of adherence and church attendance. 3.2. Social participation in CEE When examining social participation, I refer in this paper to membership in voluntary associations and organisations. The WVS reports membership in the following nine types of organisations: (1) church or religious; (2) sport or recreation; (3) art, music or educational; (4) labour unions; (5) political parties; (6) environmental; (7) professional; (8) charitable or humanitarian; (9) any other. Respondents are asked whether they are an active or inactive member or not a member at all. Membership in labour unions and sport and recreation organisations are excluded from the analyses due to their bonding character (see Section 2.3). Membership in religious organisations is examined separately for two reasons. First, religious organisations do not clearly have a bridging character (Section 2.3). Secondly, the power of religion related to the key explanatory variables churchgoing and religious adherence is considerably higher in reference to religious and church-based participation than to secular participation (Lasinska 2013, 214). This is why we analyse participation in secular bridging associations separately from participation in religious organisations. Furthermore, we build two kinds of measures of secular participation: active member and member. For democracy, active participation is the most desirable, because through active engagement in organisations, people practice their civic skills and build networks through personal interactions. However, just being a member of an organisation also provides a feeling of connectedness to a community of members. So the measure member encompasses both active and passive forms of membership in secular bridging organisations, while the measure active member refers only to individuals who declare active membership in secular bridging organisations. Both measures refer to the number of organisations to which respondents belong. Both variables can have a value from 0 not a member in any organisation to 6 a member in all six types of organisations (after excluding religious, sport associations and labour unions, there are six types of secular 12

Number of organisations per capita Number of organisations per capita bridging organisations the questionnaire refers to). We calculate the mean levels of membership in each country in order to compare participation levels across countries. Figure 6 depicts the mean levels of active membership and membership in CEE societies. Except for Poland (which is measured only in the fifth wave), membership is measured in the third wave, while three countries (Romania, Russia and Slovenia) also have participation information in the fifth wave. Obviously, levels of membership are consistently higher than active membership in all societies. On average, respondents are members of 0.4 organisations, while active in 0.16. In only one country Bosnia and Herzegovina average membership exceeds one organisation. In Croatia, Czech Republic and Slovenia it reaches 0.5, while the lowest levels are found in Ukraine, Russia and Bulgaria (below 0.2). As far as active membership is concerned, Bosnia and Herzegovina ranks first again, with 0.4 organisations per capita. Czech Republic, Slovenia and Slovakia have 0.2 active memberships per capita, and the level is considerably lower in the remaining countries. The lowest levels of active participation are found in Bulgaria, Estonia and Russia. In the three countries for which we have participation data at two points in time, we can see a rise in participation in Slovenia and Russia and a sharp decrease in Romania. Figure 6 Participation in secular bridging organisations in CEE societies 1.4 1.2 1 0.8 0.6 0.4 0.2 0 Active membership BH BU CR CZ ES HU LV LT PL RO RU SK SN UK 1.4 1.2 1 0.8 0.6 0.4 0.2 0 Membership BH BU CR CZ ES HU LV LT PL RO RU SK SN UK Wave 3 Wave 5 Wave 3 Wave 5 Source: WORLD VALUES SURVEY 1981-2008 OFFICIAL AGGREGATE v.20090901, 2009. World Values Survey Association Let us turn now to the examination of religious participation. The measure of religious membership is slightly different from secular membership at the aggregated level: we compare the percentages of respondents that are active or inactive members in a religious organisation across countries (Figure 7). As expected, average levels of active membership (15 per cent in the third wave) are lower than average levels of inactive memberships (7 per cent). The differences across the countries are considerable. While in Croatia almost 40 per cent of respondents belongs to a religious organisation, followed by Bosnia and Herzegovina (30 per cent), and Romania, Slovakia and Slovenia (which each has around 20 per cent), in Bulgaria and Russia the inactive religious participation does not exceed 4 per cent. The patterns of active religious participation are similar Croatia, Bosnia and Herzegovina, and Romania take a leading position, while the lowest levels (below 2 per cent) are reported in Bulgaria, Russia and Estonia. Changes in religious participation vary from country to country. While in 13

Active members in religious organisation (in percentage) Inactive member in religious organisation (in percentage) Bulgaria there is virtually no change in the level of religious participation, in Russia only inactive membership doubled from 4 per cent to almost 9 per cent. In Slovenia there are more active members in religious organisations, but fewer inactive members. In Ukraine the direction of change is the same: the percentage of active and inactive members rose between the third and waves. The developments in Romania are the opposite: the level of religious participation dropped considerably. Figure 7 Participation in religious organisations in CEE societies 40 30 Active religious membership 40 30 Inactive religious membership 20 20 10 10 0 BH BU CR CZ ES HU LV LT PL RO RU SK SN UK 0 BH BU CR CZ ES HU LV LT PL RO RU SK SN UK Wave 3 Wave 5 Wave 3 Wave 5 Source: WORLD VALUES SURVEY 1981-2008 OFFICIAL AGGREGATE v.20090901, 2009. World Values Survey Association In sum, the changes in religious participation reveal no consistent patterns: in two Orthodox countries there was a rise in religious inactive membership, in two there was a drop (one Orthodox and one Catholic), while in one Orthodox country the situation did not change at all. Active membership rose in one Orthodox country and one Catholic, dropped in one Orthodox and remained stable in two Orthodox societies. Let us now compare the direction of changes in religiosity (religious adherence and churchgoing) and participation (secular and religious) at the country level. Table 1 presents an overview of the direction of changes across CEE countries discussed above. Unfortunately, a comparison of changes in all five measures of religiosity and participation is possible in only three countries: Romania, Russia and Slovenia (and partly in Ukraine). Among these societies, only Russia has the same direction of changes in all five measures of religiosity and participation. Otherwise, Ukraine also experienced a rise in both churchgoing and religious participation. In Romania, one of the most religious societies in this region, there is a drop in participation, though there are no changes in churchgoing and a shift in religious adherence. In Slovakia the trend is the opposite: rising participation and no changes in religiosity. 14

Table 1 Changes in religiosity and participation in CEE Religiosity Participation Country Subjective religiosity Religious adherence Churchgoing Secular Religious Bosnia and Herzegovina + Shift No change n.a. n.a. Bulgaria n.a. + + n.a. No change Czech Republic n.a. Shift + n.a. n.a. Hungary - Shift - n.a. n.a. Poland n.a. no change No change n.a. n.a. Romania n.a. Shift No change - - Russia + + + + + Slovakia + - + n.a. n.a. Slovenia n.a. No change No change + +/(-) Ukraine n.a. No change + n.a. + n.a. = not available To conclude, we found a positive relationship between religious revival and participation in one only country (Russia) and also partly in Ukraine 9 (but only in reference to religious participation). However we did not find any negative relationships between changes in religiosity and participation. The problem we face is the lack of systematic survey data to analyse the relationship between religious revival and participation in CEE at the country level. 3.3. Churchgoing and social participation empirical individual-level examination The analyses at the macro level reveal a positive relationship between religious revival and participation in CEE, though to a very limited extent. Therefore we need to address this question by analysing the relationship between religiosity and participation at the individual level. The evidence at the micro level of analyses allows us to indirectly answer our question about whether religious revival contributes to a rise in participation: when the relationship between religiosity and social participation is positive, it is possible that an increase in the number of devoted people in a society results in more participants in social life. However, as some scholars argue, at a certain level of religiosity, a larger number of churchgoers does not further contribute to participation (Ruiter/De 9 In both countries, correlation coefficients are very high (r=0.9). 15

Graaf 2006, 204; Lasinska 2013, 167-168). The reason is a kind of saturation effect, when all potential participants have already been recruited through religious networks and religious motivation, so that a rise in religiosity has no further effect on participation. Regardless, there are only two CEE countries with very high levels of religiosity: Poland and Romania. And in these countries, religion does not need a revival, since virtually everybody is already religious. All other countries, however, seem to be far from reaching the highest levels of religiosity, so recruitment through religious networks and motivation through religious adherence in these countries is far from saturated. We continue the examination at the individual level and address the question: Is the relationship between religiosity and participation also positive at the individual level in CEE? As shown in Section 3.1, religious revivals have occurred mostly in Orthodox societies. On the other hand, Protestantism usually has a stronger effect on participation than the Catholic tradition (see Section 2.2). Therefore we ask: Does denomination affect the religious influence on participation? In order to answer these questions, we run regression analyses with participation as the dependent variable. Participation is operationalised in three ways, similar to Section 3.2. First, we analyse active membership referring to the number of different types of secular bridging organisations in which an individual is active. Second dependent variable, membership (in a broader sense) is calculated as the number of secular bridging organisations in which an individual is an active or inactive member. Third dependent variable, membership in a religious organisation can take a value from 0 (not a member), 1 (inactive member) to 2 (active member). For Romania, Russia and Slovenia we use data from the newest (fifth) wave, skipping the data from the third wave, and for the remaining countries the third wave of questioning is considered. For each dependent variable, we build three models in order to examine the changes in the power of key religious explanatory factors with the introduction of various control variables. In the first model we consider a single key explanatory variable (churchgoing) together with gender, age and education. 10 Education serves as a simple indicator of socio-economic status, which is usually considered the most important predictor of social participation (Almond/Verba 1963, 383ff.; Verba et al. 1995, 281; Gabriel et al. 2002, 98-99). Similar men are usually more engaged than women, while age has a curvilinear relationship to participation (Gabriel et al. 2002, 98-99). We expect, 10 Variable gender is dichotomized: male (1), female (0); for age we take year of birth; and education is operationalised as the highest degree attained. The question wording is: What is the highest educational level that you have attained? and answers rank from: No formal education (1) to University-level education, with degree (9). 16

churchgoing to be positively related to all three measures of participation, as described in Section 2.1, and religious participation is expected to capture more variance than secular participation. The aim of the second model is to examine the relevance of denomination adherence for explaining social participation in CEE. As discussed in Section 2.2, Protestantism raises the likelihood of social participation at the individual and macro levels. We introduce variables concerning denominations at both the individual and macro levels. At the individual level we use dummies for the three most popular denominations in the region: Protestant, Catholic and Orthodox. Operationalising of denomination at the macro level is more complicated. Some scholars use the category predominant denomination in a society (Norris/Inglehart 2005), but in societies with mixed traditions such as Hungary (Catholic and Protestant) or the Baltic societies (Protestants and Orthodox) this simple categorisation can provide distorted results, because it is difficult to assess which denomination affects society to a larger extent. This is why we introduce variables measuring the percentage of believers that belongs to each of the three main denominations in a society. This method allows us to assess how much each of the three denominations in the region influences participation, while avoiding the pitfall of categorising societies as having a predominant religious tradition. Finally, the last model also considers country dummies in order to control for country-specific influences on individual participation. These last models provide evidence of the isolated effects of churchgoing and denominations on social participation. If the effects remain significant even after considering country-specific effects, we can answer the questions of whether religious revival matters for social participation and what is the role of denominations. 3.3.1. Active membership in secular bridging organisations Let us start with the analyses of the relationship between active membership operationalised as the number of secular bridging organisations of which one is an active member and churchgoing. The results of the regression analysis using active membership as a dependent variable are presented in Table 2. 17

Table 2 Linear regression analyses of active membership in secular bridging organisations in CEE (unstandardised regression coefficient B, R 2 and N, unweighted) Independent variables Model 1 Model 2 Model 3 RELIGIOUS FACTORS Churchgoing (1 = more than once a week; 5 = never) -0.028*** -0.029*** -0.018*** Protestant (dummy) -0.005 0.115*** Catholic (dummy) -0.046 0.063** Orthodox (dummy) -0.020 0.126*** Percentage of Protestants -0.001** -0.022*** Percentage of Catholics -0.000 0.003*** Percentage of Orthodox -0.002*** 0.013*** SOCIODEMOGRAPHIC INDIVIDUAL CHARACTERISTICS Education (1 = elementary; 8 = university degree) 0.031*** 0.032*** 0.033*** Age (year of birth) -0.001* 0.000-0.000 Gender dummy (1= male) 0.053*** 0.054*** 0.040*** Constant 1.241** 0.017 1.237* N 13,780 9,709 9,709 Adjusted R 2 0.023 0.034 0.088 Note: significance levels: *p<0.05, **p<0.01,*** p<0.001 Source: WVS cumulative dataset (Waves 1 5). In general, we can see that the key explanatory variable, churchgoing, is positively related to secular participation, as expected. So the likelihood of social engagement rises with the frequency of church attendance, and this effect is statistically significant, though not very strongly. This relationship remains statistically significant over all three models, meaning that even when we control for country-specific factors that influence social participation in Model 3, church attendance still matters for explaining social participation. Furthermore, introducing factors related to individual- and country-level denomination characteristics does not really weaken the effect of church affiliation. Only after considering country dummies in Model 3 does the explanatory power of churchgoing drop by one-third. With the introduction of new variables in Models 2 and 3, the level of explained variance rises from 2 per cent to almost 9 per cent in Model 3, which contains the country dummies. Individual adherence to one of the three most popular denominations in the region (Protestantism, Catholicism or Orthodox) has no effect on social participation unless the country dummies are introduced to the model. In Model 3, all three denominations have statistically significant effects. 18

Surprisingly, of the three denominations, Orthodox tradition (rather than Protestantism) most powerfully influences the likelihood of being socially engaged. 11 Catholicism is second, while Protestantism ranks last. This evidence is different from that of Western European societies, where Protestantism is generally the most favourable denomination for social participation. The second surprising finding is that adherence to any of the three denominations has stronger effects on participation than churchgoing (when comparing standardised coefficients). Apparently, social participation can be better explained by religious-driven motivation to participate than by embeddedness in church-related networks in CEE. Similarly, the effects of denominations at the country level are more powerful in the full Model 3 when country dummies are introduced to the model. Interestingly, with a higher percentage of adherence to Protestantism, the likelihood of being socially engaged drops and the effects are statistically significant, though quite weak. In contrast we find positive effects of the country-level adherence to Orthodoxy and Catholicism. Again, this evidence is surprising when compared to the empirical evidence in the literature on Western societies, where Protestantism has a positive impact on social participation. The strongest positive effect is seen with the percentage of Orthodox in a society (when comparing standardised coefficients), which is quite surprising, too. Last but not least, individual socio-demographic characteristics have robust effects in all three models, and in general reflect the influences described in the literature: men are more likely to be socially engaged than women, and socio-economic status positively affects individuals readiness to join connected associations. On the other hand, age has no significant effect. 3.3.2. Membership in secular bridging organisations Let us turn now to the analyses of the determinants of social participation more broadly: membership irrespective of the type of activity (active or inactive), which is operationalised as the number of secular bridging organisations to which one belongs. The results of these regression analyses are presented in Table 3. We use the same three models as above to explore the situation in CEE. 11 Relative power of predictors can be compared with standarised beta coefficients, which is higher for Orthodox believers than for Protestants and Catholics. The results of regression analyses with standarised coefficients are available by the author on request. 19

Table 3 Linear regression analyses of membership in connected organisations in CEE (unstandardised regression coefficient B, R 2 and N, unweighted) Independent variables Model 1 Model 2 Model 3 RELIGIOUS FACTORS Churchgoing (1 = more than once a week; 5 = never) -0.088*** -0.082*** -0.042*** Protestant (dummy) -0.142* 0.258*** Catholic (dummy) -0.226*** 0.143** Orthodox (dummy) -0.238*** 0.256*** Percentage of Protestants -0.006*** -0.075*** Percentage of Catholics -0.002** 0.006*** Percentage of Orthodox -0.004*** 0.044*** SOCIODEMOGRAPHIC INDIVIDUAL CHARACTERISTICS Education (1 = elementary; 8 = university degree) 0.052*** 0.057*** 0.056*** Age (year of birth) 0.000 0.002* 0.000 Gender dummy (1 = male) 0.115*** 0.101*** 0.060** Constant -0.058-0.407 1.714 N 13,780 9,709 9,709 Adjusted R 2 0.020 0.039 0.144 Note: Significance levels: *p<0.05, **p<0.01,*** p<0.001 Source: WVS cumulative dataset (Waves 1-5). There are a number of similarities in the pattern of relationships between religion and participation in the present and previous sections. For example, churchgoing in both cases has a positive influence on membership. This relationship remains statistically significant even after the introduction of variables related to denominations and country dummies, though in the last model its power is weaker than in the first. The consideration of country-specific conditions raises the level of explained variance from 4 per cent to 14 per cent. The influences of the individual denominations are not stable over the models. In the second model which includes only religious-related variables and socio-demographic characteristics adherence to Protestantism, Catholicism or Orthodox traditions seems to have a negative, statistically significant impact on participation. However, the introduction of country dummies changes the situation radically. In the third model, adherence to one of three religious traditions raises statistically significant the likelihood of participation. Again, surprisingly effects of adherence to Orthodoxy are more powerful than effects of adherence to Catholicism and Protestantism (analysis of standardised 20